Vincent S.

Smentkowski Editor

Surface
Analysis and
Techniques in
Biology
Surface Analysis and Techniques in Biology
Vincent S. Smentkowski
Editor

Surface Analysis and

Techniques in Biology
Editor
Vincent S. Smentkowski, Ph.D.
Nanostructures and Surfaces Laboratory
GE Global Research
Niskayuna, NY, USA

ISBN 978-3-319-01359-6 ISBN 978-3-319-01360-2 (eBook)

DOI 10.1007/978-3-319-01360-2
Springer Cham Heidelberg New York Dordrecht London

Library of Congress Control Number: 2014931212

Springer International Publishing Switzerland 2014

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection
with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and
executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this
publication or parts thereof is permitted only under the provisions of the Copyright Law of the Publishers
location, in its current version, and permission for use must always be obtained from Springer.
Permissions for use may be obtained through RightsLink at the Copyright Clearance Center. Violations
are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

Preface

This book came about as a result of a discussion which took place at a Microscopy
and Microanalysis meeting. Although the number of research groups working in the
field of biological surface chemistry, modification and characterization have
increased during the past few decades, a number of advances have been made to
standard surface analytical instrumentation, and a number of new instruments have
been introduced, only two books on the subject of surface analysis of biological
systems have been published (see Refs. [44] and [45] in Chap. 1) and are both now
outdated. We felt the time was right for a book which went into more detail on the
main surface analysis techniques that are being used to study biological specimens
and systems.
The process of editing a book is very rewarding, as you are tasked with identify-
ing best-in-class researchers in their respective fields of study and helping them
assemble and refine the content. I very much appreciate that each of the chapter
authors took time from their busy schedules to write their chapters. The technical
content described in this book is very high. The compilation of chapters will help
the biological research community realize the benefits that surface analysis pro-
vides. We look forward to seeing a larger number of biologists and medical special-
ists start using the techniques discussed in this book.
New analysis instruments (such as the QSTAR and the Ionoptika J105 3D
Chemical Imager, which are discussed in the future outlook section of Chap. 4) are
continuously being developed and introduced to the scientific community; we look
forward to seeing what the future has in store. I am also excited to see the next gen-
eration of medical devices, which will benefit from surface analysis and will help
our society.

Niskayuna, NY, USA Vincent S. Smentkowski

v
Contents

1 Introduction ............................................................................................. 1
Vincent S. Smentkowski
2 Applications of XPS in Biology and Biointerface Analysis ................. 9
Sally L. McArthur, Gautam Mishra, and Christopher D. Easton
3 Biomolecular Analysis by Time-of-Flight Secondary
Ion Mass Spectrometry (ToF-SIMS) ..................................................... 37
Daniel Breitenstein, Birgit Hagenhoff, and Albert Schnieders
4 Cluster Secondary Ion Mass Spectrometry .......................................... 71
Joseph Kozole and Nicholas Winograd
5 Biological Tissue Imaging at Different Levels: MALDI
and SIMS Imaging Combined ............................................................... 99
J. Stauber and Ron M.A. Heeren
6 Molecular Structure and Identification Through
G-SIMS and SMILES ............................................................................. 141
F.M. Green, I.S. Gilmore, and M.P. Seah
7 Imaging with the Helium Ion Microscope ............................................ 171
John Notte and Bernhard Goetze
8 Sum Frequency Generation Vibrational Spectroscopy:
A Sensitive Technique for the Study of Biological
Molecules at Interfaces ........................................................................... 195
Andrew P. Boughton and Zhan Chen
9 Near-Field Scanning Optical Microscopy: A New Tool
for Exploring Structure and Function in Biology ................................ 225
Nicholas E. Dickenson, Olivia L. Mooren, Elizabeth S. Erickson,
and Robert C. Dunn

vii
viii Contents

10 Atomic Force Microscopy: Applications in the Field

of Biology ................................................................................................. 255
J.K. Heinrich Hoerber
11 Multi-technique Characterization of DNA-Modified
Surfaces for Biosensing and Diagnostic Applications .......................... 289
Chi-Ying Lee, Lara J. Gamble, Gregory M. Harbers,
Ping Gong, David W. Grainger, and David G. Castner

Index ................................................................................................................. 315

Contributors

Andrew P. Boughton Department of Chemistry, University of Michigan, Ann

Arbor, MI, USA
Daniel Breitenstein Tascon GmbH, Mnster, Germany
David G. Castner National ESCA and Surface Analysis Center for Biomedical
Problems, Department of Bioengineering, and Department of Chemical Engineering,
University of Washington, Seattle, WA, USA
Zhan Chen Department of Chemistry, University of Michigan, Ann Arbor, MI, USA
Nicholas E. Dickenson Department of Chemistry, Ralph N. Adams Institute for
Bioanalytical Chemistry, The University of Kansas, Lawrence, KS, USA
Robert C. Dunn Department of Chemistry, Ralph N. Adams Institute for
Bioanalytical Chemistry, The University of Kansas, Lawrence, KS, USA
Christopher D. Easton CSIRO Molecular and Health Technologies, Clayton
South, VIC, Australia
Elizabeth S. Erickson Department of Chemistry, Ralph N. Adams Institute for
Bioanalytical Chemistry, The University of Kansas, Lawrence, KS, USA
Lara J. Gamble National ESCA and Surface Analysis Center for Biomedical
Problems and Department of Bioengineering, University of Washington, Seattle,
WA, USA
I.S. Gilmore Analytical of Life Division, National Physical Laboratory,
Teddington, Middlesex, UK
Bernhard Goetze Director of Research and Development, Carl Zeiss Microscopy
in Peabody, Peabody, MA, USA
Ping Gong Department of Pharmaceutics and Pharmaceutical Chemistry,
University of Utah, Salt Lake City, UT, USA
Seventh Sense Biosystems, Cambridge, MA, USA

ix
x Contributors

David W. Grainger Department of Bioengineering and Department of Pharmaceutics

and Pharmaceutical Chemistry, University of Utah, Salt Lake City, UT, USA
F.M. Green Analytical Science Division, National Physical Laboratory, Teddington,
Middlesex, UK
Birgit Hagenhoff Tascon GmbH, Mnster, Germany
Gregory M. Harbers Replenish, Inc., Pasadena, CA, USA
Department of Pharmaceutics and Pharmaceutical Chemistry, University of Utah,
Salt Lake City, UT, USA
Ron M.A. Heeren FOM-Insitute for Atomic and Molecular Physics, Amsterdam,
The Netherlands
J.K. Heinrich Hoerber H.H. Wills Physics Laboratory, University of Bristol,
Bristol, UK
Joseph Kozole Department of Chemistry, Penn State University, University Park,
Philadelphia, PA, USA
DuPont, Philadelphia, PA, USA
Chi-Ying Lee National ESCA and Surface Analysis Center for Biomedical
Problems and Department of Chemical Engineering, University of Washington,
Seattle, WA, USA
Sally L. McArthur Biointerface Engineering Group, IRIS, Faculty of Engineering
and Industrial Sciences, Swinburne University of Technology, Hawthorn, VIC,
Australia
Gautam Mishra Kratos Analytical, Wharfside, Manchester, UK
Olivia L. Mooren Department of Chemistry, Ralph N. Adams Institute for
Bioanalytical Chemistry, The University of Kansas, Lawrence, KS, USA
John Notte Director of Research and Development, Carl Zeiss Microscopy in
Peabody, Peabody, MA, USA
Albert Schnieders Tascon USA, Inc., Chestnut Ridge, NY, USA
Managing Director, CNM Technologies, Bielefeld, Germany
M.P. Seah Analytical Science Division, National Physical Laboratory, Teddington,
Middlesex, UK
Vincent S. Smentkowski Nanostructures and Surfaces Laboratory, GE Global
Research, Niskayuna, NY, USA
J. Stauber FOM-Insitute for Atomic and Molecular Physics, Amsterdam,
The Netherlands
Nicholas Winograd Department of Chemistry, Penn State University, University
Park, Philadelphia, PA, USA
Chapter 1
Introduction

Vincent S. Smentkowski

Abstract The outer layer of bulk samples is referred to in the scientific community
as the surface of the sample or material. At the surface, the composition, microstruc-
ture, phase, chemical bonding, electronic states, and/or texture can be different than
that of the bulk material. The outer surface is where many material interactions/
reactions take place; this is especially true for biomaterials as biomaterials are
intended to interact with the biological system. Breakthroughs in biomaterials and
biomedical devices will require novel approaches to tailoring biological surfaces in
order to generate surfaces with desired properties. Analytical techniques are
required to characterize the surface of biological materials and quantify their impact
in real-world biological systems. Surface analysis of biological materials started in
the 1960s, and the number of researchers working in this area has increased rapidly
since then. Surface analysts are being asked to detect and image species present in
lower concentrations and within smaller spatial dimensions in biomaterials.
Biological samples, which can be highly irregular in shape, can be chemically and
morphologically altered during sample preparation, storage, and/or placement
under high vacuum; the molecules can become damaged when energetic surface
probes are used to analyze them. The complexity of the biological system compli-
cates the analysis of biological samples. The desire to analyze biological samples is
driving the development of new and/or improved surface analysis instrumentation
and data analysis tools. The surface analysis instruments which are most often used
to analyze biological samples are reported in this book.
The outer layer of bulk solid or liquid samples is referred to in the scientific
community as the surface of the sample or material. At the surface, the composition,
microstructure, phase, chemical bonding, electronic states, and/or texture is often
different than that of the bulk material. The outer surface is where many material

V.S. Smentkowski, Ph.D (*)

Nanostructures and Surfaces Laboratory, GE Global Research, 1 Research Circle,
Niskayuna 12309, New York, USA
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 1

DOI 10.1007/978-3-319-01360-2_1, Springer International Publishing Switzerland 2014
2 V.S. Smentkowski

interactions/reactions take place (for instance, corrosion of metals and/or catalysis)

[1, 2]. In 1984, Charles B. Duke wrote a paper entitled Atoms and electrons at
surfaces: A modern scientific revolution, where he summarized the importance of
surface and interfacial phenomena. In the abstract, he stated, Whereas in the mid-
1960s an interface was regarded merely as the boundary between two bulk media,
today it is seen as an independent entity: a state of matter determined by its history
and exhibiting its own unique composition, structure, and electronic properties [1].
Traditionally, surface scientists and analysts have been grouped into three differ-
ent categories: (1) researchers who perform a complete characterization of clean,
single crystal surfaces; (2) researchers who study the interaction of adsorbates onto
clean surfaces under ultrahigh-vacuum conditions; and (3) researchers who are ana-
lyzing as-received, real-life parts. The first two groups of scientists are often trying
to understand fundamental phenomena under well-controlled conditions. The last
group of researchers are often found in an industrial setting and are usually classi-
fied as applied surface scientists/analysts. Often times, their samples contain mul-
tiple components and because the samples are transferred from ambient conditions
into the surface analysis instrument, they are not clean, in contrast with the sam-
ples analyzed by groups 1 and 2 above.
The field of surface analysis rapidly grew in the late 1960s and early 1970s with
the introduction of commercial surface analysis instrumentation [1, 2]. It is of inter-
est and importance to note that the first surface analysis instruments, such as Auger
electron spectroscopy (AES) [35], X-ray photoelectron spectroscopy (XPS) [6],
ion scattering spectroscopy (ISS) [7], and secondary ion mass spectroscopy (SIMS)
[8, 9], were developed at, or in collaboration with, industrial research laboratories
such as General Electric[35], Hewlett-Packard [6], 3M [7], and Knolls Atomic
Power LaboratoryGeneral Electric [8, 9], respectively. These first surface analysis
instruments were designed and built in order to study practical phenomena such as
metallurgy [1013], corrosion science [1416], electronic devices [17, 18], tribol-
ogy [1922], polymers [2326], adhesives and resins [27], and catalysis [2, 2830]
in real material systems. The applications of surface analysis reported at confer-
ences, as well as in the literature, have grown (and changed) with time, reflecting
priorities and research interests. Surface analysts are now being asked to detect and
image species present in lower concentrations and within a smaller spatial dimen-
sion in a new class of materialsbiological samples. Biological samples, which can
be highly irregular in shape, can be chemically and morphologically altered during
sample preparation, storage, and/or placement under high vacuum; the molecules
can become damaged when energetic surface probes are used to analyze them. The
overall complexity of the biological system complicates the analysis of many bio-
logical samples [31]. The desire to analyze biological samples is driving the devel-
opment of new and/or improved surface analysis instrumentation and data analysis
tools. The most commonly used surface analysis instruments are reported in this
book.
A biomaterial is defined as a material used in a medical device; a biomaterial is
intended to interact with the biological system [32]. Biomaterials are fabricated into
bio-devices, which are often implanted into tissues and organs. Surfaces of
1 Introduction 3

biomaterials (synthetic or modified natural materials) are of critical importance

since the surface is typically the only part of a biomaterial/bio-device that comes in
contact with the biological system (e.g., tissues and organs for implanted biomate-
rial/bio-devices); surfaces are where reaction and/or adsorption occurs in biomedical
assemblies [31, 33]. For certain applications, adsorption is detrimental and should be
minimized [e.g., protein adhesion within microfluidic channels, and formation of a
thrombus (an aggregation of blood cells) upon insertion of a biomaterial into the
body], while for other applications, adsorption is needed to facilitate repair or growth
(e.g., adherent cells) [31, 33]. Breakthroughs in biomedical devices will require
novel approaches to tailoring biological surfaces in order to generate surfaces with
desired properties. Analytical techniques are required to characterize the surface of
biological materials and quantify their impact in real-world biological systems.
Surface analysis of biological materials started in the 1960s, and the number of
researchers working in this area has increased very rapidly since then. Today there
are journals devoted to the surface analysis of biological materials and systems [34].
There are numerous exceptional books on practical surface analysis [3543];
however, these books do not have much information regarding biological surface
analysis. Books specific to a single technique often contain a chapter on biological
surface analysis. In 1988, Buddy D. Ratner edited a book, Progress in Biomedical
Engineering 6, Surface Characterization of Biomaterials [44], which summarized a
symposium, sponsored by the American Chemical Society (ACS), that was attended
by biologists, surface scientists, chemists, physicists, materials scientists, and phy-
sicians in Ann Arbor, Michigan, in June of 1987. In the preface to the book, Ratner
compared a surface scientists perception of a surface to that of a biologist. He con-
cluded that on the whole, biologist will not invoke surface-induced effects in their
hypotheses. The surface scientist, on the other hand, will consider the problems of
biology as being too complex and disorderly to be dealt with using the tools avail-
able. There is a wide gap in understanding between these two disciplines, but there
are signs that it is narrowing. About 25 years later, many of these gaps still remain
although biologists are now realizing the importance of the outer surface of bioma-
terials and surface analysts are starting to analyze biomaterials. Most important,
research teams are being formed that have members from both disciplines, and they
are working together and cross-training each other. Since the release of the Ratner
book, there have been a number of new surface analysis instruments, as well as
numerous improvements to the traditional surface analysis instruments, that are
facilitating biological sample analysis; many of these will be outlined in this book.
In 1996, John Davies edited the book Surface Analytical Techniques for Probing
Biomaterial Processes, which summarized four techniques commonly used by biol-
ogists for probing dynamic, in situ processes at the biomaterial interfaces [45]. The
book described total internal reflection fluorescence spectroscopy (TIRFS), surface
plasmon resonance (SPR), ellipsometry, and dynamic contact angle (DCA). To
date, the Ratner and Davis books are the only two books that have been published
on the surface analysis of biological materials.
The concept for the present book originated from numerous discussions with
surface analysts who are now starting to analyze biological samples as well as
4 V.S. Smentkowski

researchers in the medical and biological communities who have indicated that they
are not aware of surface analysis techniques and that having one comprehensive
book on the most commonly used surface analysis techniques being used to analyze
biological materials would be very useful. It is anticipated that this book will result
in an increase in the use of surface analysis techniques by researchers in the medical
and biological communities. Although the examples described in this book are bio-
logical, the book will also be useful for researchers analyzing other materials (met-
als, alloys, ceramic, etc.) since this book represents a compilation of the current,
state-of-the-art surface analysis techniques.
This book starts with a chapter on X-ray photoelectron spectroscopy (XPS),
which is a widely used instrument as it detects all elements except hydrogen and
helium; changes in peak position (binding energy) provide information about the
chemical environment of the elements. The second chapter is on secondary ion mass
spectrometry (SIMS). SIMS allows for the detection of all elements, with a high
(ppm or better) chemical sensitivity; most state-of-the-art SIMS instruments allow
for 3D analysis as well as the collection of a full mass spectrum at every voxel. In
order to maximize the yield of high-mass molecular information and minimize deg-
radation of the molecular information during SIMS analysis, the time-of flight sec-
ondary ion mass spectrometry (ToF-SIMS) community is devising novel cluster ion
sources. Chapter 3 describes the benefits of these new cluster SIMS ion sources via
examples. A number of research groups have also shown that the application of a
metal or matrix to a biological specimen prior to SIMS analysis can also increase
the amount of molecular information available. Chapter 4 describes metal-enhanced
and matrix-enhanced SIMS and compares these approaches to matrix-assisted laser
desorption ionization (MALDI) mass spectroscopy. Compared to SIMS, MALDI is
able to detect higher mass species but has degraded lateral resolution. SIMS and
MALDI provide complementary information about samples. Both SIMS and
MALDI data sets are very complex and contain a wealth of information about the
sample that was analyzed. The scientific community is devising data analysis strate-
gies to facilitate data reduction. Multivariate statistical analysis (MVSA) algorithms
are proving to be very valuable in this regard; readers interested in MVSA are
referred to a special, two-issue series on MVSA that was published in Surface and
Interface Analysis (SIA) in 2009 [46, 47]. Chapter 5 introduces the reader to gentle
secondary ion mass spectrometry (G-SIMS) and SMILES (simplified molecular
input line entry specification). G-SIMS provides information about the molecular
structure that is not directly available from the as-collected mass spectrum. G-SIMS
often allows the identification of unknown materials without the need for experi-
mental library spectra. SMILES is used to simulate the fragmentation pathways that
occur in G-SIMS. Helium ion microscopy (HIM) is a relatively new technique that
is showing benefits compared to scanning electron microscopy (SEM) imaging,
including the ability to analyze insulating samples without an overcoat, a higher
lateral resolution, and a higher depth of field. HIM is described in Chap. 6. Chapter 7
summarizes scanning probe microscopy (SPM), which is a general term that is used
to describe a family of instruments that measure the interaction of a small tip with a
sample surface at close distances in order to provide the topography of the surface.
1 Introduction 5

Many SPM instruments can provide information on a materials phase, capacitance,

magnetic properties, and electrochemical properties (to name a few) while simulta-
neously recording the topography. Near-field scanning optical microscopy (NSOM),
also referred to as scanning near-field optical microscopy (SNOM), instruments use
specially fabricated probes to deliver light down to the nanometric dimension,
enabling optical microscopy with a spatial resolution of tens of nanometers. NSOM
instruments can simultaneously map sample fluorescence and topography with a
high spatial resolution and single-molecule detection limits. NSOM is summarized
in Chap. 8. Sum frequency generation (SFG) is a nonlinear optical vibrational spec-
troscopic technique with excellent sensitivity to interfacial molecules and molecular
ordering; SFG is well suited to probing biomolecules in a native interfacial environ-
ment in order to provide information on biomolecular orientation and conformation
at interfaces. SFG is summarized in Chap. 9.
Chapter 10 describes the benefits of using complementary surface analysis tech-
niquesX-ray photoelectron spectroscopy (XPS), near-edge X-ray absorption fine
structure (NEXAFS), time-of-flight secondary ion mass spectrometry (ToF-SIMS),
and surface plasmon resonance (SPR)to characterize the structure and composi-
tion of DNA-modified surfaces.
Importantly, new instruments and data analysis protocol are being designed and
tested with the aim of providing more molecular information about biological
samplesthe future looks very encouraging. A final significant point to be made is
that the most successful groups working in the field of surface analysis of biological
materials have biologists as part of their research team. The biologists are bringing
their surface analysis colleagues up to speed on their terminology, experimental
methods, and so forth, and the surface analysts are then able to relate their findings
back to the biologists using the terminology the biologists are accustomed to using.
Having a common language that is understood by both parts of the team is benefi-
cial. The biologists and the surface analysts are also working together to determine
the best sample preparation methods. The gap that Ratner mentioned between these
two disciplines [44] is indeed narrowing.

References

1. Duke CB. Atoms and electrons at surfaces: a modern scientific revolution. J Vac Sci Technol
A. 1984;2(2):139.
2. Somorjai GA, Li Y, editors. Introduction to surface chemistry and catalysis. 2nd ed. Hoboken:
Wiley; 2010.
3. Harris LA. General electric report no 67-C-201 (1967). J Appl Phys. 1968;39:1419.
4. Harris LA. Miscellaneous topics in Auger electron spectroscopy. J Vac Sci Technol.
1974;11:23.
5. Gergely G. Vacuum. Commemoration of the 25th anniversary of Auger electron spectroscopy.
1994;45(2/3):311.
6. In cooperation with Siegbahn, a group of engineers from Hewlett-Packard in the USA (Mike
Kelly, Charles Bryson, Lavier Faye, Robert Chaney) produced the first commercial monochro-
matic XPS instrument in 1969. http://en.wikipedia.org/wiki/X-ray_photoelectron_spectros-
copy. Kelly MA. J Chem Educ. 2004;81(12):1726.
6 V.S. Smentkowski

7. Smith DP. Wang PS, Moddeman WE, Haws LD, Wittberg TN, Peters JA. Surface studies of
plastic-bonded ETN and RDX by X-ray photoelectron spectroscopy (XPS) and ion-scattering
spectroscopy (ISS). J Appl Phys. 1967;38:340; http://www.osti.gov/bridge/servlets/
purl/6254094/6254094.pdf
8. McHugh JA, Sheffield JC. Secondary positive ion emission from a tantalum surface. J Appl
Phys. 1964;35(3):512.
9. McHugh JA, Sheffield JC. Mass spectrometric determination of beryllium at the sub-nano-
gram level. Anal Chem. 1967;39(3):377.
10. Lea C. Boron inhibition of oxidation on Fe-10%Cr studied by Auger electron spectroscopy.
Met Sci. 1979;13:301.
11. Edmons DV, Jones PN. Interfacial embrittlement in liquid-phase sintered tungsten heavy
alloys. Metall Trans. 1979;10A:289.
12. Stoddart CTH, Lea C, Dench WA, Green P, Pettit HR. Relationship between lead content of
Cu40Zn, machinability, and svvarf surface composition determined by Auger electron spec-
troscopy. Met Technol. 1979;6:176.
13. Waters RE, Charles JA, Lea C. Prior particle boundaries in hot isostatically pressed nickel-
based superalloy, studied by Auger electron spectroscopy. Met Technol. 1981;8:194.
14. Castle JE. The use of X-ray photoelectron spectroscopy in corrosion science. Surf Sci.
1977;68:583.
15. Stout DA, Gavelli G, Lumsden JB, Staehle RW. In Situ AES and ESCA Analysis of Iron
Oxides Formed by a Galvanic Cell. Appl Surf Anal ASTM STP. 1980;699:42.
16. Mathieu JB, Mathieu HJ, Landolt D. Electropolishing of titanium in perchloric acid-acetic
acid solution: I. Auger electron spectroscopy study of anodic films. J Electrochem Soc.
1978;125:1039.
17. Holloway PH. Characterization of electronic devices and materials by surface-sensitive ana-
lytical techniques. Surf Sci. 1980;4:410.
18. Pignataro S. Surface analysis in microelectronics. Anal Bioanal Chem. 1995;353(34):227.
19. Buckley DH. Friction-induced surface activity of some simple organic chlorides and hydrocar-
bons with iron. ASLE Trans. 1974;17:36.
20. Jones RG. Halocarbon adsorption on Fe(100): The adsorption of CCl4 studied by AES, LEED,
work function change and thermal desorption. Surf Sci. 1979;88:367.
21. Buckley DH. Surface effects in adhesion, friction, wear, and lubrication. New York: Elsevier;
1981 and Dorinson A, Ludema KC. Mechanics and chemistry in lubrication. New York:
Elsevier; 1985.
22. Smentkowski VS, Cheng CC, Yates Jr JT. Langmuir. 1990;6(1):147.
23. Clark DT. In: Briggs D, editor. Handbook of X-ray and ultraviolet photoelectron spectroscopy.
London: Heyden; 1977. p. 212.
24. Wheeler DR, Pepper SV. J Vac Sci Technol. 1982;20:226.
25. Briggs D, editor. Surface analysis of polymers by XPS and static SIMS. Cambridge, UK:
Cambridge University Press; 1998.
26. Castner DG, Ratner BD. Surface characterization of butyl methacrylate polymers by XPS and
static SIMS. Surf Interface Anal. 1990;15(8):479.
27. Ebnesajjad S. Surface and material characterization techniques, page 32. In: Ebnesajjad S,
editors. Handbook of adhesives and surface preparation: technology, applications and manu-
facturing. Elsevier; 2011.
28. Siegbahn K. Electron Spectroscopy An Outlook. J. Electron Spectrosc. Relat. Phenom. 1974;5:3.
29. Yates Jr JT. Catalysis Chem Eng News. Catalysis. 1974;1922.
30. Fischer TE. Catalysis and surfaces. J Vac Sci Technol. 1974;11:252.
31. Castner DG, Ratner BD. Surf Sci. Biomedical surface science: Foundations to frontiers.
2002;500:28.
32. http://en.wikipedia.org/wiki/Biomaterial
33. Kasemo B. Biological Surface Science. Surf Sci. 2002;500:656.
34. Grunze M, Exarhos G. A surface scientists perspective on biointerphases, or Out of the
vacuum, into the liquid. Biointerphases. 2006;1(1):CL1.
1 Introduction 7

35. Briggs D, Seah MP, editors. Practical surface analysis by Auger and X-ray photoelectron spec-
troscopy. Wiley; 1983; also see Briggs D, editors. Practical surface analysis by Auger and
X-ray photoelectron spectroscopy. 2nd ed. Wiley; 1992.
36. Briggs D, Seah MP. Practical surface analysis: volume 2ion and neutral spectroscopy. 2nd
ed. Chichester: Wiley; 1993.
37. Vickerman JC, Gilmore I, editors. Surface analysis the principal techniques. 2nd ed. Chichester:
Wiley; 2009.
38. Riviere JC, editor. Surface analytical techniques, Monographs on the physics and chemistry of
materials. Oxford, UK: Oxford Science Publications; 1990.
39. Walls JM, editor. Methods of surface analysis. Cambridge, UK: Cambridge University Press;
1989.
40. Brewis SM, Briggs D, editors. Industrial adhesion problems. New York: Wiley Interscience;
1985.
41. Casper LA, Powell CJ, editors. Industrial applications of surface analysis, ACS symposium
series, vol. 199. Washington, DC: American Chemical Society; 1982.
42. Vickerman JC, Briggs D, editors. ToF-SIMS surface analysis by mass spectrometry. Chichester:
SurfaceSpectra/IMPublications; 2001.
43. Grams J, editor. New trends and potentialities of ToF-SIMS in surface studies. New York:
Nova Science Publishers; 2007.
44. Ratner BD, editor. Surface characterization of biomaterials, Progress in biomedical engineer-
ing, vol. 6. Amsterdam: Elsevier; 1988.
45. Davies J, editor. Surface analytical techniques for probing biomaterial processes. Boca Raton:
CRC Press; 1996.
46. Surface and interface analysis, Special issue on multivariate analysis. 2009;41(2).
47. Surface and interface analysis, Special issue on multivariate analysis II. 2009;41(8).
Chapter 2
Applications of XPS in Biology
andBiointerface Analysis

Sally L. McArthur, Gautam Mishra, and Christopher D. Easton

Abstract XPS has been used extensively to characterize the surface chemistry of
materials used in bioengineering and is increasingly finding a role in biology. Its
ability to characterize both the elemental and chemical structures of the surface
makes it particularly useful, as it can be used to identify and image the chemical
functional groups present on the surface of virtually any material. This review is
intended both to profile traditional applications of XPS in bioengineering and biol-
ogy as well as to discuss advances in XPS instrumentation aimed at enabling the
characterization of biological and organic materials.

2.1 Introduction

The past 15 years have seen a rapid rise in techniques capable of probing both biol-
ogy and its interface with materials. Techniques such as surface plasmon resonance
(SPR), optical waveguide lightmode spectroscopy (OWLS), and quartz crystal
microbalance (QCM) are all used to give insight into the kinetics of proteinprotein
interactions and the interactions that occur between biomolecules and materials.
Atomic force microscopy (AFM) can be used to image single molecules,
unfold proteins, monitor surface topography, and measure the forces that hold
biological structures together. While all of these techniques give invaluable

S.L. McArthur (*)

Biointerface Engineering Group, IRIS, Faculty of Engineering and Industrial Sciences,
Swinburne University of Technology, Hawthorn, VIC 3122, Australia
e-mail: [email protected]
G. Mishra
Kratos Analytical, Wharfside, Trafford Wharf Road, Manchester, M17 1GP, UK
C.D. Easton
CSIRO Molecular and Health Technologies, Bag 10, Clayton South, VIC 3169, Australia

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 9

DOI 10.1007/978-3-319-01360-2_2, Springer International Publishing Switzerland 2014
10 S.L. McArthur et al.

insight into specific biological processes, none of them provides chemically

specific information. To fill this gap in our knowledge, surface chemical analysis
techniques more commonly associated with pure materials science have been
applied to study both biological events and the interactions that occur at the inter-
face between biology and engineering.
X-ray photoelectron spectroscopy (XPS), also called electron spectroscopy for
chemical analysis (ESCA), is the most widely used ultrahigh-vacuum (UHV) sur-
face analysis technique. In an XPS experiment, the sample is placed in a UHV
chamber and irradiated with X-rays of a specific wavelength. The adsorption of the
X-rays by atoms in the sample leads to the ejection of core and valence electrons
(photoelectrons). These photoelectrons have energies that are unique to each ele-
ment and sensitive to their chemical states. Significantly, the intensities of the pho-
toelectrons are proportional to the concentration of the element from which they are
ejected. Of course, X-rays are capable of penetrating the sample surface up to many
micrometers, but a small fraction of the photoelectrons generated relatively close to
the surface (~10nm) have sufficient energy to escape into the vacuum system with-
out being scattered. These are the photoelectrons that are detected in XPS.
There are a number of key characteristics that make XPS suitable for the analysis
of biological materials and biointerface analysis, which we discuss next.

2.1.1 Range of Elements Analyzed

XPS detects all elements except H and He. In general, for the detection of proteins,
the nitrogen content of the protein is utilized to determine the presence and quantify
the amount of protein on a surface [1]. The presence of nitrogen in the substrate can
complicate quantification; however, monitoring other elements that are present in
the protein and not in the substrate (such as Fe, Zn, or S) can overcome this compli-
cation. In addition, if the substrate contains an element not found in the protein,
signal attenuation may be used to quantify the amount of adsorbed or immobilized
protein. Similar approaches can be utilized to detect other biomolecules, including
DNA [2], lipids, and mucins [3]. Critically, XPS can be used in the detection of
adventitious contaminants such as silicones, hydrocarbons, and other chemical spe-
cies that may affect the biological interactions or the function of a biomaterial or
medical device [4].
In addition to detecting elements, XPS can be used to characterize specific func-
tional groups associated with a specific element. For carbon, CHx, CO, C=O, and
OC=O may all be differentiated due to variations in the functional groups electro-
negativity, shifting the relative energy of their ejected photoelectron. In some
instances, the binding energy differences between some functional groups are too
small to be resolved clearly by XPS; however, chemical derivatization can be used
to overcome this problem. In this case, the functional group is tagged with a unique
element that has a high photoionization cross section and is stable in the analysis
conditions, such as F or Br [5]. Common compounds utilized include trifluoroacetic
2 Applications of XPS in Biology andBiointerface Analysis 11

anhydride for hydroxyl tagging [6], trifluroethanol for carboxyl groups [7], and
pentafluorobenzaldehyde for amine detection [8]. Independent of complications
associated with small shifts in binding energy, derivatization also enables the iden-
tification and quantification of the functional group present at the interface (e.g.,
acids or amines) and can be used to assess relative activities in specific applications
or reaction conditions [7, 9].

2.1.2 Surface Sensitivity

XPS is a surface-sensitive technique because it monitors unscattered photoelec-

trons, approximately 95% of which arise from within a distance of three times the
inelastic mean free path (3) of the element being detected. For carbon, with
=3.3nm, this means that the XPS sampling depth is ~10nm when the sample
surface is positioned normal to the detector. It is the sampling depth for carbon that
is commonly quoted as the XPS sampling depth. This means that the sampling
depth is larger than the dimensions of many adsorbed proteins and biomolecules.
This allows for signals from a substrate and an adsorbed (i.e., protein) overlayer to
be detected simultaneously.
Attenuation of a specific element from the substrate, or the introduction of a
specific element by the adsorption of a protein, can be utilized to calculate the
adsorbed film thickness or determine the intercalation of protein into a porous sub-
strate. Algorithms utilizing the X-ray emission angle (), theoretical composition of
the protein film or substrate (I), and inelastic mean free path of the emitted photo-
electron of a specific element () enable the calculation of the protein film thickness
(d) using Eq.2.1.
Equation 2.1, XPS overlayer algorithm.

d
I = I exp l cos q
(2.1)

A number of other algorithms exist, but most assume that the protein film is
homogeneous and continuous. If the XPS data do not fit the form of Eq.2.1, this
indicates that the protein film is incomplete or patchy. Paynter and Ratner have
shown that it is possible to incorporate a fractional coverage term into Eq.2.1, but
the quantity of protein adsorbed to the surface must be established from another
technique, such as radiolabeling [10].

2.1.3 Angle-Dependent XPS (ADXPS) and Depth Profiling

As stated earlier, the sampling depth of XPS is dependent on the inelastic mean free
path of the emitted photoelectron of the specific element () and the X-ray emission
angle () via Eq.2.2.
12 S.L. McArthur et al.

Equation 2.2, XPS sampling depth.

d = 3l cos q (2.2)

Thus, by varying the angle between the X-ray source and the sample by simply
tilting the sample stage, one can vary the XPS analysis depth from ~2 to 10nm,
depending on the element being analyzed. ADXPS enables the detection of compo-
sitional variations as a function of depth from the sample surface in a nondestructive
fashion and allows the continuity and depth distribution of the coating to be probed.
Traditionally, ion beam depth profiling of materials in XPS has been limited to
inorganic materials due to ion beaminduced sample damage. In recent years, there
have been significant developments in the ion sources for depth profiling of organic
materials. Driven by the ToF-SIMS community, cluster ion sources (including C60+
[11] and coronene [12]) have become available for integration into XPS instruments.
While the use of these sources is still in its infancy for biology, they have been suc-
cessfully implemented to explore drug distributions in pharmaceuticals [12] and
depth profiling of organic materials [13].

2.1.4 Freeze Hydration XPS

As many of the materials used in bioengineering and all biomolecules operate in a

hydrated environment, questions about the relevance of XPS data to bioengineering
applications and biological interactions are often raised [14, 15]. Freeze hydration
XPS, developed by Lewis and Ratner in the early 1990s [16], involves the rapid
freezing of a wet sample within the XPS entry chamber and can be used to both
circumvent and investigate issues associated with sample dehydration. Once a sam-
ple is frozen, subsequent exposure to UHV at a temperature of approximately
100C etches the ice from the surface via sublimation. The sample temperature is
then lowered below 120C for XPS analysis. Both the study by Lewis[16] and a
number of subsequent investigations[17, 18] have shown that freeze hydration
XPS enables the hydrated surface chemistry of polymers to be probed and the pro-
cesses of polymer reorganization at interfaces investigated.
There were a number of XPS protein adsorption studies using cryogenic sample
preparation that predate the work by Lewis and Ratner and investigated the role of
substrate chemistry in the distribution and orientation of protein films [19, 20]. A
review of the more recent literature illustrates that there has been very little work on
biological or bioengineering applications published since 1993. The work that has
been published has focused on the analysis of bacterial cells and cell membranes
[21]. Some of this may be due to the complex nature of the sample preparation and
the extended amount of instrument time required in preparing for and performing
the experiments. In addition, as noted by Lewis, surface contamination is a major
issue as cryogenic temperatures increase the rate of condensation of contaminants
onto the sample surface in both air and vacuum [16]. As cryogenic sample prepara-
tion becomes more commonplace for the ToF-SIMS analysis of biological samples,
there may be renewed interest in cryogenic XPS.
2 Applications of XPS in Biology andBiointerface Analysis 13

2.1.5 XPS Imaging and Mapping

The increasing interest in spatial control of cells and biomolecules for applications
such as tissue engineering and array technologies has seen the requirement for small
spot analysis and chemical state imaging move to the forefront of XPS research.
Recent advances in instrumentation have seen the spatial resolution of XPS imaging
improve significantly. Today, imaging can be achieved with spatial resolutions
below 10m as the standard; in some cases, a near-micron resolution can be
achieved [22]. A number of groups have used standard spot size analysis
(300700m) to characterize both chemical and thickness gradients over relatively
large areas (~10mm) [7, 23], producing two- and three-dimensional maps of the
material. Unlike ToF-SIMS, quantitative XPS imaging is also possible. While the
process is not trivial, there are an increasing number of papers being published that
detail methods for both producing and processing quantifiable images [2426].
This chapter aims to explore the increasing number of areas where XPS is applied
to address issues in biology and its engineering cousin, bioengineering. After an
introduction to the instrumentation involved, we will focus on a number of the spe-
cific applications where XPS is in use today to investigate biological processes and
the interface between materials and biology.

2.2 Instrumentation

XPS instrumentation has advanced significantly in the last three to four decades,
largely driven by the demand of quantitative surface analysis in the fields of cataly-
sis [27, 28], microelectronics [2931], corrosion science [32, 33], and biomaterials
[3436]. In most common cases, the samples studied by this surface analysis method
are durable and robust. When interest developed in applying XPS to organic and
biological systems, special instrument designs and experimental considerations
were needed. In particular, concerns about the effect of the ultrahigh-vacuum envi-
ronment on traditionally hydrated biological surfaces needed to be addressed.
Alongside this, issues including X-ray and electron damage, surface charge accu-
mulation, narrow X-ray line widths required for higher energy resolution and finer
chemical discrimination, high sample throughput, and better signal-to-noise ratios
needed to be addressed. Advances in instrument manufacturing and the availability
of cutting-edge electronics led to the development of instruments that are precisely
tuned to address many of these specific issues. Nowadays, monochromatic X-ray
probes are standard in most commercial laboratory instruments, with a probe diam-
eter down to less than 10m. The energy resolution and signal-to-noise ratio of the
analyzer system have improved with the introduction of position-sensitive multi-
channel detectors. Imaging analyzers and sophisticated lens systems, such as a mag-
netic immersion lens, have contributed significantly to the improvement in
small-area imaging XPS performance and, most importantly, surface charge
14 S.L. McArthur et al.

neutralization. A 13-m chemical state image resolution can now be routinely

achieved on a laboratory XPS instrument, thus enabling researchers to examine
small features often studied in the fields of biology and medicine.

2.2.1 Overview of XPS Instrumentation

Most commercial XPS systems are divided into two main chambers, one in which
the XPS analysis is carried out (sample analysis chamber, or SAC). This is directly
linked to a second chamber (sample entry chamber, or SEC), a sample introduction
chamber consisting of anything from a simple introduction vacuum lock or a more
complicated vacuum lock with other specialized preparation or treatment equip-
ment attached to it. Stepper motor-controlled automatic or manual mechanisms are
available to transfer the sample from one chamber to another.
As shown in Fig.2.1, the sample introduction chamber is usually pumped by
turbomolecular pumps capable of attaining vacuum levels of ~5
108Pa
(or~51010mbar). Roughing and backing of the turbomolecular pumps are pro-
vided by well-trapped rotary pumps, although novel dry backing and roughing
pumps such as diaphragm and scroll pumps are now becoming more common, par-
ticularly where hydrocarbon contamination must be avoided. The sample is intro-
duced directly into the load lock by venting this chamber to dry nitrogen. For more
complicated load lock designs, often the sample is introduced, as shown in Fig.2.1,
in a relatively smaller chamber (often referred to as fast-entry load lock), where
specific treatments, such as sample cooling, are performed. Temperature measure-
ments are made by a thermocouple in direct contact with the sample stub in the
introduction chamber. Once a good level of vacuum is achieved, the sample is then
transferred into the main analysis chamber or other attached preparation or treat-
ment chambers. The use of fast-entry lock arrangements allows the main load lock
to remain under ultrahigh vacuum at all times. The fast-entry lock arrangements can
also be fitted with glove boxes for specialized sample treatment. Stainless steel is
generally used for the vacuum lock and preparation chambers, although high-grade
aluminum alloys are being used more, especially for locks. The sample-handling
arrangement in the analysis chamber is also equipped with heating and cooling
arrangements. The temperature is monitored by a thermocouple in direct contact
with the sample stubs, and a feedback loop is used to program rates of heating and
cooling. The temperature range for heating and cooling in the analysis chamber is
typically in the range of +600to 150C.
The samples are often transferred between the introduction and analysis chamber
by using an automated or manual sample insertion probe. Various vacuum inter-
locks operate between the connecting chambers in a fail-safe mode to prevent loss
of UHV conditions.
The SAC is generally constructed of mu-metal for effective magnetic screening
and is designed with numerous vacuum ports that have a line of sight to the
2 Applications of XPS in Biology andBiointerface Analysis 15

Fig. 2.1 An example of a commercial XPS instrumentthe Kratos Axis Ultra DLD. This version
of the instrument has been fitted with a radial distribution chamber (RDC) permitting the transfer
of a sample to and from other equipment under vacuum. The computerized user interface is not
shown in this figure (Reproduced with permission from Kratos Analytical Ltd.)

sample in order to accept a comprehensive set of accessories, including ion guns.

The analysis chamber is generally pumped by ion and titanium sublimation pumps,
and the pressure is maintained below ~1108Pa during analysis.

2.2.2 Charge Neutralization

As discussed earlier, the basic XPS experiment involves bombardment of a material

in vacuum with soft X-rays that are capable of penetrating the sample surface up to
many micrometers. Absorption of X-ray energy by an atom in a solid leads to ejec-
tion of an electron, in a process termed photoionization, from either from the core
level or valance bands. A small fraction of these photoelectrons generated relatively
close to the surface (depth~10nm) have sufficient energy to escape into the vacuum
system (i.e., photoemission); the process is termed the photoelectric effect. The
photoelectric emission is the energy analyzed to produce a signature spectrum of
electron intensity as a function of energy.
16 S.L. McArthur et al.

For a conducting sample, conservation of energy leads to the following equation:

Ek = hn EBF j sp

where Ek is the measured kinetic energy of the emitted photoelectron, h is the
energy of the exciting X-ray photon, EBF is the electron binding energy relative to
the Fermi level (EF) of the sample, and sp is the work function of the spectrometer.
Since the binding energy is generally of interest to the user, the spectrometer is set
up to record the spectrum on this energy scale directly.
For insulating samples that do not a have well-defined Fermi level, and to deal
with uncertainty in surface potential developed by emission of the photoelectrons,
we modify the above equation as

E k = hn E B j

where is now the term that captures these surface potential uncertainties and the
actual reference point for EB. If the binding energies for different insulating samples
are to be meaningfully compared, a common reference point needs to be estab-
lished. For polymers, biological samples, and organic samples, the hydrocarbon
component (CC/CH) of the C 1s peak is typically set to 285.0eV and used as an
internal reference.
Non-monochromatic XPS instruments do not generally require a dedicated
charge compensation system. The X-ray flux in these systems is relatively uniform,
and there are usually enough secondary electrons in the vicinity of the sample
induced by the X-ray beams striking the X-ray window and the spectrometer sur-
face to produce adequate charge compensation.
For monochromatic XPS systems, two major advances in charge compensation
systems are commonly available in commercial instruments. The first operates with
spectrometers using a magnetic immersion lens often referred to as a snorkel lens
[37]. The magnetic field intersecting the sample traps photoemitted electrons, caus-
ing them to spiral about the field lines. The charge balance plate causes lower-
energy electrons to be reflected back to the sample, greatly reducing the number of
additional electrons needed to reach equilibrium. As all of the reflected electrons
come from the sample, the surface never becomes overcharged. The design of the
neutralizer enables excess electrons to become trapped in the magnetic field until
required for the neutralization process. This creates a sufficiently high flux of elec-
trons and ensures that the whole analyzed area has a uniform surface charge, provid-
ing optimal XPS performance. A second approach of charge compensation relies on
the use of a low-energy ion beam for neutralization. A stream of low-energy ions
discharges the peripheral regions of the sample, and so electrons from the flood gun
are not reflected before they can reach the illuminated area. In absence of the ion
flux, the sample surface approaches the potential of the most energetic electron
striking it, thus producing an unstable, repulsive potential. This process ensures that
the potential of this peripheral region, which can be charged several volts negative,
is kept close to the potential of the illuminated region so more flooding electrons can
2 Applications of XPS in Biology andBiointerface Analysis 17

reach the illuminated region. When the ion beam energy has been kept below 50eV,
little or no sample damage or implantation has been observed. It has also been
shown that by using a low-work-function metal oxide cathode in the flood gun, a
narrow energy distribution of flooding electrons can be used, making the potential
across the illuminated region more uniform [38].
Biological samples can often have surface topography, and this can be a major
issue for charge compensation. In the case where collimated low-energy ion beams
are used, one would only expect partial neutralization of the sample surface, because
of surface roughness and shadowing effects. This can be overcome using a magnetic
immersion lens system, as the spiral trajectory of the neutralizing electrons allows
them to evenly reach the samples surface irrespective of topography.

2.2.3 Depth Profiling

As discussed earlier, the sampling depth achieved by XPS is approximately 10nm.

However, it is common to have samples composed of a much larger (<10-nm)
compositional gradient. Furthermore, the chemical information generated from the
top ~10nm is a convolution of information from all the layers contained within this
region. The chemical distribution as a function of depth from the outermost ~10nm
of the surface can be converted into depth profiles by using data acquired in an
angular-dependent XPS (ADXPS) experiment. A number of different algorithms
exist, but many, like that shown in Eq.2.2, assume that the overlayer is homoge-
neous and continuous. More realistic models that do not rely on oversimplifying
assumptions exist (e.g., Tougaard [39, 40]) and allow noncontinuous coatings and
nanostructured surfaces to be characterized. Compositional depth profiles can also
be established using these methods. By combining ARXPS with 125I radiolabeling,
the homogeneity and distribution of protein on the surface can also be assessed. If
the surface is porous, the distribution of protein within the outer 10nm of the mate-
rial may also be monitored. Paynter produced an algorithm for biological systems in
the early 1980s [10], and software is now readily available for producing depth
profiles from ARXPS data (e.g., The National Physics Laboratory (NPL) ARCtick
freeware available from http://www.npl.co.uk/nanoanalysis/arctick.html).
To investigate chemical distributions at greater depths (i.e., more that 20nm
from the top surface), destructive depth profile experiments (i.e., ion etching) are
performed. Conventionally in this approach, monoatomic ions such as Ar+ or Cs+
are used to etch a few nm (210nm) of the sample surface, and the bottom of the
etching crater is analyzed at regular intervals. Although this depth profile approach
was found to be very useful for inorganic systems, application in the field of organic,
polymer, and biological materials is limited, as the structural information of the
organic systems is very susceptible to damage from the monoatomic ion beams. The
accuracy of analysis is further compromised by intermixing and knock-in of atoms
at the bottom of the crater. Recently, it has been shown that minimally destructive
XPS depth profiles of organic materials can be obtained by sputtering with large
cluster ions beams (e.g., C60 [11], coronene [12]).
18 S.L. McArthur et al.

Fig. 2.2 The C 1s spectra of spun-cast PLGA during a depth profile experiment using the coro-
nene gun developed by Kratos

Coronene (C24H12) is a polycyclic aromatic hydrocarbon (PAH) consisting of

six carbon rings. Within the ion gun, coronene powder is loaded into the oven
assembly and heated to sublime coronene molecules into the gas phase. Coronene
vapor passes into the source region and is ionized by electron-impact ionization.
Coronene ions are formed into a beam by the condenser lens and projected into a
high-resolution Wien filter to mass-/energy-select a specific ion species. The ion
gun is designed to operate with ion acceleration voltages up to 20keV, but the volt-
age may be set to choose either singly or doubly charged ions. If doubly charged
ions are selected, then the effective ion energy available can be increased to 40keV.
Results show that organic systems can be successfully depth-profiled using the
coronene cluster ion beam. Success is defined in terms of two requirements: (1) a
depth profile with a constant etch rate through the sputtered layer; and (2) a mini-
mum amount of chemical damage to the sample during the sputter depth profile.
Model polymers [e.g., PLGA, poly(lactic-co-glycolic acid)] spun-cast onto silicon
substrate (thickness 85nm) were used to better understand the performance of this
cluster ion gun on organic systems. C 1s spectra recorded at regular intervals during
coronene depth profile are shown in Fig.2.2. Prior to exposure to the coronene
source, three distinct chemical environments were detected in the C 1s spectrum, a
hydrocarbon peak at 285.0eV, a carbon singly bonded to oxygen peak at 286.8eV
and a C peak for the ester group at 289.2eV. During the coronene profile (16-keV
beam energy), the chemical structure of the polymer was found to be conserved
throughout the film to the sample-substrate interface.
In another study, the coronene cluster ion source was used to study the distribu-
tion of a model drug codeine (C18H21NO3) in poly(L-lactic) acid matrix as a model
for a drug-loaded polymer coating [12]. The controlled release of such active
2 Applications of XPS in Biology andBiointerface Analysis 19

90
80 C 1s
70
Atomic concentration (%)

Si 2p
60
50 Zone of codeine depletion Calculated normal
distribution of N 1s
40 O 1s contribution based
30 on uniform drug
distribution x 10
20 N 1s x 10

10
0
0 20 40 60 80 100 120
Depth (nm)

Fig. 2.3 XPS depth profile from coronene-etched drug-loaded PLA film showing C 1s, O 1s,N
1s10 and Si 2p signal and calculated nitrogen concentration assuming a uniform drug distribu-
tion through the film thickness. The N 1s signal was used as a unique indicator for the presence of
drug in the sample (Reprinted with permission from Ref. [12]. Copyright Elsevier Ltd. 2009)

pharmaceutical ingredients from polymers over prolonged periods of time is vital

for their application in drug-eluting stents and other drug-loaded delivery devices.
Figure2.3 shows the coronene ion depth-profile data from the drugpolymer binary
system where the N 1s signal was used to monitor the distribution of the drug as a
function of depth. XPS analysis of the bottom of the sputter crater with sputter time
indicated that codeine was depleted from the surface and segregated to the bulk of
the polymer films by comparison with a uniform distribution calculated from the
bulk loading. This serves to illustrate that surface depletion of drug occurs, which
poses important implications for drug-loaded polymer delivery.

2.2.4 Small Spot Spectroscopy

It is often necessary to analyze a small feature of interest on the surface of a speci-

men. For the analysis to be effective and accurate, the signal from the surrounding
area should be eliminated. Two main approaches are used in current research-grade
instruments to achieve this.
The first approach relies on flooding the sample with X-rays but limiting the
area from which the photoelectrons are collected, generally by using the transfer
lens. In most spectrometers, the electron optic lenses are fitted to the analyzer and
operated in a way to produce a photoelectron image at some point in the electron
optical column. By placing a small aperture at this point, only the electrons emitted
from a defined area are allowed to pass through the aperture and reach the analyzer.
20 S.L. McArthur et al.

Usually in these instruments, a choice of aperture is available for selection, thus

enabling analysis from a wide range of areas. Another variant of this instrument
uses an iris to provide a continuous range of analysis areas. Quite often fixed aper-
tures or an additional iris is positioned in the lens column to correct for spherical
aberrations in the electron optical system by limiting the acceptance angle of the
lens. Using this technique, commercial instruments can provide small-area analysis
down to about 10m.
In the second approach, a monochromatic beam of the X-ray is focused into a
smaller area to analyze the feature of interest on the sample surface. Here, a quartz
crystal is bent so that it can focus a beam of X-rays and provide monochromatic
X-rays by diffraction. In this respect, it behaves rather like a concave mirror. The
focusing is usually achieved using a magnification of unity, which means that the size
of the X-ray spot on the specimen is approximately equal to the size of the electron
spot on the X-ray anode. Analysis areas down to about 10m can be achieved in
commercially available instruments using this method.

2.2.5 XPS Imaging and Mapping

Another area where XPS instrumentation has improved significantly is photoelec-

tron imaging. Manufacturers use two distinct approaches to obtain XPS maps: serial
acquisition, in which each pixel of the image is collected in turn (mapping mode);
and parallel acquisition, where data from the entire analysis area is collected (direct
or real-time imaging).
Serial acquisition of images is based on a 2D rectangular array of small-area
XPS analysis. In this approach, the micro-focused beam of X-rays is scanned over
the sample and an image is built one pixel at a time in a point-by-point acquisition.
The ultimate spatial resolution of the image is thus determined by the size of the
smallest analysis area. Typically, the spatial resolution obtained by this method of
image acquisition is limited to the focus of the X-ray beam, such as ~10m based
on 8020% (or 8416%) edge resolution measurements. Serial acquisition is gen-
erally slower than the parallel acquisition but has the advantage that one can collect
a range of energies.
In parallel acquisition of photoelectron images, the entire field of view is imaged
simultaneously without scanning voltages being applied to any component of the
spectrometer (imaging mode). Obtaining images in this way requires additional
lenses in the spectrometer and the use of a 2D detector in the image plane. The
image resolution has been significantly improved by limiting the angular acceptance
of the lens, thus reducing spherical aberrations. The use of a magnetic immersion
lens in the specimen region also reduces aberrations and therefore allows higher
sensitivity at a given resolution. This method of imaging is relatively fast; commer-
cial instruments can produce images with an image resolution as low as 3m.
Parallel imaging clearly provides better image resolution and is faster than the
serial methods, but it collects an image at a single energy only. Obtaining more
2 Applications of XPS in Biology andBiointerface Analysis 21

accurate image measurements using the parallel image acquisition mode thus
requires a second measurement at some energy remote from the peak, where the
signal intensity is approximately equal to the estimated background signal under the
peak maximum. By mathematically subtracting the background signal from the sig-
nal at the peaks maximum, we can routinely make more accurate measurements.

2.3 Characterization of Biomolecules

2.3.1 Proteins and Peptides

A significant amount of work exists on the use of XPS to study proteins. Due to
similarities in the chemical composition of most proteins, XPS cannot be used to
differentiate individual components within a complex mixture of proteins. Coupled
with the fact that analyzing these biomolecules requires them to be immobilized
onto a surface, which can contribute to the analyzed spectra, XPS is not ideal for a
fundamental chemical analysis of proteins and peptides. The strength of XPS is
realized when applied to examine the immobilization of these biomolecules onto
surfaces. Fundamental knowledge of the adsorption process of proteins to surfaces
at the molecular level is vital for the development of interfaces that exhibit specific
biophysical properties [41]. XPS provides the opportunity to examine adsorbed pro-
teins and elucidate information regarding the orientation, surface coverage, and
layer thickness [1, 10, 42].
Proteins are composed of mostly carbon, oxygen, and nitrogen and can contain
low levels of other elements, including sulfur, phosphorus, and metals. Detection
using XPS is typically achieved by monitoring changes in the nitrogen signal and
attenuating the substrate signal. The highest sensitivity to protein adsorption can be
achieved on nitrogen-free substrates or in situations where the substrate is rich in an
element not found in the protein, such as fluorine. Biochemical assays, such as
radiolabeling using 125I or detection using a fluorescent or enzyme-labeled antibody,
are often used in conjunction with XPS to provide a complete picture of the absorp-
tion characteristics of the system under examination [1, 43, 44]. Figure2.4 demon-
strates the correlation of 125I data with XPS results, for the adsorption of fibrinogen
onto mica [1]. These results compare the XPS detection limit for proteins as calcu-
lated by using the At% N signal (open circles) or attenuation of a unique substrate
signal (Al) shown here as the N(1s)/Al(2s) ratio (x). In both cases, the detection
limit for the protein is shown to be ~10ng/cm2. Much of the early work and related
theory about the detection of proteins has been summarized by Paynter and Ratner
[45] and updated in papers by Castner and Ratner [42] and Wagner etal. [1].
Even when the substrate contains nitrogen, it is still possible to detect protein
although the limits of detection are reduced. Figure2.5 shows the high-resolution C
1s spectrum of a nitrogen-containing plasma-polymerized heptylamine thin film
(HApp) before (solid line) and after the adsorption of IgG (dashed line). While the
22 S.L. McArthur et al.

Fig. 2.4 Surface nitrogen detected by XPS analysis of fibrinogen adsorbed onto mica. N/Al
atomic ratios are also given for mica. Asterisks (*) indicate samples that are significantly different
from the control samples (p0.01) (Reprinted with permission from Ref. [1]. Copyright
Koninklijke Brill NV 2002)

HApp contains nitrogen, it is in the form of amines and imines, whereas the protein
contains amides as well as amines. XPS is able to differentiate between the amine
and amide groups present on the surface, as the photoelectrons produced by the
carbons found in these species have distinct energy shifts. The CN related species
are shifted to 286.5eV, while the electronegativity of the amide (OCN) groups
shifts the binding energy of the carbon associated with these groups to a higher
binding energy of ~288eV. Changes in the amine and amide content of the surface
have been utilized to detect proteins on nitrogen-containing surfaces with a sensitiv-
ity of ~100ng/cm2 [1].
Peptides provide a simplified model system for examining protein adsorption
processes to various surfaces. In addition, adsorbed peptides provide an opportunity
to create a biologically specific interface for tissue engineering [46, 47] and bio-
sensing [48]. As a result, there has been a significant increase in the research on
peptide adsorption in recent years to surfaces such as gold, self-assembled mono-
layers, TiO2, solgel silica, and a range of polymers [4653]. The analysis of pep-
tides on surfaces creates its own challenges due to the size of the molecules and the
often low density of their surface coverage. These factors combine to create what
are often small changes in the chemical composition of a surface [46]. To some
extent, ARXPS can be used to reduce the analysis depth and increase the relative
2 Applications of XPS in Biology andBiointerface Analysis 23

Fig. 2.5 High-resolution XPS C 1s spectra recorded on a plasma-polymerized heptylamine

(HApp)-coated fluorinated ethylene propylene sample before (solid line) and after (dashed line)
immersion in an IgG solution (Reprinted with permission from Ref. [43]. Copyright Elsevier
Ltd. 2000)

proportion of the surface chemistry that is due to the peptide, enabling the peptide
to be differentiated from the substrate, which aids in limiting their impact on the
surface chemistry [54]. An alternative approach is to use a process similar to
derivatization to label the peptide with a specific chemical marker either prior to
immobilization or once it is on the surface [48].
While the examples to date have been from model or laboratory systems, the
same approaches have been used to implement XPS in the characterization of bio-
materials and devices after in vivo experimentation [3, 5558]. Table2.1 presents
the XPS atomic concentration data for Etafilcon A (Acuvue) contact lenses before
and after 10min and 1h of patient wear. These results show that a significant
increase in nitrogen occurs after 10min of wear, while after 1h, a further increase
in nitrogen in addition to oxygen was detected. An analysis of the high-resolution C
1s spectra from the lenses shown in Fig.2.6 clearly demonstrates at 10min the
introduction of peak shifts associated with CO/CN species (~286.5eV) and
OCN species at 288eV. After 1h, the contributions have altered slightly again,
with an increase in the OC=O contributions at 289eV, in line with the increased
level of oxygen detected in the survey spectrum. From this data, researchers deduced
that the adsorption processes were initially dominated by proteins (e.g., after 10min
of wear), but as the wear time increased to an hour, oxygen-rich species (e.g.,
mucins or polysaccharides) adsorbed alongside the proteins [3].
24 S.L. McArthur et al.

Table 2.1 XPS atomic concentrations of Etafilcon A (Acuvue) contact lenses before and after
wear for 10min and 1h. Numbers in parentheses are standard deviations (Reprinted with
permission from Ref. [3]. Copyright Elsevier Ltd. 2001)
Atomic concentration (%) Atomic ratios
C O N Si N:C O:C
Unworn (n=4) 72.5 (1.2) 27.0 (1.6) 0.4 (0.2) 0.1 (0.1) 0.01 (0.00) 0.37 (0.03)
10min (n=10) 72.6 (2.3) 24.5 (2.7)a 2.3 (0.8)a 0.1 (0.0) 0.03 (0.01) 0.34 (0.04)
1h (n=8) 67.9 (1.0)a 28.2 (1.7)a 3.7 (1.6)a 0.2 (0.3) 0.05 (0.02) 0.42 (0.03)
a
Indicates an atomic concentration statistically different from the unworn lens (Student t-test,
p<0.01)

Fig. 2.6High-resolution
XPS C 1s spectra from an
Etafilcon A (Acuvue)
contact lens before (dotted
Relative Intensity (%)

line) and after 10min (solid

line) and 1h of wear (thin
line) (0 take-off angle)
(Reprinted with permission
from Ref. [3]. Copyright
Elsevier Ltd. 2001)

292 290 288 286 284 282

Binding Energy (eV)

2.3.2 Other Biomolecules: Lipids, Mucins, Enzymes, and DNA

Proteins and peptides are not the only biomolecules that play a significant role in
various biological processes. As a result, a number of studies have utilized XPS to
detect and analyze the behavior of a range of biomolecules, including lipids [59
63], mucins and sugars [64, 65], enzymes [6668], and DNA [2, 6973]. Examining
the chemical functionality and adsorption of these biomolecules to different sur-
faces is critical for advancing our fundamental understanding of their behaviors,
developing bioarrays [2, 6971, 73] and biosensors [66, 74], and controlling medi-
cal biofouling [3, 56, 75].
Table2.2 demonstrates the elemental compositions of typical biomolecules [3].
From this data, it is apparent that adsorption of these biomolecules to a surface
would induce a specific and measurable change in the surface composition of the
substrate. As mentioned previously, a change in the nitrogen-to-carbon ratio (N:C)
is often used as a detection marker for proteins, while a change in the oxygen-to-
carbon ratio (O:C) is expected when lipids (ratio decrease) and mucins and sugars
2 Applications of XPS in Biology andBiointerface Analysis 25

Table 2.2 Elemental compositions of common biomolecules (Reprinted with permission from
Ref. [3]. Copyright Elsevier Ltd. 2001)
Atomic % composition Atomic ratios
C 1s O 1s N 1s N:C O:C
Proteina 63.0 20.1 16.0 0.25 0.32
Proteinb 65.3 18.1 14.2 0.22 0.28
Lipidc 95.0 5.0 0.0 0.00 0.05
Mucind 58.0 31.0 9.8 0.17 0.53
a
Theoretical composition of human albumin
b
Data derived from XPS spectra of thick human albumin film
c
Theoretical composition of cholesterol (C27H48O)
d
Data derived from XPS spectra of the glycosylated region of porcine submaxillary mucin (PSM,
MUC1)

(ratio increase) adsorb to surfaces. The contrast between the adsorbed biomolecules
and the substrate can be further enhanced using angle-resolved XPS (ARXPS),
where the depth of analysis is varied to analyze a larger portion of the adsorbed or
immobilized material.
The introduction of DNA microarray technology provides researchers a tool to
study multiple cellular processes in parallel and thus is significant to fundamental
biological research and biomedical applications [7678]. In papers that have
appeared between 2007 and 2009, Lee [2], Liu [69], and others [73, 78] have exam-
ined how XPS could be used to interrogate DNA arrays and provide a quantitative
interpretation of DNA hybridization efficiencies and base-pair mismatch detection.
Overlays of XPS images of phosphorus, nitrogen, and sodium onto the silicon
image before and after hybridization from Lees study clearly demonstrated an
increase in signal intensity upon hybridization for microspots containing comple-
mentary probe sequences (Fig.2.7) [2]. Lius work was able to detect a single base
mismatch as well as give valuable insight into hybridization efficiencies of the
arrays [69]. These studies, together with others from groups around the world, have
clearly demonstrated the applications of XPS in both the development and analysis
of DNA arrays without the need for radioactive or fluorescent labels [73, 78, 79].
Significantly, many of these studies have also demonstrated the ability of XPS to
detect variations and surface-induced problems within the hybridization processes
of a wide range of commercial and noncommercial arrays [73, 79].

2.4 Cell, Bacteria, and Tissue Analysis

There are many types of cells, each containing a unique combination of components
within the cell membrane, including phospholipids, proteins, glycosaminoglycans
(GAGs), and cholesterol. XPS provides the opportunity to probe the cell surface and
elucidate information regarding their composition and responses to environmental
stimulation. As detailed earlier, all of the individual components of a cell have been
26 S.L. McArthur et al.

Fig. 2.7 XPS overlay of phosphorus (P 2p), nitrogen (N 1s), and sodium (Na 1s) with the substrate
silicons (Si 2p) signal intensity images (800800m) from printed DNA probes on CodeLink
microarray slides (a) before and (b) after target hybridization (Reprinted with permission from
Ref. [2]. Copyright ACS 2007)

investigated on surfaces using XPS. When combined within a cell membrane, it

becomes difficult to distinguish the individual components within the resultant XPS
spectrum and quantitatively determine the amount of each component. As a result,
the total element compositions obtained from survey spectra are typically employed
to provide a semiquantitative comparison of cells [80]. In addition, the peak fitting
2 Applications of XPS in Biology andBiointerface Analysis 27

of high-resolution components (e.g., C 1s, N 1s, and O 1s) is typically undertaken.

Postprocessing techniques can then be used to gain further insight. For example, the
surface composition of bacteria can be modeled by three classes of components,
specifically polysaccharides, peptides, and hydrocarbon-like compounds (lipidic
compounds) [21].
The major focus of XPS analysis of cells has been n the investigation of micro-
bial cells, including yeast, fungi, diatoms, and bacteria [8184]. A number of stud-
ies have used XPS to explore changes in the chemical composition of the cell wall
with changes in media (for suspended cells) or changes in substrate and media (for
adhesive cells) [85, 86]. These types of studies are highly relevant for the minerals
[87] and engineering industries, where specific microbial events are central to the
success or failure of a process. In some instances, it enables industries to understand
and prevent events such as biocorrosion [8890], while in others it enables them to
understand how surface characteristics of the bacteria or their extracellular poly-
meric substance (EPS) change in response to the environment [91, 92]. Properties
that have been investigated include bacteria attachment and adhesion, with the aim
of addressing biofouling [75, 85, 88], bioremediation [87, 93], and environmental
applications [84, 94, 95].
An example of employing XPS to investigate the surface chemical structure of
microbial cells is found in the work of Ahimou etal. [96] The authors focused on
the fitting of high-resolution peaks to elucidate the chemical functions of nine
strains of Bacillus subtilis. Figure2.8 presents high-resolution spectra of oxygen,
nitrogen, and carbon for three B. subtilis strains. It is apparent that the spectral com-
ponents assigned to the carbon peak vary considerably among strains, while the
degree of variation is less significant for the nitrogen and oxygen peaks. Based on
the peak position of the spectra components, the authors were able to determine the
presence of individual components within the cell membrane. For example, the
peak centered at 285eV [carbon bound to carbon and hydrogen; C(C,H)] was
considered to originate from lipids or from side chains of amino acids, while the
peak at ~289eV (carboxylate and carboxyl groups) was the result of proteins and
uronic acids. Further data interpretation of molar concentrations provided addi-
tional evidence; however, the authors did acknowledge that such modeling can lose
significance as the number of unknown variables increases.
Another example of utilizing XPS to analyze B. subtilis was reported by Leone
etal. [80]. In this study, a single strain (ATCC code: 6633) was analyzed using the
standard freeze-dried technique and compared to frozen wet-paste samples sus-
pended at different pH values. Applying this alternative sample-preparation tech-
nique allowed the authors to investigate protonation/deprotonation reactions of
amine groups at the cellwater interface and elucidate information regarding the
activity of the cell at specific pH values. It is interesting to note that different sample-
preparation techniques were explored in this study, as the choice of sample-
preparation and sample-handling techniques for XPS analysis of cells has long been
a subject of some debate [14].
As XPS is a UHV technique, there has always been considerable concern as to
whether the preparation techniques required and resultant dehydration disrupt the
28 S.L. McArthur et al.

Fig. 2.8 Representative O 1s, N 1s, and C 1s XPS peaks of three B. subtilis strains: (a) ATCC
code: 7058; (b) ATCC code: 15476; (c) S 499 (Reprinted with permission from Ref. [96]. Copyright
Elsevier Ltd. 2007)

cell surface [14, 97]. In light of these concerns, a substantial amount of work has
been undertaken to confirm XPS results by drawing comparisons with complemen-
tary data, including infrared absorption, water contact angle, and electrophoretic
mobility, which are techniques that do not require ultrahigh vacuum [85, 86, 98].
These studies have shown a correlation between changes in the cell wall chemical
composition with changes in the media or substrate and media, dependent on
whether the cell was suspended or adhered to a surface prior to analysis. Despite
these results, it is apparent that a significant amount of care must be taken when
preparing microbial cells for XPS analysis. A number of reviews on the field exist
and include critical assessment of the issues associated with sample preparation as
well as details of the cell types that have been analyzed to date [14, 86].

2.4.1 Hard Tissue

Hard tissues include bone, dentin, cementum, and dental enamel; these represent
the group of tissues that have undergone various degrees of mineralization during
development. This mineralization is typically in the form of hydroxyapatite, which
2 Applications of XPS in Biology andBiointerface Analysis 29

is often applied as a biomaterial for bone implants. Within the literature, application
of XPS to the analysis of hard tissue has focused on preventive and restorative den-
tistry. This has included analyzing the effects of various treatments, including laser
irradiation [99, 100] and mouth rinses [101], and the effects of environmental expo-
sure to chemical agents on the health of tooth enamel [102]. XPS has also been
employed to examine bonding of materials and mechanisms of interactions between
hard tissues and biomaterials [57, 103, 104] In 2008, Lou etal. used XPS to inves-
tigate the surface chemical composition of 98 human maxillary first premolars to
determine the compositional differences between right and left premolars [104].
The descriptive statistics of the percentage atomic concentration for all 12 elements
detected were determined. The percentage of carbonate (CO3) was derived from
high-resolution carbon spectra; the ratio CO3/P is considered to be a measure of the
susceptibility of enamel to acid attack. While no statistically significant difference
was observed between right and left premolars, the authors considered the observed
variation in CO3/P potentially significant in terms of bond strength, as acid etching
is typically performed prior to orthodontic bonding.

2.5 Biointerface Engineering

Biointerface engineering aims to bring together biology with surface engineering/

science to unlock the fundamental properties driving biomolecule interactions with
surfaces and to address key issues relating to biomaterials, tissue engineering, medi-
cal diagnostics, and many biotechnology applications, including sensors, bioarrays,
and microfluidics. As XPS is a surface-sensitive technique, it is no surprise that it
has become a core characterization tool within the discipline.
Differences between surface and bulk properties can arise via a number of differ-
ent processes, including surface contamination, processing additives, blooming of
plasticizers, and oxidation. Adventitious contaminants such as silicones and hydro-
carbons can prevent adhesion of components, may influence the biological interac-
tions of a biomaterial or tissue scaffold, and can interfere with the function of
surface-immobilized proteins, peptides, and DNA. As such, it is critical that the
surface chemistry of a material is verified at some point in the testing cycle prior to
ascribing relationships between biological performance and surface chemistry or
evaluating a specific immobilization strategy. There is a dearth of literature related
to the detection of surface contamination via XPS, and a number of reviews exist
that specifically focus on biomedical applications and issues [42, 105].
A key issue facing the integration of biomaterials is the stability of the surface
when exposed to biological media; thus, an examination of how biological systems
impact surfaces is paramount. XPS allows changes in the chemical composition of
interfaces to be examined and has been used previously to examine the influence of
microbiological corrosion (MIC) by marine aerobic bacterium on stainless steel
[89] and the effects of cell culture medium on Ti, stainless steel, and a range of
polymers, to name but a few examples [58, 106, 107].
30 S.L. McArthur et al.

As materials used in biology and bioengineering are often chosen for their
mechanical or optical properties, there is an array of techniques and treatments
designed to modify the materials surface for specific applications [42, 108].
Because the modifications are generally designed not to affect the bulk properties,
the surface sensitivity of XPS is paramount, and angle-dependent XPS (ADXPS) is
also regularly used to produce compositional depth profiles.

2.5.1 The Future of XPS

While XPS is a fundamental method for probing interfacial interactions in bioen-

gineering, research is increasingly focusing on using XPS as part of a suite of
characterization tools [42]. Obvious synergies exist between XPS and ToF-SIMS,
as evidenced by the large number of papers currently in the literature that use
both of these techniques. More fundamental insight into, and improvements in,
devices and technology are progressively coming from combining UHV surface
analysis with techniques commonly used in colloids and surface science (e.g.,
AFM[109]) and biological assays, such as ELISA [110], immunostaining [111],
and polymerase chain reactions (PCR) [94]. This is where XPS can be used for
its strengths in quantifying surface contamination, verifying surface chemistry
[47, 109], and determining changes in surface chemistry after biological contact
[94, 95, 111].
However, this is not to say that there are no opportunities for developments in
XPS. Today multivariate statistical analysis (MVSA) routines are increasingly
being developed today to assist in the interpretation of XPS data, particularly with
results from imaging studies. Multivariate image analysis (MIA) methods such as
scatter diagrams, principal component analysis (PCA), and classification methods
are used to extract maps of pure components from degradation and images-to-
spectra data sets [112]. Walton and Fairley have shown that by maintaining the
relationship between images and spectra, it is possible to progress beyond the appli-
cation of spectroscopic processing to multispectral imaging data sets, by utilizing
the three-dimensional information contained in such data sets, to therefore improve
both the processing and the visualization of the data [113]. With the ongoing devel-
opment of depth profiling of biological materials being made possible by the intro-
duction of the polyatomic ion guns, groups are just beginning to explore the
applications of MVSA to explore biological systems. Studies from Artyushkova
have used principal component analysis (PCA) to analyze quantitative XPS data,
combining elemental and chemical species data as a function of sputter time to
explore the structure of a yeast cell, with the final aim of exploring cell-directed
assembly [113]. Of course, the ongoing close relationship between XPS and ToF-
SIMS development will be of significant benefit as the sample-preparation tech-
niques and cryogenic stages that have been developed for ToF-SIMS can be directly
translated to XPS analysis.
2 Applications of XPS in Biology andBiointerface Analysis 31

2.6 Conclusions

Today XPS has established itself as a workhorse tool for biointerface analysis.
While the key features of surface sensitivity and wide elemental analysis are at the
heart of the technique, improvements in the imaging and depth-profiling capabilities
of the instruments are increasingly drawing in new biological applications. But at
the core, it is the relative ease of XPS data interpretation compared to ToF-SIMS
and spectrometry techniques that will ensure that XPS retains its position well into
the future.

References

1. Wagner MS, McArthur SL, Shen MC, Horbett TA, Castner DG. Limits of detection for time
of flight secondary ion mass spectrometry (ToF-SIMS) and X-ray photoelectron spectros-
copy (XPS): detection of low amounts of adsorbed protein. J Biomater Sci-Polym Ed.
2002;13:40728.
2. Lee CY, Harbers GM, Grainger DW, Gamble LJ, Castner DG. Fluorescence, XPS, and TOF-
SIMS surface chemical state image analysis of DNA microarrays. J Am Chem Soc.
2007;129:942938.
3. McArthur SL, McLean KM, St John HAW, Griesser HJ. XPS and surface-MALDI-MS char-
acterisation of worn HEMA-based contact lenses. Biomaterials. 2001;22:3295304.
4. Foley JO, Fu E, Gamble LJ, Yager P. Microcontact printed antibodies on gold surfaces: func-
tion, uniformity, and silicone contamination. Langmuir. 2008;24:362835.
5. St John HAW, Gengenbach TR, Hartley PG, Griesser HJ. Surface analysis of polymers. In:
OConnor D, Sexton B, Smart RC, editors. Surface analysis methods in material science.
Heidelberg: Springer-Verlag Berlin; 2000.
6. Beamson G, Alexander MR. Angle-resolved XPS of fluorinated and semi-fluorinated side-
chain polymers. Surf Interface Anal. 2004;36:32333.
7. Alexander MR, Whittle JD, Barton D, Short RD. Plasma polymer chemical gradients for
evaluation of surface reactivity: epoxide reaction with carboxylic acid surface groups. J
Mater Chem. 2004;14:40812.
8. Ratner BD, Castner DG. Advances in X-ray photoelectron spectroscopy instrumentation and
methodology: instrument evaluation and new techniques with special reference to biomedical
studies. Colloid Surf B. 1994;2:33346.
9. Whittle JD, Barton D, Alexander MR, Short RD. A method for the deposition of controllable
chemical gradients. Chem Commun. 2003;14:17667.
10. Paynter RW, Ratner BD. The study of interfacial proteins and biomolecules by X-ray photo-
electron spectroscopy. In: Andrade JD, editor. Surface and interfacial aspects of biomedical
polymers. New York: Plenum Press; 1985. p. 189216.
11. Chen Y-Y, Yu B-Y, Wang W-B, Hsu M-F, Lin W-C, Lin Y-C, et al. X-ray photoelectron spec-
trometry depth profiling of organic thin films using C60 sputtering. Anal Chem. 2007;
80:5015.
12. Rafati A, Davies MC, Shard AG, Hutton S, Mishra G, Alexander MR. Quantitative XPS
depth profiling of codeine loaded poly(l-lactic acid) films using a coronene ion sputter source.
J Control Release. 2009;138:404.
13. Yu BY, Chen YY, Wang WB, Hsu MF, Tsai SP, Lin WC, et al. Depth profiling of organic
films with X-ray photoelectron spectroscopy using C-60(+) and Ar+ co-sputtering. Anal
Chem. 2008;80:34125.
32 S.L. McArthur et al.

14. Pembrey RS, Marshall KC, Schneider RP. Cell surface analysis techniques: what do cell
preparation protocols do to cell surface properties? Appl Environ Microbiol. 1999;
65:287794.
15. Vogler EA. On the biomedical relevance of surface spectroscopy. J Electron Spectrosc Relat
Phenom. 1996;81:23747.
16. Lewis KB, Ratner BD. Observation of surface restructuring of polymers using ESCA. J
Colloid Interface Sci. 1993;159:7785.
17. Lukas J, Sodhi RNS, Sefton MV. An XPS study of the surface reorientation of statistical
methacrylate copolymers. J Colloid Interface Sci. 1995;174:4217.
18. Magnani A, Barbucci R, Lewis KB, Leachscampavia D, Ratner BD. Surface-properties and
restructuring of a cross-linked polyurethane-poly(amido-amine) network. J Mater Chem.
1995;5:132130.
19. Paynter RW, Ratner BD, Horbett TA, Thomas HR. XPS studies on the organisation of
adsorbed protein films on fluoropolymers. J Colloid Interface Sci. 1984;101:23345.
20. Ratner BD, Thomas TA, Shuttleworth D, Horbett TA. Analysis of the organisation of protein
films on solid surfaces by ESCA. J Colloid Interface Sci. 1981;83:63042.
21. Dufrene YF, VanderWal A, Norde W, Rouxhet PG. X-ray photoelectron spectroscopy analy-
sis of whole cells and isolated cell walls of gram-positive bacteria: comparison with bio-
chemical analysis. J Bacteriol. 1997;179:10238.
22. Vohrer U, Blomfield C, Page S, Roberts A. Quantitative XPS imaging new possibilities
with the delay-line detector. Appl Surf Sci. 2005;252:615.
23. Wang XJ, Haasch RT, Bohn PW. Anisotropic hydrogel thickness gradient films derivatized to
yield three-dimensional composite materials. Langmuir. 2005;21:84529.
24. Tougaard S. Algorithm for automatic X-ray photoelectron spectroscopy data processing and
X-ray photoelectron spectroscopy imaging. J Vac Sci Technol A. 2005;23:7415.
25. Walton J, Fairley N. Quantitative surface chemical-state microscopy by X-ray photoelectron
spectroscopy. Surf Interface Anal. 2004;36:8991.
26. Walton J, Fairley N. Transmission-function correction for XPS spectrum imaging. Surf
Interface Anal. 2006;38:38891.
27. Steinmiller EMP, Choi K-S. Photochemical deposition of cobalt-based oxygen evolving cata-
lyst on a semiconductor photoanode for solar oxygen production. Proc Natl Acad Sci U S A.
2009;106:206336.
28. Grnert W, Stakheev AY, Mrke W, Feldhaus R, Anders K, Shpiro ES, et al. Reduction and
metathesis activity of MoO3/Al2O3 catalysts: I. An XPS investigation of MoO3/AI2O3 cata-
lysts. J Catal. 1992;135:26986.
29. Finster J, Klinkenberg ED, Heeg J, Braun W. ESCA and SEXAFS investigations of insulating
materials for ULSI microelectronics. Vacuum. 1990;41:15869.
30. Casper L. Microelectronics processing: inorganic materials characterization. Washington
DC: American Chemical Society; 1986.
31. Stingeder G. The challenge of microelectronics for analytical chemistry. Fresenius J Anal
Chem. 1992;343:7712.
32. Heung WF, Yang YP, Wong PC, Mitchell KAR, Foster T. XPS and corrosion studies on zinc
phosphate coated 7075-T6 aluminium alloy. J Mater Sci. 1994;29:136873.
33. Windisch C, Baer D, Engelhard M, Jones R. Analyzing localized corrosion in lon-implanted
metals via XPS/AES. JOM J Miner Met Mater Soc. 2001;53:3741.
34. Chang MC, Tanaka J. XPS study for the microstructure development of hydroxyapatite-col-
lagen nanocomposites cross-linked using glutaraldehyde. Biomaterials. 2002;23:387985.
35. Pan J, Thierry D, Leygraf C. Electrochemical and XPS studies of titanium for biomaterial
applications with respect to the effect of hydrogen peroxide. J Biomed Mater Res.
1994;28:11322.
36. Ratner B. Surface characterization of biomaterials by electron spectroscopy for chemical
analysis. Ann Biomed Eng. 1983;11:31336.
37. Walker AR. Charged particle energy analyser. US Patent: 4810879. 1989.
2 Applications of XPS in Biology andBiointerface Analysis 33

38. Paul EL, Michael AK. Surface charge neutralization of insulating samples in X-ray
photoemission spectroscopy. J Vac Sci Technol A. 1998;16:34839.
39. Tougaard S. Universality classes of inelastic electron scattering cross-sections. Surf Interface
Anal. 1997;25:13754.
40. Hajati S, Tougaard S. XPS for non-destructive depth profiling and 3D imaging of surface
nanostructures. Anal Bioanal Chem. 2010;396:274155.
41. Horbett TA, Brash JL. Proteins at interfaces II: fundamentals and applications. Washington
D. C.: American Chemical Society; 1995.
42. Castner DG, Ratner BD. Biomedical surface science: foundations to frontiers. Surf Sci.
2002;500:2860.
43. McArthur SL, McLean KM, Kingshott P, St John HAW, Chatelier RC, Griesser HJ. Effect of
polysaccharide structure on protein adsorption. Colloid Surf B-Biointerfaces. 2000;
17:3748.
44. Salim M, OSullivan B, McArthur SL, Wright PC. Characterization of fibrinogen adsorption
onto glass microcapillary surfaces by ELISA. Lab Chip. 2007;7:6470.
45. Johnson G, Jenkins ML, McLean KM, Griesser HJ, Kwak J, Goodman M, et al. Peptoid-
containing collagenmimetics with cell binding activity. J Biomed Mater Res.
2000;51:61224.
46. Massia SP, Stark J. Immobilized RGD peptides on surface-grafted dextran promote biospe-
cific cell attachment. J Biomed Mater Res. 2001;56:3909.
47. Chow E, Wong ELS, Bocking T, Nguyen QT, Hibbert DB, Gooding JJ. Analytical perfor-
mance and characterization of MPA-Gly-Gly-His modified sensors. Sens Actuator B-Chem.
2005;111:5408.
48. Weidner T, Samuel NT, McCrea K, Gamble LJ, Ward RS, Castner DG. Assembly and struc-
ture of alpha-helical peptide films on hydrophobic fluorocarbon surfaces. Biointerphases.
2010;5:916.
49. Apte JS, Collier G, Latour RA, Gamble LJ, Castner DG. XPS and ToF-SIMS investigation
of alpha-helical and beta-strand peptide adsorption onto SAMs. Langmuir. 2010;
26:342332.
50. Feyer V, Plekan O, Tsud N, Chab V, Matolin V, Prince KC. Adsorption of histidine and his-
tidine-containing peptides on Au(111). Langmuir. 2010;26:860613.
51. Iucci G, Battocchio C, Dettin M, Gambaretto R, Polzonetti G. A NEXAFS and XPS study of
the adsorption of self-assembling peptides on TiO2: the influence of the side chains. Surf
Interface Anal. 2008;40:2104.
52. Jedlicka SS, Rickus JL, Zemyanov DY. Surface analysis by X-ray photoelectron spectros-
copy of solgel silica modified with covalently bound peptides. J Phys Chem B.
2007;111:118507.
53. Charnley M, Fairfull-Smith K, Haldar S, Elliott R, McArthur SL, Williams NH, et al.
Generation of bioactive materials with rapid self-assembling resorcinarene-peptides. Adv
Mater. 2009;21:290915.
54. Shen MC, Martinson L, Wagner MS, Castner DG, Ratner BD, Horbett TA. PEO-like plasma
polymerized tetraglyme surface interactions with leukocytes and proteins: in vitro and in vivo
studies. J Biomater Sci-Polym Ed. 2002;13:36790.
55. Thissen H, Gengenbach T, du Toit R, Sweeney DF, Kingshott P, Griesser HJ, et al. Clinical
observations of biofouling on PEO coated silicone hydrogel contact lenses. Biomaterials.
2010;31:55109.
56. Zinelis S, Thomas A, Syres K, Silikas N, Eliades G. Surface characterization of zirconia
dental implants. Dent Mater. 2009;26:295305.
57. Leitao E, Barbosa MA, deGroot K. XPS characterization of surface films formed on surface-
modified implant materials after cell culture. J Mater Sci-Mater Med. 1997;8:4236.
58. Liu YT, Li K, Pan J, Liu B, Feng SS. Folic acid conjugated nanoparticles of mixed lipid
monolayer shell and biodegradable polymer core for targeted delivery of Docetaxel.
Biomaterials. 2010;31:3308.
34 S.L. McArthur et al.

59. Daniel C, Sohn KE, Mates TE, Kramer EJ, Radler JO, Sackmann E, et al. Structural
characterization of an elevated lipid bilayer obtained by stepwise functionalization of a self-
assembled alkenyl silane film. Biointerphases. 2007;2:10918.
60. Kim HK, Kim K, Byun Y. Preparation of a chemically anchored phospholipid monolayer on
an acrylated polymer substrate. Biomaterials. 2005;26:343544.
61. Michel R, Subramaniam V, McArthur SL, Bondurant B, DAmbruoso GD, Hall HK, et al.
Ultra-high vacuum surface analysis study of rhodopsin incorporation into supported lipid
bilayers. Langmuir. 2008;24:49016.
62. McArthur SL, Halter MW, Vogel V, Castner DG. Covalent coupling and characterization of
supported lipid layers. Langmuir. 2003;19:831624.
63. Russell BG, Moddeman WE, Birkbeck JC, Wright SE, Millington DS, Stevens RD, et al.
Surface structure of human mucin using X-ray photoelectron spectroscopy. Biospectroscopy.
1998;4:25766.
64. Lundin M, Sandberg T, Caldwell KD, Blomberg E. Comparison of the adsorption kinetics
and surface arrangement of as received and purified bovine submaxillary gland mucin
(BSM) on hydrophilic surfaces. J Colloid Interface Sci. 2009;336:309.
65. Libertino S, Giannazzo F, Aiello V, Scandurra A, Sinatra F, Renis M, et al. XPS and AFM
characterization of the enzyme glucose oxidase immobilized on SiO2 surfaces. Langmuir.
2008;24:196572.
66. Abbas A, Vercaigne-Marko D, Supiot P, Bocquet B, Vivien C, Guillochon D. Covalent
attachment of trypsin on plasma polymerized allylamine. Colloid Surf B-Biointerfaces.
2009;73:31524.
67. Debenedetto GE, Malitesta C, Zambonin CG. Electroanalytical X-ray photoelectron-spec-
troscopy investigation on glucose-oxidase adsorbed on platinum. J Chem Soc-Faraday Trans.
1994;90:14959.
68. Liu ZC, Zhang X, He NY, Lu ZH, Chen ZC. Probing DNA hybridization efficiency and single
base mismatch by X-ray photoelectron spectroscopy. Colloid Surf B-Biointerfaces.
2009;71:23842.
69. Zhang XC, Kumar S, Chen JH, Teplyakov AV. Covalent attachment of shape-restricted DNA
molecules on amine-functionalized Si(111) surface. Surf Sci. 2009;603:244557.
70. May CJ, Canavan HE, Castner DG. Quantitative X-ray photoelectron spectroscopy and time-
of-flight secondary ion mass spectrometry characterization of the components in DNA. Anal
Chem. 2004;76:111422.
71. Petrovykh DY, Kimura-Suda H, Tarlov MJ, Whitman LJ. Quantitative characterization of
DNA films by X-ray photoelectron spectroscopy. Langmuir. 2004;20:42940.
72. Graf N, Gross T, Wirth T, Weigel W, Unger WES. Application of XPS and ToF-SIMS for
surface chemical analysis of DNA microarrays and their substrates. Anal Bioanal Chem.
2009;393:190712.
73. Dhayal M, Ratner DA. XPS and SPR analysis of glycoarray surface density. Langmuir. 2009;
25:21817.
74. Tyler BJ. XPS and SIMS studies of surfaces important in biofilm formation three case stud-
ies. In: Prokop A, Hunkeler D, Cherrington AD, editors. Bioartificial organs science, medi-
cine, and technology. New York: New York Acad Sciences; 1997. p. 11426.
75. Bejjani BA, Shaffer LG. Application of array-based comparative genomic hybridization to
clinical diagnostics. J Mol Diagn. 2006;8:52833.
76. Lamartine J. The benefits of DNA microarrays in fundamental and applied bio-medicine.
Mater Sci Eng C-Biomim Supramol Syst. 2006;26:3549.
77. Wu P, Castner DG, Grainger DW. Diagnostic devices as biomaterials: a review of nucleic acid
and protein microarray surface performance issues. J Biomater Sci-Polym Ed.
2008;19:72553.
78. Gong P, Harbers GM, Grainger DW. Multi-technique comparison of immobilized and hybrid-
ized oligonucleotide surface density on commercial amine-reactive microarray slides. Anal
Chem. 2006;78:234251.
2 Applications of XPS in Biology andBiointerface Analysis 35

79. Leone L, Loring J, Sjooberg S, Persson P, Shchukarev A. Surface characterization of the

Gram-positive bacteria Bacillus subtilis an XPS study. Surf Interface Anal. 2006;38:2025.
80. Rouxhet PG, Mozes N, Dengis PB, Dufrene YF, Gerin PA, Genet MJ. Application of X-ray
photoelectron-spectroscopy to microorganisms. Colloid Surf B-Biointerfaces. 1994;2:34769.
81. Tesson B, Genet MJ, Fernandez V, Degand S, Rouxhet PG, Martin-Jezequel V. Surface
chemical composition of diatoms. ChemBioChem. 2009;10:201124.
82. Dague E, Delcorte A, Latge JP, Dufrene YF. Combined use of atomic force microscopy,
X-ray photoelectron spectroscopy, and secondary ion mass spectrometry for cell surface
analysis. Langmuir. 2008;24:29559.
83. Ojeda JJ, Romero-Gonzalez ME, Bachmann RT, Edyvean RGJ, Banwart SA. Characterization
of the cell surface and cell wall chemistry of drinking water bacteria by combining XPS,
FTIR spectroscopy, modeling, and potentiometric titrations. Langmuir. 2008;24:403240.
84. Boonaert CJP, Rouxhet PG. Surface of lactic acid bacteria: relationships between chemical
composition and physicochemical properties. Appl Environ Microbiol. 2000;66:254854.
85. van der Mei HC, de Vries J, Busscher HJ. X-ray photoelectron spectroscopy for the study of
microbial cell surfaces. Surf Sci Rep. 2000;39:324.
86. Sharma PK, Rao KH. Surface characterization of bacterial cells relevant to the mineral indus-
try. Miner Metall Process. 2005;22:317.
87. Beech IB, Zinkevich V, Tapper R, Gubner R, Avci R. Study of the interaction of sulphate-
reducing bacteria exopolymers with iron using X-ray photoelectron spectroscopy and time-
of-flight secondary ionisation mass spectrometry. J Microbiol Methods. 1999;36:310.
88. Yuan SJ, Pehkonen SO. Microbiologically influenced corrosion of 304 stainless steel by aero-
bic Pseudomonas NCIMB 2021 bacteria: AFM and XPS study. Colloid Surf B-Biointerfaces.
2007;59:8799.
89. Sahrani FK, Aziz MA, Ibrahim Z, Yahya A. Surface analysis of marine sulphate-reducing
bacteria exopolymers on steel during biocorrosion using X-ray photoelectron spectroscopy.
Sains Malays. 2008;37:1315.
90. Omoike A, Chorover J. Spectroscopic study of extracellular polymeric substances from
Bacillus subtilis: aqueous chemistry and adsorption effects. Biomacromolecules.
2004;5:121930.
91. Rodrigues LR, Teixeira JA, van der Mei HC, Oliveira R. Isolation and partial characterization
of a biosurfactant produced by Streptococcus thermophilus A. Colloid Surf B-Biointerfaces.
2006;53:10512.
92. Fantauzzi M, Rossi G, Elsener B, Loi G, Atzei D, Rossi A. An XPS analytical approach for
elucidating the microbially mediated enargite oxidative dissolution. Anal Bioanal Chem.
2009;393:193141.
93. Magnuson TS, Neal AL, Geesey GG. Combining in situ reverse transcriptase polymerase
chain reaction, optical microscopy, and X-ray photoelectron spectroscopy to investigate min-
eral surface-associated microbial activities. Microb Ecol. 2004;48:57888.
94. Kalinowski BE, Liermann LJ, Brantley SL, Barnes A, Pantano CG. X-ray photoelectron
evidence for bacteria-enhanced dissolution of hornblende. Geochim Cosmochim Acta.
2000;64:133143.
95. Ahimou F, Boonaert CJP, Adriaensen Y, Jacques P, Thonart P, Paquot M, et al. XPS analysis
of chemical functions at the surface of Bacillus subtilis. J Colloid Interface Sci.
2007;309:4955.
96. Marshall KC, Pembrey R, Schneider RP. The relevance of X-ray photoelectron-spectroscopy
for analysis of microbial cell-surfaces a critical-view. Colloid Surf B-Biointerfaces.
1994;2:3716.
97. Leone L, Ferri D, Manfredi C, Persson P, Shchukarev A, Sjoberg S, et al. Modeling the acid
base properties of bacterial surfaces: a combined spectroscopic and potentiometric study of
the gram-positive bacterium Bacillus subtilis. Environ Sci Technol. 2007;41:646571.
98. Mine A, Yoshida Y, Suzuki K, Nakayama Y, Yatani H, Kuboki T. Spectroscopic characteriza-
tion of enamel surfaces irradiated with Er: YAG laser. Dent Mater J. 2006;25:2148.
36 S.L. McArthur et al.

99. Ziglo MJ, Nelson AE, Heo G, Major PW. Argon laser induced changes to the carbonate
content of enamel. Appl Surf Sci. 2009;255:67904.
100. Busscher HJ, Vandermei HC, Genet MJ, Perdok JF, Rouxhet PG. XPS determination of the
thickness of adsorbed mouthrinse components on dental enamel. Surf Interface Anal.
1990;15:3446.
101. Taube F, Ylmen R, Shchukarev A, Nietzsche S, Noren JG. Morphological and chemical char-
acterization of tooth enamel exposed to alkaline agents. J Dent. 2010;38:7281.
102. Yoshida Y, Van Meerbeek B, Nakayama Y, Snauwaert J, Hellemans L, Lambrechts P, et al.
Evidence of chemical bonding at biomaterial-hard tissue interfaces. J Dent Res.
2000;79:70914.
103. Lou L, Nelson AE, Heo G, Major PW. Surface chemical composition of human maxillary
first premolar as assessed by X-ray photoelectron spectroscopy (XPS). Appl Surf Sci.
2008;254:67069.
104. Andrade JD. X-ray photoelectron spectroscopy. In: Andrade JD, editor. Surface and interfa-
cial aspects of biomedical polymers. New York: Plenum Press; 1985.
105. Burgos-Asperilla L, Garcia-Alonso MC, Escudero ML, Alonso C. Study of the interaction of
inorganic and organic compounds of cell culture medium with a Ti surface. Acta Biomater.
2010;6:65261.
106. Zelzer M, Alexander MR. Nanopores in single- and double-layer plasma polymers used for cell
guidance in water and protein containing buffer solutions. J Phys Chem B. 2010;114:56976.
107. Chung DJ. Surface modification of polymers for biomaterials. J Dent Res. 2001;80:1348.
108. Hartley PG, McArthur SL, McLean KM, Griesser HJ. Physicochemical properties of poly-
saccharide coatings based on grafted multilayer assemblies. Langmuir. 2002;18:248394.
109. Griesser HJ, Hartley PG, McArthur SL, McLean KM, Meagher L, Thissen H. Interfacial
properties and protein resistance of nano-scale polysaccharide coatings. Smart Mater Struct.
2002;11:65261.
110. Canavan HE, Cheng XH, Graham DJ, Ratner BD, Castner DG. Surface characterization of
the extracellular matrix remaining after cell detachment from a thermoresponsive polymer.
Langmuir. 2005;21:194955.
111. Artyushkova K, Fulghum JE. Multivariate image analysis methods applied to XPS imaging
data sets. Surf Interface Anal. 2002;33:18595.
112. Walton J, Fairley N. XPS spectromicroscopy: exploiting the relationship between images and
spectra. Surf Interface Anal. 2008;40:47881.
113. Artyushkova K. Structure determination of nanocomposites through 3D imaging using
laboratory XPS and multivariate analysis. J Electron Spectrosc Relat Phenom. 2010;

178:292302.
Chapter 3
Biomolecular Analysis by Time-of-Flight
Secondary Ion Mass Spectrometry
(ToF-SIMS)

Daniel Breitenstein, Birgit Hagenhoff, and Albert Schnieders

Abstract The use of time-of-flight secondary ion mass spectrometry (ToF-SIMS)

for biomolecular analysis has made tremendous progress in recent years. This chap-
ter will outline the principal aspects of ToF-SIMS as well as recent technical devel-
opments that have made surface mass spectrometry so valuable for this field.
Furthermore, an overview on relevant biochemical applications based on the four
essential operational modesspectrometry, imaging, depth profiling, and 3D anal-
ysiswill be given. The applications range from the analysis of LangmuirBlodgett
films and tissue sections to the analysis of whole cells. With these results in mind,
we will discuss the chances and limitations of the technique.

3.1 Introduction

Time-of-flight secondary ion mass spectrometry (ToF-SIMS) enables the analysis of

the elemental as well as molecular composition of a samples surface. The funda-
mental process involved is the impact of primary ions, leading to, among other
things, the desorption of secondary ions, which are characteristic for the composition
of the analyzed surface. Recent progress in instrumentation development has facili-
tated the considerable use of this technique for the analysis of biological samples.
The tremendous development of time-of-flight secondary ion mass spectrometry
(ToF-SIMS) in the field of biological applications within recent becomes clear when
one compares the comprehensive review by Pacholski and Winograd published in

D. Breitenstein (*) B. Hagenhoff

Tascon GmbH, Heisenbergstr. 15, Mnster 48155, Germany
e-mail: [email protected]; [email protected]
A. Schnieders
Tascon USA, Inc., 100 Red Schoolhouse Road, Chestnut Ridge, NY 10977, USA
Managing Director, CNM Technologies, Bielefeld, Germany
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 37

DOI 10.1007/978-3-319-01360-2_3, Springer International Publishing Switzerland 2014
38 D. Breitenstein et al.

1999 [1] with other, more recent reviews [24], including this chapter. Pacholski
and Winograd had to point out just over a decade ago that in so-called static
ToF-SIMS experiments, images are usually acquired with only a few counts/pixel
(p. 2979). They emphasized that most of the examples [at that time] involved the
detection of elemental species () rather than molecular species () (p. 3001).
However, they also stated that this trend will surely change in the future (p. 3001).
The aim of this chapter is to outline the principal aspects of ToF-SIMS as well as
recent technical developments that made this prediction become true and that have
made ToF-SIMS a useful method in the field of biochemistry.
Among the aspects that make ToF-SIMS in principle a beneficial tool for bio-
analysis are the following: First, ToF-SIMS just probes the uppermost three
monolayers of a sample. Second, it is especially well suited for the detection and
identification of molecules with a mass of several hundred atomic mass units (u),
which make up a majority of biologically relevant compounds. Third, it offers
imaging capabilities with a lateral resolution as good as a few hundred nanome-
ters. (Recently, even a 3D-imaging option for biological samples was established,
which allows a vertical resolution of 30 nm [5].) The most important technical
development in the last decade making these principal aspects of SIMS accessi-
ble to the field of biochemistry is the use of polyatomic primary ions for desorption.
Additional details on polyatomic primary ion sources can be found in Sect. 3.2.1.2
as well as Sect. 3.3.1.
Besides a short review of the fundamental SIMS process and the presentation of
the current technology of ToF-SIMS, this chapter will also give an overview of rel-
evant biochemical applications based on the four essential operational modes: spec-
trometry, imaging, depth profiling, and 3D analysis. The applications range from the
analysis of LangmuirBlodgett films and tissue sections to the analysis of whole
cells. With these results in mind, we discuss the chances and limitations of the
technique.
Due to limited space, we cannot address a number of aspects: One of them is the
historical development of ToF-SIMS. This aspect has been described in detail by
Benninghoven [6]. Furthermore, as stated in the title of this chapter, biomolecular
analysis by ToF-SIMSnot the detection of elements or isotopesis the focus of
this chapter. Therefore, ToF-SIMS studies focusing on this aspect, such as classical
dynamic SIMS studies (SIMS-microscopy, recently reviewed by Guerquin-Kern
et al. [7]), are excluded from this review.
The chapter is divided into seven sections: Following the introduction (Sect. 3.1),
the basic principle of the ionization process is explained (Sect. 3.2). More detailed
information on the technical setup of ToF-SIMS instruments that makes them
uniquely useful for biomolecular applications are available in the third section.
Examples of applications are discussed using the four ToF-SIMS operational modes
in Sect. 3.4. Section 3.5 deals with strategies to improve the data quality. Following
a brief summary of the information provided (Sect. 3.6), the final section gives an
outlook of the future development of ToF-SIMS in the field of biomolecular
analysis.
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 39

3.2 Principle of the Ionization Process

The three fundamental processes of any mass-spectrometric technique are the

formation, separation, and detection of analyte ions. Since in the case of surface
mass spectrometry, the analyte is present on or in a solid or liquid sample, the ion
formation not only consists of ionization, but also involves the desorption of the
respective particles from the surface. In ToF-SIMS, the formation of secondary ions
is a result of the impact of primary ions onto the surface of the sample, whereas the
separation and detection are performed by a time-of-flight (ToF) analyzer. Because
of its importance, the ion-formation process will be discussed in more detail in the
following section.

3.2.1 Formation of Secondary Ions

The fundamental event in a ToF-SIMS experiment is the desorption (sputtering) and

ionization of surface components after the impact of primary ions. Whereas the
desorption of particles (neutral and charged) can be described fairly well on the
basis of the results of practical experiments as well as molecular dynamics calcula-
tions [8], the intrinsic effects leading to the ionization of some of these sputtered
species (secondary ions) are not yet understood.
In the following section, first the linear collision cascade model of Sigmund
[911] and Thompson [12] will be shortly introduced, because this concept has had
the greatest success in quantitatively explaining observed features of sputtering,
such as total yields, energy distributions of ejected particles, and the yield depen-
dence on energy, atomic number, angle of incidence, and mass of the primary par-
ticle, as well as mass and atomic number of the target atoms. Even if the model
strictly applies to the bombardment of mono-elemental, polycrystalline targets, its
concepts and ideas are the basis for many other models of sputtering, including of
organic systems. The question of the charge of the sputtered particles is not within
the scope of Sigmunds theory. In an additional section, a short overview of the use
of polyatomic primary ions will be given because of its importance for the analysis
of organic, especially biomolecular, samples. For a more detailed introduction, see
the next chapter in this book.

3.2.1.1 Collision Cascade (Monoatomic Bombardment)

It is consensually believed that elastic collisions of a primary ion with surface atoms
displace these surface atoms (primary recoils) and that this movement induces a
movement of further particles (secondary recoils) [13]. In the target, the recoils will
first have a directional distribution peaked in the direction of the incoming primary
particle. After a few collisions, however, their motion in the cascade will be
40 D. Breitenstein et al.

essentially isotropic. Some recoils arrive at the target surface. They can leave the
sample when their energy is high enough to overcome the surface-binding energy of
the solid. Therefore, 95% of the particles desorbed by the impact of a monomolecu-
lar primary ion originate from the first to third outermost monolayers of the sample
[14]. Under conditions prevailing in SIMS, the lifetime of a cascade is of the order
of 1011 to 1012 s and its dimension is of the order of 10 nm [15]. The collision
cascades may be considered independent events under typical conditions (primary
ion currents and beam sizes) used in ToF-SIMS.
As long as the energy of the moving particle in the collision cascade is large
relative to the binding energy of the target atoms, the collision partners can be con-
sidered free particles. Therefore, close to the track of the impacting projectile pen-
etrating into the bulk of the sample, a massive fragmentation of molecules takes
place, resulting in the destruction of organic material and the emission of mainly
atoms. As the distance from the point of impact increases, more intact organic mol-
ecules survive. Some of these molecules may be emitted, mainly as (secondary)
neutrals but partly as positively or negatively charged secondary ions [16]. Most of
the secondary ions obtained in ToF-SIMS experiments are singly charged [15].
More details on the collision cascade under monoatomic bombardment are provided
by Urbassek [17].
As detailed before, emitted particles could be atoms, fragments of molecules,
intact molecules, or even clusters of these particles. The type of particles present in
the final mass spectrum is the same as in classical mass spectrometry, such as elec-
tron impact, or gas chromatography mass spectrometry. However, due to the addi-
tionally needed step of desorption by sputtering, secondary ion mass spectra tend to
show more fragmentation than those particles. A good review of the rules governing
the formation and stability of ions present in mass spectra is given by McLafferty
and Turecek [18].
Especially with the use of polyatomic primary ions, ToF-SIMS is well suited for
the detection and identification of molecules in a medium mass range of several
hundred microns. This makes ToF-SIMS particularly well suited for biomolecular
work, as many important molecules (e.g., drugs) are in this mass range.
Considering the massive fragmentation of the near-surface molecules under
monoatomic bombardment, it is mandatory for an artifact-free measurement to
avoid collecting data from already damaged surface areas. In other words, the prob-
ability for a particular surface site to receive more than one primary ion has to be
negligibly small. This can statistically be ensured by keeping the primary ion dose
density below a value guaranteeing that only 1% of the surface structure is influ-
enced. The value of this so-called static SIMS limit depends on the particular sec-
ondary ion species but typically ranges from a primary ion dose density of 1011-1013
ions/cm2. Upon exceeding this static limit, one reaches dynamic SIMS condi-
tions, including operational modes that are based on the removal of surface material
by sputtering [19, 20]. However, with monoatomic bombardment, most organic sec-
ondary ions are highly fragmented in such an approach [21, 22]. Therefore, these
dynamic SIMS experiments allow predominantly the detection of elements and
their distribution within the sample [7].
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 41

Fig. 3.1 ToF-SIMS mass-resolved images of a blue pigment (m = 416 u, top row) and a green pig-
ment (m = 641 u, bottom row) in a color filter array. The numbers indicate the total counts inte-
grated over the whole image. All images were acquired with the respectively indicated singly
charged primary ion species until the signal intensity dropped to 1/e of the initial valuecorre-
sponding to identical sample consumption. Without going into details, which can be found else-
where [24], this figure may serve as a qualitative example for the statements given in the text on
the different types of primary ion species and primary ion sources

3.2.1.2 Effect of Cluster Primary Ions (Polyatomic Bombardment)

Cluster ions, such as SF5+ [23, 24], C60+ [2527], Au3+ [28], or Bi3+ [29], consist of
multiple atoms. Upon impact on the sample, these clusters break apart [30, 31],
leading to a multiple, simultaneous bombardment of a small spot. One of the obser-
vations possible under these conditions is that the secondary ion yield, namely, the
number of secondary ions generated per impacting primary ion, is drastically
increased compared to monoatomic bombardment [32, 33]. The increase is nonlin-
ear with respect to the number of projectile particles: For instance, changing the
primary ion from Au1+ to Au3+ on an organic polymer additive sample (Irganox
1010) does not lead to a threefold but to an almost 45-fold increase of the secondary
ion yield for the intact molecule (quasi-molecular ion) on thick organic substrates
[34]. The key observation for the use of cluster primary ions in molecular ToF-
SIMS analysis is that the yield enhancement is not accompanied by a likewise
increase in surface damage [31, 3438]. This increase in efficiency corresponds to
an increase in the ultimate number of secondary ions, which can be desorbed from
a surface area before they are completely destroyed by ion bombardment. A com-
parison of the use of monoatomic and polyatomic primary ions on identical organic
samples appears in Fig. 3.1. The example is from a study highlighting the differ-
ences in performance of various primary ion species [24]. One can clearly see how
the image quality improves from monoatomic to polyatomic primary ions.
42 D. Breitenstein et al.

The nonlinear effects (increase in yield and efficiency) are assumed to originate
on the one hand from the deposition of the collision energy closer to the surface, as
cluster ions have a lower kinetic energy per constituent than atomic primary ions
have [31, 39]. On the other hand, it is additionally and/or alternatively explained by
the overlap of the collision cascades caused by the individual constituents of the
polyatomic projectile [8]. Finally, there are results indicating a more efficient pro-
tonation (i.e., ionization by attachment of a proton) of originally neutral molecules
[40, 41]. The anticipated accumulation of available protons on the sample surface is
expected to be the result of differences in bond-breaking mechanisms in polyatomic
bombardment compared to monoatomic bombardment [40, 42].
A number of groups (such as Gillen and Roberson [23]; for a more complete list
and discussion, see the review of Mahoney [43]) have reported the detection of
molecular secondary ions during depth profiling, that is, sputter removal of sample
surface layers (see Sects. 3.4.4 and 3.4.5), in favorable cases. The results can be
explained simplistically by assuming that all molecules damaged during sputtering
are removed under polyatomic sputtering in the same event, therefore exposing an
undamaged surface [44]. The traditional distinction between static and
dynamic ToF-SIMS conditions therefore does not play such a critical role for
polyatomic projectiles and thick organic layers.
More details on the effect of the use of cluster ions on ToF-SIMS results can be
found in the substantial review by Wucher [32] and in the next chapter in this book.
Nevertheless, the physics and chemistry of cluster ion bombardment are far from
being well understood and are currently widely discussed in the SIMS community.

3.3 Instrumentation

This section will discuss some technological aspects of ToF-SIMS instrumentation.

We place special emphasis on the consequences of certain design features on the
analysis of biomolecular samples.

3.3.1 Primary Ion Guns

To enable the sputtering and ionization process on the surface, primary ions have to
be generated, accelerated, and focused onto the target surface. Two types of ion
sources are used for primary ion generation in ToF-SIMS: the liquid metal ion
source and the electron impact ion source (see Sect. 3.3.1.1). Other ion sources,
including surface ionization sources and duoplasmatron sources, which are also
used on SIMS instrumentations, are of minor importance for biological applica-
tions. Their description goes beyond the scope of this review and can be found
elsewhere [45]. Section 3.3.1.2 discusses the mass separation and pulsing of pri-
mary ions, and Sect. 3.3.1.3 describes the focusing of primary ion beams.
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 43

3.3.1.1 Primary Ion Sources

3.3.1.1.1 Liquid Metal Ion Source (LMIS)

Primary ion guns using a liquid metal ion source for ion generation offer the highest
spatial resolution of all available ion guns. In an LMIS, the source consists of a
needle with a tip radius of a few microns. The needle is covered with a thin layer of
a liquid metal provided from a reservoir. If the metal is not liquid at room tempera-
ture, it can be melted by a heating filament. Primary ions are extracted from the tip
of the needle by an extraction voltage (in the range of 510 keV) applied to an
opposing electrode. The emitter itself is typically at a potential of 530 keV above
ground, resulting in a corresponding energy of the primary ions when they reach the
target. Similar to the ionization process in electrospray mass spectrometry, the liq-
uid metal forms a Taylor cone at the tip of the needle and the primary ions can be
continuously extracted toward the sample.
The advantage of the LMIS against other ion sources lies in the well-defined
point of primary ion extraction. Ions are formed from the Taylor cone, which is
much better defined than the tip itself. The ion source therefore has a virtual size
down to 10 nm in diameter [45]. This small virtual size allows exceptional focusing
of the primary ion beam onto the target. Focusing is optimized by an ion-optical
lens system inside the primary ion column (see Sect. 3.3.1.3). In state-of-the-art
liquid metal ion guns, the beam can be focused down to some 10 nm in DC opera-
tion or approximately 100 nm in pulsed operation (see also Sect. 3.3.1.2).
The metal and therefore the primary ion species used in an LMIS depend on
several critical parameters: First, the material has to be ionized easily in the field-
emission process. Moreover, as the primary ions have to be created from a liquid
phase, a low melting point is of high interest from a technological point of view.
Furthermore, as a consequence of the time-of-flight principle within the ion gun
(see Sect. 3.3.1.2), the element used as primary ion should either be monoisotopic
or be easily available in isotopically separated species. In this perspective, the price
of the material is also of interest. Finally, it has been shown (as discussed above)
that a high mass and a polyatomic nature of the primary ion are of high interest for
improved results on organic samples.
The traditional metal used in an LMIS is gallium [69Ga, two Ga isotopes, melt-
ing point (mp): 302.9 K]. It has many benefits, including a low melting point and the
availability as isotopically enriched 69Ga. Furthermore, it is widely used in other
applications, such as focused ion beam (FIB) bombardment for material processing.
However, Ga has a relatively low atomic mass, resulting in a low efficiency for sec-
ondary ion emission. Indium (115In, two In isotopes, mp: 429.8 K) can be seen as a
developing step because of its higher mass. More recently, gold (197Au, monoiso-
topic, used as a eutectic mixture with germanium; mp: ~400 K [28]; or pure gold,
mp: 1,337.3 K) and bismuth (209Bi, monoisotopic, mp: 544.4 K) were introduced;
gold and bismuth primary ion sources have been a breakthrough for the analysis of
both organic and biological samples. Gold and bismuth sources already profit from
the high mass of the monoatomic primary ion; however, the main improvement
44 D. Breitenstein et al.

compared to the other elements listed above is the formation of polyatomic cluster
ions (Myx+, with x = 1, 2 and y = 1, 3, 5, 7). The benefits of both aspects (higher
primary ion mass and polyatomic primary ion species) for organic analysis are
visualized in Fig. 3.1 and were discussed in Sect. 3.2.1.2. The use of a Bi-cluster
LMIS typically is preferable to an Au-LMIS because of a better performance in
terms of achievable cluster currents and beam quality (focus, pulse width) and the
wider variety of available cluster species [29].

3.3.1.1.2 Electron Impact Ion Source

In an electron impact (EI) ion source, gas-phase atoms or molecules are impacted
by electrons with an energy high enough to ionize them. This process requires the
continuous leaking of a gas or vapor into the ionization chamber of the source.
Electrons are emitted from a heated filament and are accelerated into this chamber,
where they interact with the gas or vapor particles. Ionized gas or vapor particles are
captured by an electrical extraction field and are accelerated through an ion-optical
system onto the target.
Traditionally, noble gases, oxygen, and SF5+ were used in EI sources. Oxygen
became one of the dominant primary ions often used for depth-profiling of inor-
ganic samples [46]. In the field of biological applications, the recent development of
a C60 source [27] is of great interest: Applying the ionization process to sublimed
buckminsterfullerenes opens the possibility to generate a high-mass (720-u) multi-
atomic ion species.
While the polyatomic nature of the C60+, C60++, and C60+++ ions is reported to be
advantageous for biochemical applications, the ionization process in the EI source
only allows focusing of the primary ion beam down to a diameter of a few microns
in the best case. Thus, the imaging capabilities of a C60 source are (to date) inferior
to the LMIG sources (see also Fig. 3.1) [29].

3.3.1.2 Separation and Pulsing of Primary Ions

The continuous extraction of primary ions from an ion source without further
manipulation of the beam would lead to a continuous bombardment of the sample
surface and therefore to the continuous generation of secondary particles. In such a
scenario, the mass separation of secondary ions by a ToF analyzer (see Sect. 3.3.2)
does not easily work. Thus, for ToF-SIMS analysis, the surface bombardment with
primary ions is pulsed, which is often performed by deflecting the primary ion beam
in and out of the optical axis of the primary ion gun and blanking it with appropri-
ately placed apertures.
In most primary ion sources, a broad range of ions are generated. For example,
an Au-LMIS operating with an AuGe alloy produces Gexy+ and Auxy+ ions, with x
and y being a natural number between 1 and 5 and 1 and 3, respectively, as well as
mixed AuGe cluster ions. Similarly, a C60-EI source produces C60y+, with y being a
natural number between 1 and 3, as well as a wide variety of fragments of the
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 45

buckminsterfullerene molecules. All these ions are accelerated by the extraction

field with the same energy into the primary ion gun column. Due to their different
mass-to-charge ratios, they then have various velocities. Thus, different species in a
pulsed ion beam would hit the sample surface at different times, leading to a widen-
ing of the primary ion pulse. Three methods can be used (sometimes together) to
separate the different species in a primary ion pulse. First, the time-of flight princi-
ple (see also Sect. 3.3.2) can be used. After a beam is pulsed, an additional set of
blanking plates further down the ion column is used to select a distinct ion package
from a primary ion pulse at the time when these ions pass the blanking plates. At all
other times, the beam would be blanked. A second method uses a Wien filter, which
consists of perpendicular electric and magnetic fields. Only ions of a certain veloc-
ity (i.e., charge-to-mass ratio) can pass through such a filter. The third method uses
a 90-deflection unit for pulsing and mass-separating the beam [47].
In order to achieve the highest possible mass resolution in a time-of-flight mass
analyzer, primary ion pulses should be as short as possible. Ion packages of several
tens of nanoseconds in length can be further compressed in time by using an axial
buncher (two plates perpendicular to the ion-optical axis). When the ion package is
in the center of the two plates, a high-voltage pulse is switched to the back plate,
accelerating the ions at the end of the package more than those in the front. By
adjusting the amplitude of the bunching pulse, it is possible to shorten the pulse
width at the target to less than 1 ns. This allows one to achieve mass resolutions
(FWHM) of more than 10,000 on commercially available ToF-SIMS instruments
[48]. However, the shortest pulse lengths are achieved at the expense of either the
focus quality or the measuring time.

3.3.1.3 Focusing of Primary Ions

The ion beam is focused by an ion-optical lens systems. As the setup of the lens sys-
tems varies among manufacturers and depends on the analytical goal of the analysis,
just a brief general overview is given here. More details can be found elsewhere [45].
The main components of the ion-optical system are the apertures and the electro-
static lenses. Apertures are used to blank part of the beam, while electrostatic lenses
are used to focus the primary ion beam. Similar to optical lenses, aberrations should
be avoided upon use of the lens system. In this review, only two types of aberrations
are discussed, spherical and chromatic.
The most important aberration in an ion-optical system is the spherical aberra-
tion, meaning that ions with a movement parallel with but far from the axis of the
beam are focused to a different spot than those moving on the axis of the beam. To
avoid this type of aberration, one can use a circular aperture on the axis of the beam.
Such a pupil allows only those primary ions that are close to the beam axis to pass.
Thus, ions that would lead to a spherical aberration of the focus (i.e., those far away
from the beam axis) are blanked. It should be noted that removing these primary
ions increases the focus quality but decreases the number of primary ions hitting the
sampleand thus increases the measuring time needed for the same sensitivity or
dynamic range of a measurement.
46 D. Breitenstein et al.

The second important aberration to deal with is the chromatic aberration: Ions
with different velocities are brought to a different focus. One instance where the
chromatic aberration limits the focus of the analysis beam occurs during the electro-
dynamic bunching of an ion pulse in order to reduce its pulse length (see
Sect. 3.3.1.2). Bunching leads to an increase in the energy spread, making it difficult
to optimize the focus and mass resolution at the same time.
In dealing with these limitations, it is important to choose the correct operational
mode of a primary ion gun for a certain application. In general, it is possible to
optimize only two of the three following beam characteristics at the same time:
pulse current (a high pulse current corresponds to high sensitivity, high dynamic
range of a measurement, or short acquisition times), pulse width (a short pulse
width is necessary for a good mass resolution), and focus. A typical operational
mode is the spectroscopy or bunched mode (pulse width less than 1 ns; pulsed
ion currents of a few pA, focus of some microns). If a better lateral resolution is
needed, one can either sacrifice beam current while still maintaining a short pulse
width (good mass resolution) or work with long ion pulses (effectively unit mass
resolution) while maintaining a high pulse current.

3.3.2 Secondary Ion Separation and Detection:

Time-of-Flight Analyzer

In general, surface material is limited. There are only approximately 106 molecules
in an area of a square micron, and the analyte of interest is usually a minor compo-
nent in the sample (as it is often in biological applications, such as when tracing the
active ingredient of a drug in a cell). Knowing that the secondary ion yieldthe
portion of molecules that are successfully transformed from their state in the sample
into secondary ions, which then can be detected by a mass spectrometeris in the
best cases in the percent range, but more often several orders of magnitude smaller,
makes the task of ion separation and detection even more demanding. In order not
to lose any information, an efficient and parallel detection of almost all generated
secondary ions is of high interest. For this purpose, the time-of-flight (ToF) analyzer
is superior to most other mass-spectrometric analyzer systems [49].
A linear ToF analyzer typically consists of an accelerating section and a drift
section. In the accelerating section, all generated secondary ions of a given polarity
are accelerated to the same nominal kinetic energy. The resulting velocities of these
secondary ions depend on their mass-to-charge ratio. The accelerated ions are then
transferred into the drift section. The drift section allows for a field-free drift of the
ions. As the velocity of the secondary ions differs, they are mass-separated on their
way to the detector: Faster (i.e., lighter) secondary ions arrive earlier at the detector
than slower (i.e., heavier) ions.
Due to the strength and the direction of the individual recoils responsible for
desorption, secondary ions already have initial velocities, which are added to the
velocity they gain in the acceleration section of the mass spectrometer. As these
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 47

Fig. 3.2 Schematic designs of a reflectron (electrostatic mirror) analyzer (left) and a sector field
analyzer (right). Secondary ions with a higher initial kinetic energy (dark trace) are sent on a lon-
ger path compared to ions with a lower initial kinetic energy (light trace)

initial velocities are not all the same, the separation process is hampered. To
compensate for this effect, most ToF-SIMS instruments today are equipped with an
energy-focusing analyzer. The basic idea of such a device is to increase the flight
path of ions with a higher initial velocity compared to ions of the same mass-to-
charge ratio with a lower initial velocity. For this purpose, electrostatic fields can be
used that act as an ion mirror (a so-called reflectron; Fig. 3.2). As high-energy ions
of one species penetrate deeper into this electrostatic field, their flight paths are
longer than those of ions of a lower initial energy. Under optimal alignment of the
reflectron, all secondary ions of the same mass-to charge ratio arrive at the detector
after the same flight time. A second option for energy focusing used in ToF-SIMS
is the electrostatic sector analyzer (Fig. 3.2). Within the radius of the electrostatic
sector analyzer, the drift paths of higher-energy ions are longer than those of lower-
energy ions [49]. Modern ToF-SIMS instruments equipped with either of these
devices achieve mass resolutions (FWHM) of more than 10,000.
Ion detection is performed using a multichannel plate configuration. The indi-
vidual designfor instance, a double-channel plate arrangement (Chevron) or a
combination of channel plate, scintillator, and photomultiplier (DalyThompson)
can vary, but the basic principle is the same: Each impact of an incoming ion causes
an electrical signal, which is amplified and then registered.

3.3.3 Charge Compensation

The impact of primary ions and the desorption of charged particles (secondary ions
and even to a greater extent secondary electrons) lead to charging of the bombarded
area. On conductive samples, this net charge is instantly dissipated. However,
48 D. Breitenstein et al.

insulating samples can keep a localized charge for a long time, which affects the
analysis. Not only is the primary beam deflected, which can degrade the lateral
resolution, buteven more importantcharging affects the extraction of secondary
ions by nonuniformly changing the extraction field close to the analyzed surface.
Depending on the samples conductivity, the path of secondary ions is deflected
from the default path through the mass spectrometer, resulting in a poor mass reso-
lution or, in the worst case, in a complete loss of transmission to the detector.
To avoid such artifacts, charge compensation is included in modern ToF-SIMS
instruments. The sample is flooded with low-energy electrons supplied by a heated
filament between the primary ion pulses. The low-energy electrons are attracted by
positively charged sample regions, leading to a self-adjustment of the whole surface
to a uniform potential. It should be noted that even electrons of some 10 eV of
energy can destroy certain organic bonds and change the sample being analyzed.
Thus, minimizing the charge compensation is desired [50, 51].

3.3.4 Technology of Sputtering (Depth Profiling)

Depth-profiling methods are used to sample subsurface regions of samples. Depth

profiles are generated by using a sputter source to erode the topmost layer of the
sample surface using massive ion bombardment between analysis cycles. It is of
importance to note that the pulsed nature of ToF-SIMS spectral or image analysis
does not (significantly) erode the sample; thus, erosion during a depth profile is
often performed using a quasi-DC ion beam. For the sputter erosion, either the pri-
mary ion beam utilized for analysis can be used or, in the more convenient case, the
instrument is equipped with a second ion beam system, which is optimized for
material removal (dual-beam approach [46]). There are two advantages of dual-
beam depth profiling compared to single-beam depth profiling, described next.
First, the number of primary ions in an analysis mode is not adequate for a quick
sputtering of the surface. Thus, in the single-beam setup, a second setting for the
voltages and timing of the primary ion gun has to be used, allowing an enhanced
bombardment of the surface. Constant switching from one setting to another is
time-consuming. In contrast, in dual-beam mode, the time needed for analyzing the
secondary ions (while they are drifting through the ToF-MS) can already be used in
many cases for sputtering (interlaced mode).
Second, the sputter ion species, energy, and angle of incidence can be chosen
independently from the analysis ion in the dual-beam approach. Thus, low-energy
ions or a cluster ion beam such as C60 can be used for sputtering in order to obtain a
high depth resolution and/or low damage of the remaining material on the sputtered
surface, whereas high-energy ions (e.g., from an LMIS), which can be well focused,
are used to analyze the center of the crater. A typical setup for the analysis of bio-
logical systems uses C60y+ ions for sputtering and Bi3+ for high-lateral resolution
analysis of the sample [52] (see also the advantages and disadvantages of the differ-
ent ion sources as discussed in Sect. 3.3.1.1).
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 49

3.4 Operational Modes

3.4.1 Overview

As shown in Table 3.1, ToF-SIMS analysis can be performed in four different

operational modes, which are demonstrated on different samples in Figs. 3.3, 3.4,
3.5, and 3.6. One differentiation among operational modes is between those where
(almost) no sample erosion takes place (static SIMS) and those where sample ero-
sion is performed (dynamic SIMS). Furthermore, one can distinguish between mea-
surements where lateral distributions of analytes are probed or not.
The simplest operational mode is the spectrometry mode (low sample erosion,
no lateral probing). The result of this type of measurement is displayed as a mass
spectrum, as is well known from other mass-spectrometric techniques (Fig. 3.3).
Information on the chemical composition of a surface can be gained from such a
spectrum. All ToF-SIMS measurements start by using this mode.
More complex is the imaging mode (low sample erosion, lateral probing). In
most cases, a focused primary ion beam is rastered in both the x- and y-directions
over the sample surface. A full mass spectrum is collected at every pixel, and a few
user-selected species are monitored during data collection. A square of 256 256
pixels leads to a data set of 65,536 individual spectra. The raw data taken from this
kind of experiment can be presented in various ways; for instance, spectra can be
retrospectively generated from any subset of pixels. Most frequently, the data are
presented in mass-resolved images (Fig. 3.4). The intensity of one secondary ion
species in each of the pixels is represented according to a color scale, leading to a
two-dimensional image of the surface. It is noteworthy that ToF-SIMS allows the
simultaneous generation of a mass-resolved image for any detected secondary ion.
When a depth profile is being performed, the spectrometric data acquisition is
alternately interrupted by a sputter process eroding the surface. Thus, a small region
of the sample is destroyed when this process is used: The erosion process leads to
the formation of a sputter crater. The analysis area is typically centered in this crater.
For every analysis cycle, the complete spectrum is stored in a raw data stream.
Therefore, mass spectra from different depths of the sample can be gained. The
intensity of an analyte can be plotted against the time needed for sputtering
(Fig. 3.5). Thus, the vertical distribution of any detected secondary ion within the
sample can be probed.
Three-dimensional analysis is performed by combining the imaging mode with
sputtering of the sample. Thus, a three-dimensional representation of the detected

Table 3.1 Operational modes in ToF-SIMS

With erosion (sputtering)
With low erosion (static SIMS) (dynamic SIMS)
Without lateral resolution Spectrometry (Fig. 3.3) Depth profiling (Fig. 3.5)
With lateral resolution Imaging (Fig. 3.4) 3D microarea analysis (Fig. 3.6)
50 D. Breitenstein et al.

Fig. 3.3 ToF-SIMS spectrum of positively charged secondary ions detected on a pure monolayer
of the lipid DPPC (dipalmitoyl phosphatidylcholine) on a gold- and tin-containing substrate

analytes can be obtained (Fig. 3.6). The amount of data generated in this approach
is enormous: For each depth, a complete set of image data (e.g., 65,536 spectra for
a raster of 256 256 pixels) is acquired. In the example shown in Fig. 3.6, image
acquisition was performed in 120 depths. Thus, almost 7.9 million spectra were
generated. To present the data, virtual xy-, xz-, and yz-sections can be reconstructed
from the raw data stream. One must keep in mind when interpreting the 3D data that
the sections are not one-to-one representations of the spatial distribution of analytes.
This is due to the fact that even in the ideal case of uniform erosion by the sputter
process, which is not always true especially for nonhomogeneous materials, the
original sample surface is quite often not flat. Thus, for example, the sodium signal
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 51

Fig. 3.4 The DNA-staining fluorophore ethidium homodimer (EthDI) was detected by confocal
scanning microscopy (a, b) as well as by ToF-SIMS imaging (c). Due to the low information depth,
the lipids covering the nuclei had to be removed by appropriate fixation protocols in order to allow
a ToF-SIMS 2D imaging of the nuclei (Reprinted from Ref. [79] with permission)

Fig. 3.5 Depth profile of a layered polymer system (5 Irganox 3114 delta layers in an Irganox
1010 matrix) obtained in dual-beam mode, sputtering with a 20-keV C60 beam, analysis with a
25-keV Bi3 beam, sputter angle: 45; sample rotation
52 D. Breitenstein et al.

Fig. 3.6 Three-dimensional ToF-SIMS analysis of glutardialdehyde-fixated normal rat kidney

(NRK) cells (redgreenblue color overlay for a correlation analysis). (a)(d) Horizontal
xy-sections. The scale bar in (d) corresponds to 20 mm. Pooled signals of amino acid fragment ions
are represented in red (b), those of phospholipids in green (c), and substrate-derived secondary
ions are depicted in blue (a). (e) Vertical xz-sections; and (f) mathematically corrected xz-section
showing the topography of the cells on the substrate, which is assumed to be flat (Reproduced with
permission from Ref. [169])

in Fig. 3.6e derived from the plain glass substrate does not show up in an even
z-layer but rather reflects the initial inhomogeneous topography of the surface.
Software algorithms are being developed to correct for these effects.
In the following sections, the operational modes will be discussed in more detail.
The discussion will focus on their use for biochemical applications.
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 53

3.4.2 Spectrometry

Analyzing the sample in spectrometry mode is reasonable if the aim of the analysis
is to obtain laterally non-resolved information on the first three outermost monolay-
ers of the sample. A ToF-SIMS spectrum of a pure monolayer of the lipid dipalmi-
toyl phosphatidylcholine (DPPC) on an Au- and Sn-containing substrate is shown
as an example in Fig. 3.3. Besides the inorganic substrate ions (e.g., Au+ and Sn+),
quasi-molecular ions (e.g., DPPC + H+) and fragment ions of the analyte can be
detected. The strength of ToF-SIMS analysis in spectrometry mode is that it allows
a simultaneous detection of various types of analytes. It is limited to neither one-
component systems nor to a certain type of species. In the following, a short over-
view of spectrometric studies of biological systems with ToF-SIMS is given.
Since the molecular mass of typical lipid molecules is ideal for detection with
ToF-SIMS, the analysis of all types of lipids is a focus of ToF-SIMS research.
Whereas high-mass biopolymers usually fragment during ToF-SIMS analysis, and
most often statistical evaluation procedures have to be used for identification, lipids
can routinely be detected as quasi-molecular ions. More information on lipid detec-
tion can be found in the imaging section (see Sect. 3.4.3).
ToF-SIMS spectra of proteins are typically dominated by amino acid fragments
(NH2CHR+, with R = amino acid residue) [5355].1 However, protein analysis is not
limited to the detection of these fragments; larger ions [56] and even intact small
proteins and peptides up to a mass of approximately 10,000 u can be detected when
adequate sample preparation and analysis protocol are followed. Examples for
intact protein detection are found for angiotensin (M = 1,045 u) and its derivatives
[54], the B-chain (M = 3,496 u) as well as the complete molecule (M = 5,778 u) of
insulin [57], porcine renin (M = 1,759 u), and bovine ubiquitin (M = 8,563 u) [58].
Besides these single-component systems, protein mixtures of adsorbed multicom-
ponent protein films were analyzed by Wagner et al. [59]. A very interesting
approach is the detection and identification of tryptic peptides of a protein by
ToF-SIMS [6062].
Although DNA oligomers were found to fragment [58, 63, 64], May et al. [65]
were able to distinguish representatives of this molecular species by statistical eval-
uation procedures (see Sect. 3.5). In addition, they give a comprehensive overview
of the ToF-SIMS fragmentation patterns of nucleobases, nucleosides, and nucleo-
tides. Another example for analyzing multiple-component systems of nucleic acids
is the analysis of combined DNA and peptide nucleic acid (PNA) samples by
Arlinghaus et al. [66].
In addition to lipid, protein, and DNA samples, other biotic and xenobiotic sub-
stances have also been analyzed: The analysis of different monosaccharides leads to
the detection of quasi-molecular ions as well as typical fragmentation patterns [67].
In other studies, spectra of fluorophores [6870], bone minerals [71], or pharmaceu-
ticals [29, 72, 73] are presented.

1
Lists of the composition and mass-to-charge ratio of typical secondary ions characteristic of
amino acids can be found, for example, in a study by Michel et al. [55].
54 D. Breitenstein et al.

In principle, any biochemically relevant substance that is vacuum-sustainable

can be analyzed. Whereas molecules larger than 2,000 u most often fragment, the
chance to detect quasi-molecular ions of smaller molecules is high when polyatomic
primary ion sources are used. Of note, in some cases spectrometric data can be used
for semiquantitative evaluation. A striking example is the determination of an
enzyme activity in correlation with the ToF-SIMS signal intensities of its substrate
and product [74]. Good examples are also found for self-assembled monolayers
[75], ethanolamin [76], and proteins [77].

3.4.3 Imaging

Mass-resolved images are most often interpreted to represent at least semiquantita-

tively the distribution of the analyte on the surface of the sample. In doing so, it has
to be kept in mind that the simple approach of assuming a linear correlation of the
secondary ion signal with the surface concentration of a species may lead to misin-
terpretation. This is due to the fact that in SIMS the number of desorbed secondary
ions is a function not only of the surface concentration but also of the respective
chemical environment (matrix effect), which can influence the ionization probabil-
ity [15] as well as the sputter yield [78] by up to several orders of magnitude.
However, a semiquantitative interpretation of images in biomolecular applications
is still often possible, because the general matrix does not radically change over the
area of an image. Since secondary ion yields of different species can vary by several
orders of magnitude, the information accompanying a secondary ion image should
include not only the field of view but also the number of ions detected in the com-
plete image as well as in the brightest pixel in order to allow data interpretation with
respect to statistical consideration.
An example for mass-resolved secondary ion images is shown in Fig. 3.4 [79].
The DNA of epithelial cells [normal rat kidney (NRK) cells] has been stained with
the fluorophore ethidium homodimer (EthDI), which allows the localization of
the nuclei by confocal scanning microscopy (Fig. 3.4a, b). The corresponding sec-
ondary ion maps (Fig. 3.4c) demonstrate that ToF-SIMS is also able to detect EthDI,
showing the same lateral distribution. Due to the low information depth, the lipids
covering the nuclei had to be removed by appropriate fixation protocols in order to
allow a ToF-SIMS 2D imaging of the nuclei. For a more detailed discussion of this
example, see [79]. It is of importance to note that staining or labeling is not required
for ToF-SIMS. For the study mentioned here, it was required for the confocal scan-
ning microscopy analysis only.
In the following, three areas of biochemical imaging applications in which ToF-
SIMS has been delivering important results are discussed in more detail: the analy-
sis of thin monolayer films (LangmuirBlodgett films or self-assembled monolayers,
Sect. 3.4.3.1); the analysis of tissue sections (Sect. 3.4.3.2); and the analysis of cell
culture samples (Sect. 3.4.3.3).
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 55

3.4.3.1 LangmuirBlodgett Films/Self-Assembled Monolayers

LangmuirBlodgett (LB) films and self-assembled monolayers (SAM) are

comparatively well-defined organic systems that came early in the scope of molecu-
lar imaging by ToF-SIMS [1] because of the perfect match of the surface sensitivity
of ToF-SIMS with the monolayer nature of LB films and SAMs.
Generating LangmuirBlodgett films is comparatively easy: Lipid mixtures are
spread onto the water or buffer surface of a film balance, where their hydrophilic
end is oriented to the water/buffer. These layers are subsequently transferred onto a
substrate [80]. Usually, gold-coated glass [81, 82] or mica [83, 84] substrates are
used for ToF-SIMS applications in order to allow an easier ionization (see
Sect. 3.5.1) and to avoid a charging of the surface. Alternatively, SiO2/Si substrates
can be used [85]. ToF-SIMS analysis of LB systems allows a view of the mixing
behavior as well as on the phase behavior of the involved materials: One early pub-
lication observing both phenomena was published by Leufgen et al. [69], which
describes the distribution of a native lipid (DPPC) and one of its fluorescent dye
analogs. Thus, the distribution of the dye could be probed by two different tech-
niques: by fluorescent light microscopy as well as by ToF-SIMS. As can be seen in
this and other studies, fluorescence microscopy [81, 86, 87] and also atomic force
microscopy [86, 88, 89] can be used to validate ToF-SIMS results.
A close correlation of a model system to the real-world environment is found for
studies on LB films of pulmonary surfactants [81, 82, 86, 90, 91]. A pulmonary
surfactant is a lipid/protein mixture preventing alveolar collapse in the mammalian
lung. These studies aim to detect a protein-induced lipid demixing, necessary for
lipid-enrichment processes in the mammalian body. They are accompanied by
experiments where the complete extract of lipids and hydrophobic proteins from
bovine lungs is analyzed [87].
The use of deuterated lipids in LangmuirBlodgett studies has also been demon-
strated [83, 86]. Here, hydrogen atoms in certain lipids have been replaced with the
heavier isotope of hydrogen (deuterium), thereby placing a unique marker into one
of the lipids. Using this strategy, researchers have distinguished similar lipids dur-
ing ToF-SIMS analysis by their isotopic label, using care so that the mixing and
phase behavior do not significantly change upon isotopic labeling [86].
The so-called raft theory, which, in short, predicts that cell membrane contains
cholesterin-rich micro domains [92], increased the interest in protein/lipid model
systems. Therefore, recent studies have dealt with lipid and lipid/protein model sys-
tems expected to be relevant for raft formation [93]. Nevertheless, it has to be
pointed out that the expected size of rafts (down to 25 nm [94]) is smaller than the
lateral resolution of ToF-SIMS (down to 100 nm, [29]).
It is noteworthy that artifacts were observed using ToF-SIMS for imaging and
semiquantitative analysis of LangmuirBlodgett films. Sostarecz et al. reported higher
yields of DPPC-specific fragments in an expanded film, where the molecules are
loosely packed, compared to a condensed film, where the molecules are densely
packed in a well-ordered manner [85]. This finding does not correlate with the expected
molecule density, which is higher in a condensed film. Biesinger et al. reported an
56 D. Breitenstein et al.

inversion of yield contrast upon exceeding the static limit on LangmuirBlodgett films
[84]. Both studies were performed under monoatomic bombardment, and thus further
studies under polyatomic bombardment would be of interest.
Self-assembled mono- and multilayer systems offer another opportunity to benefit
from the surface sensitivity of ToF-SIMS [9599]. Most often they are used as model
systems for the determination of ToF-SIMSspecific features. Nevertheless, other
analytical questions, such as the temperature dependence of the stability of these
systems [100], can be addressed. Finally, it is also possible to analyze freeze-dried
supported bilayers and compare the respective ToF-SIMS results to those obtained
from LangmuirBlodgett films and bulk material of identical composition [101].

3.4.3.2 Tissue Sections

Tissue sections are found to be good ToF-SIMS samples because they expose the
area of interest and the analyte and make them directly available for analysis. The
avoidance of sample contamination by lubricants or other unwanted substances dur-
ing the preparation of the tissue section is critical for a successful analysisthis is
not unique to ToF-SIMS. For this reason, the use of fixatives and cryoprotectives is
also not reasonable. As summarized by Richter et al. [102], to date two preparation
methods are available: (1) cryofixation followed by freeze-fractioning and freeze-
drying or (2) plunge-freezing and cryostat slicing followed by freeze-drying are
applied to the sample. From a technological point of view, the recent development
of a combined cryosectioning and ToF-SIMS instrument is interesting [103] because
the dehydration step can be avoided.
Meanwhile, a noticeable number of ToF-SIMS imaging studies on tissue sec-
tions have been published: For example, the distributions of triglycerides, phospho-
lipids, cholesterol, and fatty acids in mouse brain [104106] and freshwater snail
brain [107] sections were analyzed. In other studies, a part of a rat brain, the cere-
bellar cortex, was probed by ToF-SIMS imaging. It has been possible to detect the
distribution of cholesterol and galactosylceramides in different cell types of these
tissue sections [48, 108]. In an earlier work, Nygren et al. were able to visualize the
distribution of phosphatidylcholine [109], among other substances, and with the
help of metal-enhanced SIMS (see Sect. 3.5.1) the distribution of cholesterol [110]
in rat kidney. Another group obtained information on the cellular metabolism in the
leg section of model mice with Duchenne muscular dystrophy by identifying the
distribution of various lipids within the sample [111]. Furthermore, the identifica-
tion of different tissue types in human adipose tissue gained by biopsy was possible
by determining the lipid distribution [112]. Finally, the distribution of different
types of diacylglycerides was detected inside and outside certain vesicles in stea-
totic (i.e., fatty) liver samples [113].
A noteworthy paper by Brunelle et al. [3], which also includes a review of bio-
logical tissue imaging by ToF-SIMS, describes a study comparing ToF-SIMS and
MALDI-MS results on a single rat brain section. This study and a later study by
Monroe on a spinal cord section [114] show the complementary nature of these two
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 57

techniques: Whereas first a number of (small-sized) lipids could be probed by

ToF-SIMS, MALDI-MS was able to detect (larger-sized) peptides on the same sam-
ple. It is noteworthy that while ToF-SIMS was performed without further sample
treatment and basically nondestructively, MALDI analysis required the application
of a matrix and consumed the sample by ablation of surface material.

3.4.3.3 Single-Cell Analysis

Similar to the analysis of tissue sections, the adequate preparation of the sample is
the crucial step in the analysis of single cells. One approach is the freeze-drying of
cells, with subsequent analysis of their surfaces [115, 116], resulting in the visual-
ization of the distribution of cholesterol on the respective cell surfaces. However,
such approaches may lead to artifacts due to contaminations derived from the cell
culture medium. Additional washing steps, as shown by Berman et al. [117], may
better help to circumvent this challenge.
In another approach, Fartmann et al. analyzed freeze-fractured and freeze-dried
osteoblasts, which were cultured on silicon, thus obtaining hydrocarbon signals
[118]. The freeze fracturing was performed using a sandwich method and is similar
to the protocol used in other studies: Polystyrene beads were added to the buffer
solution before a second silicon wafer was placed on top of the sample. After rapid
cool down in nitrogen-cooled propane, the fracturing was performed by disruption
of the two silicon wafer pieces [70, 119, 120]. It is worth noting that some research
groups [121] can perform the critical preparation steps under vacuum conditions
within a ToF-SIMS instrument, minimizing the risk of contamination.
Vaidyanathan et al. [122] go to the other extreme: They were able to image heat-
fixed bacteria populations. Other groups preserved the cell by vitrification in treha-
lose [123]. Monroe et al. successfully detected the distribution of vitamin E on a
single neuron [124], not giving information on their sample preparation procedure.
Sjvall and coworkers used an imprinting method for the preparation of leuko-
cyte samples. Following a rinsing and freeze-drying procedure, Sjvall and his
group obtained imprints of cells by applying pressure to a silver foil covering the
sample. Thus, they were able to detect phosphocholine- and cholesterol-derived
signals [125].
Finally, chemical fixation procedures allow a ToF-SIMS view into the cell. The
incubation of methanol/acetic acid into cells flushes away the lipidic membrane
components [126]. This approach was first introduced by Levi-Setti for dynamic
SIMS [22, 127]. In 2009, a combined analysis of a xenobiotic fluorophore inside a
cell by confocal laser scanning microscopy and (static) ToF-SIMS [79] was reported.
The choice of a specific preparation procedure is often based on the analytical
question. However, quite often it is also governed by the abilities and equipment of
the respective laboratories. Unfortunately, comparative studies on preparation pro-
cedures are not yet available. Those studies could improve our understanding of
how the different abilities of individual preparation procedures influence analytical
results. In particular, they would give valuable information on possible chemical
changes of the specimen during preparation.
58 D. Breitenstein et al.

3.4.3.4 Other Imaging Applications

A number of studies deal with model systems or real-world samples different than
those discussed in the previous sections. Among those are the analyses of pulp [128]
and paper [129], which are of interest for their respective industries. Also, wood-
containing artwork has been analyzed [130]. Other sample systems analyzed with
ToF-SIMS comprise biochips [131, 132] and micropatterned biomolecular surfaces
[133]. Protein adsorption on surfaces relevant for medical applications [134] or cell
culturing [135] has also been studied.

3.4.4 Depth Profiling

In ToF-SIMS, the term depth profiling is used for the in-depth probing of a sam-
ple without obtaining information on the lateral distribution of the analyte. Thus,
this approach is useful for well-defined layered samples. As most biological sam-
ples do not belong to this class, this application mode is rarely used in the biochemi-
cal field. To get an idea on the type of data usually obtained by depth profiling, the
results of the analysis of a layered polymer system are shown in Fig. 3.5 [136]. The
sample consists of five Irganox 3114 delta layers at depths of 50 nm, 100 nm,
200 nm, 300 nm, and 400 nm in an Irganox 1010 matrix. It has been produced and
provided by NPL and is described in detail elsewhere [137]. The profile was
acquired under sample rotation while sputtering with C60 and performing parallel
analysis with Bi3 cluster primary ions. Many different parameters (sputter ion spe-
cies and energy, sample temperature, etc.) determine the success of an organic depth
profile. To describe the details is beyond the scope of this chapter. An up-to-date
review was given by Mahoney [43].
Althoughas stateddepth profiling is not ideally suited for biochemical anal-
ysis, a limited number of publications dealing with it are available. Most of these
studies focus on the investigation of ToF-SIMSspecific properties. For example,
Sostarecz et al. [138] analyzed multiple LangmuirBlodgett film of barium arachi-
date by C60 depth profiling, finding the sputter yield to be higher by a factor of 100
compared to that of gallium. Cheng and Winograd depth-profiled a peptide-doped
trehalose film with a C60 source that was spin-coated on silicon [44, 139]. They were
able to detect intact peptide ions up to a mass of 500 u after sputtering. Although the
intensities of the secondary ions characteristic for the peptide decrease upon con-
tinuous C60 bombardment, these results suggest the possibility of 3D analysis of
biological samples.

3.4.5 3D Microarea Analysis

In recent years, the first examples for three-dimensional analyses of biological sam-
ples have been presented. Most studies show the general possibility for the detection
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 59

of organic molecules in deeper layers without being able to accurately represent the
expected distribution of molecular components in a cell, for example [140].
Breitenstein et al. [52] developed and published a protocol to provide the correct 3D
distribution of components in glutardialdehyde-fixated normal rat kidney (NRK)
cells (see Fig. 3.6). The spatial distributions of amino acids (fragments of proteins
in the nuclei) and phospholipids (cell walls) as detected by ToF-SIMS correspond
well to control experiments by confocal light microscopy, a well-established tech-
nique in the biological field after the ToF-SIMS data are corrected in 3D. There are
a number of other 3D analyses of biological samples: In one study, Vaidyanathan
et al. [122] show images of the surface as well as of a deeper layer of a bacteria
sample. Another example deals with the analysis of thyroid tumor cells [141].
There are certainly a number of aspects that can be criticized in all 3D analyses
of biological material with ToF-SIMS. To date, the scientific output is limited.
While it is possible to detect the phosphatidylcholine and amino acid distribution in
cells, other components were not detected or at least not shown in the publications.
In addition, recent studies on trehalose films indicate that the sputtering process
does not uniformly erode the samples [142]. As the representation of the data in all
studies is based on the estimation of uniform erosion, the results are most likely
displayed in a distorted way. Nevertheless, from a technological point of view, these
first results are very promising. A condensed overview of 3D molecular imaging is
given by Delcorte [143].

3.5 Improvement of Data Quality

In addition to further developments in instrumentation, two major strategies for

improving data acquisition and data evaluation are currently being considered: On
the one hand, the secondary ionization probability of ToF-SIMS is relatively low.
An enhancement of ionization would be beneficial for ToF-SIMS analysts. On the
other hand, the enormous wealth of dataeven with these low ionization probabili-
tiesis hampering an easy (manual) evaluation of the data. Therefore, statistical
methods are applied to facilitate data evaluation and interpretation.

3.5.1 Enhancement of Secondary Ionization Probability

Only approximately 1%and in many cases even lessof the ejected particles
upon primary ion impact are charged and thus available for mass-spectrometric
analysis. The increase in the secondary ionization probability is therefore a useful
strategy to obtain better results. Three approaches will be briefly discussed in this
section. However, the reader should keep in mind that these approaches are not
generally applicable and are therefore limited to special conditions and samples.
60 D. Breitenstein et al.

One option for improvement is given by the postionization of the ejected neutral
particles, which constitute the majority of the sputtered particles. Postionization can
be achieved either by electrons (e-beam or plasma) or, more promisingly, by pho-
tons (laser secondary neutral mass spectrometry, laser-SNMS). While laser-SNMS
is often able to selectively increase the ionization probability of certain species by
several orders of magnitude (e.g., one element or a certain class of molecules in a
resonant ionization scheme), it has not yet been possible to universally increase the
ionization probability for all species in parallel. Since a full discussion is beyond the
focus of this article, the reader is referred to some (rather old but still comprehen-
sive) book sections [144, 145] and a more recent research article [146] for further
information.
The following two strategies can be employed using standard ToF-SIMS instru-
mentation. An enhancement of the secondary ionization probability can be achieved
in several cases using matrices known from MALDI (matrix-enhanced SIMS,
ME-SIMS) [57, 73, 107, 116, 147]. Another approach to enhance the secondary
ionization probability is the use of metal cationization, where a metal ion combines
with a sputtered (neutral) molecule ([molecule + metal]+) (meta-SIMS) [48, 72, 115,
116, 148151]. This can be achieved with the use of thin gold or silver coatings on
the samples or by the preparation of an analyte as a (sub-) monolayer on a suitable
metal substrate. Whereas in meta-SIMS the cationization of a molecule by a sput-
tered metal ion leads to a higher efficiency, the effects of MALDI matrices are more
multifaceted [152]. Nevertheless, many of these effects can be directly or indirectly
allocated to an increased sputter desorption of surface material. Most of the ME-
and meta-SIMS studies were performed under monoatomic primary ion bombard-
ment. Therefore, the question of whether the enhancement is still detectable under
polyatomic bombardment is currently under investigation. The state of the art in
ME- and meta-SIMS was described by Delcorte in 2006 [152].

3.5.2 Statistical Evaluation

The extraordinary wealth of data collected in each ToF-SIMS experiment requires

elaborate evaluation routines. While in most cases data evaluation is still performed
manually and relies heavily on the users expertise, the use of multivariate statistical
methods is becoming more and more relevant. An idea about the different mathe-
matical approaches used for ToF-SIMS data evaluation can be gained by reading
the introductory articles by Graham et al. [153], Tyler [154, 155], Milillo and
Gardella [156], as well as Smentkowski et al. [157]. In the following, some exam-
ples for biochemical analytical questions, which were addressed using statistical
tools, are given.
Kulp et al. were able to differentiate cell lysates by principal component analysis
(PCA) [158]. Berman et al. could distinguish a number of monosaccharides from
their spectra using PCA [67]. May et al. performed PCA on DNA oligomers [65],
and Thompson et al. were able to discern bacilli by statistical methods [159].
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 61

Another interesting approach is the determination of conformational changes in

proteins. For example, in native proteins, the hydrophobic amino acids are more
commonly found in the interior of the protein than in denatured proteins. Thus, the
spectral intensities of characteristic ions of these amino acids give an indication of
the conformational state of a protein [160]. Nowadays, this kind of analysis can be
nicely addressed by PCA [55].
The statistical tools are useful in the hands of an expert who understands the
SIMS process and the mathematics behind the data evaluation, in order to validate
or generate hypotheses. However, it has to be pointed out that without a thorough
understanding, the tools might lead to misinterpretation of the data.

3.6 Specifications of TOF-SIMS

A scientist should always be aware of the physical and practical limits of an analysis
by ToF-SIMS when addressing a problem with this technique. Many of these limits
have been mentioned earlier in this chapter. Nevertheless, a condensed overview
might (have) save(d) time for the scan and skim reader. Most of the limitations men-
tioned are matters of further scientific and engineering research and development.
ToF-SIMS is a vacuum technique. Therefore, the sample has to sustain the vac-
uum environment inside the instrument. For many volatile samples, this can be
achieved by cooling. The sample size is usually limited by the size of the introduc-
tion chamber and the analysis chamber of the instrument. It is typically in the range
of several cm in length and width and up to some 10 mm in height. However, for
outgassing samples, the pumping rate of the vacuum system has to be considered,
too. The actual analysis area is routinely up to 500 500 m2. In macro scan mode
(when the stage with the sample is scanned under the primary ion beam and the
mass analyzer), the analysis area can be up to several cm2 in size. However, samples
have to be flat for this analysis mode. The mass range of ToF-SIMSas of all time-
of-flight mass-spectrometric techniquesis principally unlimited although the
desorption and ionization process limits the practical achievable mass range to
about 10,000 u [15]. The real strength of ToF-SIMS is in the detection of molecules
in the range of several 100 to 2,000 u. The detection limit for these kinds of species
is routinely in the femtomol to attomol per m2 range. That means, for example, that
2 10-20 mol on an area of 100 100 nm can be detected [28]. Some authors even
claim to be able to detect an amount of a substance in the zeptomol range (4 attomol
on a spot approximately 2 mm in diameter) [160]. Others have shown that proteins
on a surface can be detected in an amount of size 0.1 ng/cm2 [161]. Thus, it can be
estimated that detection limits in the picomol/m2 range can be routinely obtained.
ToF-SIMS is a surface-sensitive analysis technique. The information depth is
limited to the three uppermost monolayers [13], making ToF-SIMS useful for many
analytical problems. However, this fact also necessitatesespecially for biochemi-
cal samplesa contamination-free sample preparation. Quantification is difficult
with ToF-SIMS. In general, semiquantitative information can be gained upon
62 D. Breitenstein et al.

comparison of two similar samples or different positions on one sample. The mass
resolution [defined as the ratio of a mass m to the mass difference m of the next
neighboring peak that can be resolved (R = m/m)] is in the range of 10,000. This
means that a peak at mass 100.00 u can be distinguished from a peak at mass
100.01 u. The lateral resolution, which can be routinely obtained in imaging modes,
is in the order of a few hundred nanometers [105]. Meanwhile, unpublished studies
show better resolutions of down to 50 nm. Ultimately, the size of the collision cas-
cade limits the lateral resolution in SIMS to approximately 30 nm [14]. The vertical
resolution in depth profiling and 3D imaging was found to be some 10 nm in
organic samples [135, 162]. For well-defined inorganic samples (e.g., in the semi-
conductor industry), a depth resolution of less than 1 nm can be obtained [45].
Therefore, it is likely that the ultimate limit of the vertical resolution in organic
samples has not yet been reached.
It is rarely mentioned that mass resolution, lateral resolution, and analysis time
(which is often equivalent to sensitivity) are part of the so-called magic triangle in
ToF-SIMS: For technical reasons, only two of these three parameters can be opti-
mized to an acceptable value in one experiment.

3.7 Perspectives

The use of ToF-SIMS for biomolecular analysis has made tremendous progress in
the last 10 years. This was fueled to a great extent by technical developments, such
as the widespread introduction of polyatomic primary ion sources. Developments in
instruments are still continuing, focusing, on the one hand, on the design of primary
ion sources with new species and better specifications. In particular, the develop-
ment of gas cluster ion beam (GCIB) sources appears to be opening new perspec-
tives on the field of biomolecular analysis [163, 164]. On the other hand, the
instrumentation development is focused on the introduction of different analyzer
systems (e.g., MS/MS [165]). This work is accompanied by more sample-focused
research, including approaches to simplify cell tissue sample handling or treatments
to increase ionization efficiencies (see also Sect. 3.5.1). The use of more sophisti-
cated data evaluation routines should also be mentioned. Also promising is the
increasing number of instruments available that combine ToF-SIMS with other
surface-analytical techniques, making the immediate control of experimental results
by complementaryand sometimes more establishedtechniques possible.
Combination instruments with fluorescence microscopy [78], MALDI-ToF-MS [3,
113], AFM [166, 167], FTIR-microspectroscopy [112], and XPS [88] are currently
available; others will certainly follow.
Beside these experimental improvements, it is agreed that there is a discrepancy
between the already existing potentials of the technique and their exhaustion in the
field of life sciences [168]. Certainly, biomolecular ToF-SIMS still suffers from the
relatively low number of scientists applying this technique to biological samples.
This is to some extent caused by the low availability of modern instruments, which
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 63

is likely related to their high cost. In addition, highly skilled personnel are needed
to adequately prepare the sample and perform the ToF-SIMS measurements as well
as the data workup and interpretation. In recent years, cooperation of experts from
each particular field has become more common. It is likely that this type of coopera-
tion will promote the field in upcoming years, as the different backgrounds offer a
high synergy for progress.
The judgment of the potentials of ToF-SIMS in the area of life science applica-
tions is diverse: On the one hand, there are skeptics who doubt the ability of ToF-
SIMS to become a standard application for answering complex questions in the life
sciences. On the other hand, very optimistic views can be found, such as that in the
paper entitled The magic of cluster SIMS in Analytical Chemistry [31]: Every
once in a while, a breakthrough propels a mature field into new dimensionsjust as
the discoveries of MALDI and ESI opened MS to biologists and, incidentally,
racked up Nobel Prizes for their inventors. This sort of metamorphosis is currently
under way in bioimaging because of the remarkable properties of cluster ion beam
sources being used with secondary ion MS (SIMS). It is to hope that such ambi-
tious goals rather propel than thwart the progress in this exciting but difficult field.

3.8 Acknowledgment

The authors thank Dr. Reinhard Kersting (from Tascon GmbH, Muenster) for
additional proofreading. The samples shown in this review were prepared by
Dr. Joachim Wegener, Dr. Christina Rommel, and Michael Seifert (all from the
University of Mnster in Germany). Writing this review has been financially sup-
ported by the European Union Framework Program 6 under grant number FP6-
513698 (Toxdrop-Project).

References

1. Pacholski ML, Winograd N. Imaging with mass spectrometry. Chem Rev. 1999;99:29773005.
2. Belu AM, Graham DJ, Castner DG. Time-of-flight secondary ion mass spectrometry: tech-
niques and applications for the characterization of biomaterial surfaces. Biomaterials.
2003;24(21):363553.
3. Brunelle A, Touboul D, Laprevote O. Biological tissue imaging with time-of-flight secondary
ion mass spectrometry. J Mass Spectrom. 2005;40:98599.
4. McDonnell LA, Heeren RM. Imaging mass spectrometry. Mass Spectrom Rev. 2007;26(4):
60643.
5. Breitenstein D, et al. The chemical composition of animal cells and their intracellular com-
partments reconstructed from 3D mass spectrometry. Angew Chem Int Ed Engl. 2007;46(28):
53325.
6. Benninghoven A. The history of static SIMS: a personal perspective. In: Vickerman JC,
editor. ToF-SIMS. Surface analysis by mass spectrometry. Huddersfield: IM Publications;
2001. p. 4174.
64 D. Breitenstein et al.

7. Guerquin-Kern J-L, et al. Progress in analytical imaging of the cell by dynamic secondary ion
mass spectrometry (SIMS microscopy). Biochim Biophys Acta. 2005;1724:22838.
8. Garrison BJ. Atoms, clusters and photons: energetic probes for mass spectrometry. Appl Surf
Sci. 2006;252:640912.
9. Sigmund P. Theory of sputtering. I. Sputtering yield of amorphous and polycrystalline tar-
gets. Phys Rev. 1969;184:38.
10. Sigmund P. Erratum: theory of sputtering. I. Sputtering yield of amorphous and polycrystal-
line targets. Phys Rev. 1969;187:768.
11. Sigmund P. Collision theory of displacement damage, ion ranges, and sputtering. Revue
Roumaine de Physique. 1972;1106(17):823.
12. Thompson MW. The energy spectrum of ejected atoms during the high energy sputtering of
gold. Philos Mag. 1968;414(18):377.
13. Hagenhoff B, Rading D. Ion beam techniques: surface mass spectrometry. In: Rivire JC,
Myhra S, editors. Handbook of surface and interface analysis. New York/Basel/Hong Kong:
Marcel Dekker; 1998. p. 209346.
14. Vickerman JC. TOF-SIMS: an overview. In: Vickerman JC, Briggs D, editors. ToF-SIMS
surface analysis by mass spectrometry. Manchester/Chichester: IM Publications; 2001.
15. Benninghoven A, Rdenauer FG, Werner HW. Secondary ion mass spectrometry. Basic con-
cepts, instrumental aspects, applications and trends. New York: Wiley; 1987. p. 761949.
16. Hagenhoff B. High resolution surface analysis by TOF-SIMS. Mikrochim Acta. 2000;132:
25971.
17. Urbassek HM. Status of cascade theory. In: Vickerman JC, editor. ToF-SIMS. Suface analysis
by mass spectrometry. Huddersfield: IM Publications; 2001. p. 13960.
18. McLafferty FW, Turecek F. Interpretation of mass spectra. Mill Valley: University Science
Books; 1993.
19. Chandra S. 3D subcellular imaging in cryogenically prepared single cells. Appl Surf Sci.
2004;231232:4679.
20. Galle P, et al. Subcellular localization of aluminium and indium in the rat kidney. Appl Surf
Sci. 2004;231232:4758.
21. Chandra S, Smith DR, Morrison GH. Subcellular imaging by dynamic SIMS ion microscopy.
Anal Chem. 2000;72(3):104A14.
22. Levi-Setti R, et al. Ion microprobe imaging of 44Ca-labeled mammalian chromosomes. Appl
Surf Sci. 2004;231232:47984.
23. Gillen G, Roberson SV. Preliminary evaluation of an SF-(5) + polyatomic primary ion beam
for analysis of organic thin films by secondary ion mass spectrometry. Rapid Commun Mass
Spectrom. 1998;12:130312.
24. Ktter F, Benninghoven A. Secondary ion emission from polymer surfaces under Ar+,
Xe + and SF-(5) + ion bombardment. Appl Surf Sci. 1998;133:4757.
25. Weibel DE, Lockyer N, Vickerman JC. C60 cluster ion bombardment of organic surfaces.
Appl Surf Sci. 2004;231232:14652.
26. Wong SCC, et al. Development of a C60+ ion gun for static SIMS and chemical imaging. Appl
Surf Sci. 2003;203204:21922.
27. Weibel D, et al. A C60 primary ion beam system for time of flight secondary ion mass spec-
trometry: its development and secondary ion yield characteristics. Anal Chem. 2003;75(7):
175464.
28. Davies N, et al. Development and experimental application of a gold liquid metal ion source.
Appl Surf Sci. 2003;203204:2237.
29. Kollmer F. Cluster primary ion bombardment of organic materials. Appl Surf Sci. 2004;
231232:1538.
30. Russo MF, Wojciechowski IA, Garrison BJ. Sputtering of amorphous ice induced by C60 and
Au3 clusters. Appl Surf Sci. 2006;252:64235.
31. Winograd N. The magic of cluster SIMS. Anal Chem. 2005;77(7):143A9A. doi:10.1021/
ac053355f.
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 65

32. Wucher A. Molecular secondary ion formation under cluster bombardment: a fundamental
review. Appl Surf Sci. 2006;252:64829.
33. Benguerba M, et al. Impact of slow gold cluster on various solids: non-linear effects in
secondary emission. Nucl Instrum Meth B. 1991;62:822.
34. Kersting R, et al. Influence of primary ion bombardment conditions on the emission of
molecular secondary ions. Appl Surf Sci. 2004;231/232:2614.
35. Castner D. Surface science: view from the edge. Nature. 2003;422:12930.
36. Ktter F, Benninghoven A. Secondary ion emission from polymer surfaces under Ar+, Xe+
and SF5+ ion bombardment. Appl Surf Sci. 1998;133:4757.
37. Stapel D, Brox O, Benninghoven A. Secondary ion emission from arachoidic acid LB-layers
under Ar+, Xe+, Ga+ and SF-(5) + primary ion bombardment. Appl Surf Sci. 1999;140:
15667.
38. Stapel D, Benninghoven A. Application of atomic and molecular primary ions for TOF-
SIMS analysis of additive containing polymer surfaces. Appl Surf Sci. 2001;174:26170.
39. Appelhans AD, Delmore JE. Comparison of polyatomic and atomic primary beams for sec-
ondary ion mass spectrometry of organics. Anal Chem. 1989;61:108793.
40. Conlan XA, Lockyer NP, Vickerman JC. Is proton cationization promoted by polyatomic
primary ion bombardment during time-of-flight secondary ion mass spectrometry analysis of
frozen aqueous solutions? Rapid Commun Mass Spectrom. 2006;20:132734.
41. Wucher A, et al. Molecular depth profiling in ice matrices using C-60 projectiles. Appl Surf
Sci. 2004;231232:6871.
42. Cheng J, et al. Direct comparison of Au3+ and C60+ cluster projectiles in SIMS molecular
depth profiling. J Am Soc Mass Spectrom. 2007;18:40612.
43. Mahoney CM. Cluster secondary ion mass spectrometry of polymers and related materials.
Mass Spectrom Rev. 2009;29(2):24793.
44. Cheng J, Wucher A, Winograd N. Molecular depth profiling with cluster ion beams. J Phys
Chem B. 2006;110:832936.
45. Hill R. Primary ion systems. In: Vickerman JC, editor. ToF-SIMS. Surface mass spectrome-
try. Huddersfield: IM Publications; 2001. p. 95112.
46. Niehuis E, Grehl T. Dual beam depth profiling. In: Vickerman JC, Briggs D, editors. TOF-
SIMS: surface analysis by mass spectrometry. Charlton: IM Publications; 2001. p. 75378.
47. Niehuis E, et al. Design and performance of a reflectron based time-of-flight secondary ion
mass spectrometer with electrodynamic primary ion mass separation. J Vac Sci Technol A.
1987;A5:12436.
48. Nygren H, et al. Localization of cholesterol, phosphocholine and galactosylceramide in rat
cerebellar cortex with imaging TOF-SIMS equipped with a bismuth cluster ion source.
Biochim Biophys Acta. 2005;1737(23):10210.
49. Schueler BW. Time-of-flight mass analysers. In: Vickerman JC, editor. ToF-SIMS. Surface
analysis by mass spectrometry. Huddersfield: IM Publications; 2001. p. 7594.
50. Hagenhoff B, et al. Time-of-flight secondary ion mass-spectrometry of insulators with pulsed
charge compensation by low-energy electrons. J Vac Sci Technol A. 1989;10(5):305664.
51. Gilmore IS, Seah MP. Electron flood gun damage in the analysis of polymers and organics in
time-of-flight SIMS. Appl Surf Sci. 2002;187:89100.
52. Breitenstein D, et al. The chemical composition of animal cells and their intracellular com-
partments reconstructed from 3D mass spectrometry. Angew Chem Int Edit. 2006. Submitted
for publication.
53. Mantus DS, et al. Static secondary ion mass spectrometry of adsorbed proteins. Anal Chem.
1993;65:14318.
54. McArthur SL, et al. Methods for generating protein molecular ions in ToF-SIMS. Langmuir.
2004;20:37049.
55. Michel R, et al. Influence of PEG architecture on protein adsorption and conformation.
Langmuir. 2005;21(26):1232732. doi:10.1021/la051726h. Research article.
56. Sole-Domenech S, et al. Analysis of opioid and amyloid peptides using time-of-flight sec-
ondary ion mass spectrometry. Anal Chem. 2010;82(5):196474.
66 D. Breitenstein et al.

57. Wittmaack K, et al. Time-of-flight secondary ion mass spectrometry of matrix-diluted oligo-
and polypeptides bombarded with slow and fast projectiles: positive and negative matrix and
analyte ion yields, background signals, and sample aging. J Am Soc Mass Spectrom. 2000;11:
55363.
58. Wu KJ, Odom RW. Biological material analysis by matrix-enhanced SIMS. In: Gillen G
et al., editors. Secondary ion mass spectrometry XI. Chichester: Wiley; 1998. p. 5258.
59. Wagner MS, Horbett TA, Castner DG. Characterizing multicomponent adsorbed protein
films using electron spectroscopy for chemical analysis, time-of-flight secondary ion mass
spectrometry, and radiolabeling: capabilities and limitations. Biomaterials. 2003;24(11):
1897908.
60. Jabs HU, et al. High performance liquid chromatography and time-of-flight secondary ion
mass spectrometry: a new dimension in structural analysis of apolipoproteins. J Lipid Res.
1986;27(6):61321.
61. von Eckardstein A, et al. Site-specific methionine sulfoxide formation is the structural basis
of chromatographic heterogeneity of apolipoproteins A-I, C-II, and C-III. J Lipid Res.
1991;32(9):146576.
62. von Eckardstein A, et al. Apolipoprotein A-I variants. Naturally occurring substitutions of
proline residues affect plasma concentration of apolipoprotein A-I. J Clin Invest. 1989;84(6):
172230.
63. Cheran L-E, Vukovich D, Thompson M. Imaging TOF-SIMS analysis of oligonucleotide
microarrays. Analyst. 2003;128:1269.
64. Lee C-Y, et al. Evidence of impurities in thiolated single-stranded DNA oligomers and their
effect on DNA self-assembly on gold. Langmuir. 2005;21:513441.
65. May CJ, Canavan HE, Castner DG. Quantitative X-ray photoelectron spectroscopy and time-
of-flight secondary ion mass spectrometry characterization of the components in DNA. Anal
Chem. 2004;76(4):111422.
66. Arlinghaus HF, et al. DNA sequencing with ToF-SIMS. Surf Interface Anal. 2002;33:359.
67. Berman ES, et al. Distinguishing monosaccharide stereo- and structural isomers with TOF-
SIMS and multivariate statistical analysis. Anal Chem. 2006;78(18):6497503.
68. Roddy TP, et al. Imaging of freeze-fractured cells with in situ fluorescence and time-of-flight
secondary ion mass spectrometry. Anal Chem. 2002;74(16):40119.
69. Leufgen KM, et al. Imaging time-of-flight secondary ion mass spectrometry allows visualiza-
tion and analysis of coexisting phases in LangmuirBlodgett films. Langmuir. 1996;12:
170811.
70. Roddy TP, et al. Identification of cellular sections with imaging mass spectrometry following
freeze fracture. Anal Chem. 2002;74(16):40206.
71. Malmberg P, et al. Analysis of bone minerals by time-of-flight secondary ion mass spectrom-
etry: a comparative study using monoatomic and cluster ions sources. Rapid Commun Mass
Spectrom. 2007;21(5):7459.
72. Adriaensen L, Vangaever F, Gijbels R. Metal-assisted secondary ion mass spectrometry:
influence of Ag and Au deposition on molecular ion yields. Anal Chem. 2004;76(22):
677785.
73. Adriaensen L, et al. Matrix-enhanced secondary ion mass spectrometry: the influence of
MALDI matrices on molecular ion yields of thin organic films. Rapid Commun Mass
Spectrom. 2005;19:101724.
74. Kim YP, et al. Activity-based assay of matrix metalloproteinase on nonbiofouling surfaces
using time-of-flight secondary ion mass spectrometry. Anal Chem. 2008;80(13):5094102.
75. Breitenstein D, Vonhren B, Reihs K. Semiquantitative analysis of self assembled monolay-
ers by LEIS as well as SIMS. Submitted.
76. Hull JR, Tamura GS, Castner DG. Interactions of the streptococcal C5a peptidase with human
fibronectin. Acta Biomater. 2008;4(3):50413.
77. Kim Y-P, et al. Protein quantification on dendrimer-activated surfaces by using time-of-flight
secondary ion mass spectrometry and principal component regression. Appl Surf Sci.
2008;255(4):11102.
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 67

78. Schnieders A, Mllers R, Benninghoven A. Molecular secondary particle emission from

molecular overlayers under 10 keV Ar+ primary ion bombardment. Surf Sci. 2001;471:17084.
79. Breitenstein D, et al. The chemical composition of animal cells reconstructed from 2D and
3D ToF-SIMS analysis. Appl Surf Sci. 2008;46(28):53325.
80. Ulman A. Ultrathin organic films. San Diego: Academic; 1994.
81. Bourdos N, et al. Analysis of lung surfactant model systems with time-of-flight secondary ion
mass spectrometry. Biophys J. 2000;79(1):35769.
82. Bourdos N, et al. Imaging of domain structures in a one-component lipid monolayer by time-
of-flight secondary ion mass spectrometry. Langmuir. 2000;16:14814.
83. Biesinger MC, et al. Imaging lipid distributions in model monolayers by ToF-SIMS with
selectively deuterated components and principal component analysis. Appl Surf Sci.
2006;252:695765.
84. Biesinger MC, et al. Principal component analysis of TOF-SIMS images of organic monolay-
ers. Anal Chem. 2002;74(22):57116.
85. Sostarecz AG, et al. Influence of molecular environment on the analysis of phospholipids by
time-of-flight secondary ion mass spectrometry. Langmuir. 2004;20(12):492632.
86. Breitenstein D, et al. Lipid specificity of surfactant protein B studied by time-of-flight sec-
ondary ion mass spectrometry. Biophys J. 2006;91(4):134756.
87. Harbottle RR, et al. Molecular organization revealed by time-of-flight secondary ion mass
spectrometry of a clinically used extracted pulmonary surfactant. Langmuir. 2003;19:
3698704.
88. Sostarecz AG, et al. Phosphatidylethanolamine-induced cholesterol domains chemically
identified with mass spectrometric imaging. J Am Chem Soc. 2004;126(43):138823.
89. Hansson M, et al. Iodine content and distribution in extratumoral and tumor thyroid tissue
analyzed with X-ray fluorescence and time-of-flight secondary ion mass spectrometry.
Thyroid. 2008;18(11):121520.
90. Seifert M, et al. Solubility vs. electrostatics: what determines the lipid/protein interaction in
the lung surfactant. Biophys J. 2007;93(4):1192203.
91. Keating E, et al. Effect of cholesterol on the biophysical and physiological properties of a
clinical pulmonary surfactant. Biophys J. 2007;93(4):1391401.
92. Simons K, van Meer G. Lipid sorting in epithelial cells. Biochemistry. 1988;27(17):6197202.
93. McQuaw CM, et al. Investigating lipid interactions and the process of raft formation in cel-
lular membranes using ToF-SIMS. Appl Surf Sci. 2006;252:67168.
94. Munro S. Lipid rafts: elusive or illusive? Cell. 2003;115(4):37788.
95. Wolf KV, Cole DA, Bernasek SL. High-resolution TOF-SIMS study of varying chain length
self-assembled monolayer surfaces. Anal Chem. 2002;74(19):500916.
96. Wong SCC, Lockyer NP, Vickerman JC. Mechanisms of secondary ion emission from self-
assembled monolayers and multilayers. Surf Interface Anal. 2005;37:72130.
97. Francis JT, et al. ToF-SIMS investigation of octadecylphosphonic acid monolayers on a mica
substrate. Langmuir. 2006;22:924450.
98. Schrder M, et al. Influence of primary ion species on the secondary cluster ion emission
process from SAMs of hexadecanethiol on gold. Appl Surf Sci. 2006;252(19):65669.
99. Rading D, Kersting R, Benninghoven A. Secondary ion emission from molecular overlayers:
Thiols on gold. Proc SIMS XI Conf. 1997;11:4558.
100. Francis JT, et al. ToF-SIMS investigation of octadecylphosphonic acid monolayers on a mica
substrate. Langmuir. 2006;22(22):924450.
101. Prinz C, et al. Structural effects in the analysis of supported lipid bilayers by time-of-flight
secondary ion mass spectrometry. Langmuir. 2007;23(15):803541.
102. Richter K, et al. Localization of fatty acids with selective chain length by imaging time-of-
flight secondary ion mass spectrometry. Microsc Res Tech. 2007;70(7):6407.
103. Mller J, et al. Introduction of a cryosectioning-ToF-SIMS instrument for analysis of non-
dehydrated biological samples. Appl Surf Sci. 2006;252(19):670911.
104. Touboul D, et al. Tissue molecular ion imaging by gold cluster ion bombardment. Anal
Chem. 2004;76(6):15509.
68 D. Breitenstein et al.

105. Sjvall P, Lausmaa J, Johansson B. Mass spectrometric imaging of lipids in brain tissue. Anal
Chem. 2004;76(15):42718.
106. Touboul D, et al. Improvement of biological time-of-flight secondary ion mass spectrometry
imaging with bismuth cluster ion source. J Am Soc Mass Spectrom. 2005;16:160818.
107. McDonell LA, et al. Subcellular imaging mass spectrometry of brain tissue. J Mass Spectrom.
2005;1005(40):1608.
108. Borner K, et al. Distribution of cholesterol and galactosylceramide in rat cerebellar white
matter. Biochim Biophys Acta. 2006;1761(3):33544.
109. Nygren H, et al. Imaging TOF-SIMS of rat kidney prepared by high-pressure freezing.
Microsc Res Tech. 2005;68(6):32934.
110. Nygren H, et al. Bioimaging TOF-SIMS: localization of cholesterol in rat kidney sections.
FEBS Lett. 2004;566(13):2913.
111. Touboul D, et al. Lipid imaging by gold cluster time-of-flight secondary ion mass spectrom-
etry: application to Duchenne muscular dystrophy. J Lipid Res. 2005;46(7):138895.
112. Malmberg P, et al. Imaging of lipids in human adipose tissue by cluster ion TOF-SIMS.
Microsc Res Technol. 2007;70(9):82835. doi:10.1002/jemt.20481.
113. Le Naour F, et al. Chemical imaging on liver steatosis using synchrotron infrared and ToF-
SIMS microspectroscopies. PLoS One. 2009;4(10):e7408.
114. Monroe EB, et al. SIMS and MALDI MS imaging of the spinal cord. Proteomics.
2008;8(18):374654.
115. Nygren H, Malmberg P. Silver deposition on freeze-dried cells allows subcellular localization
of cholesterol with imaging TOF-SIMS. J Microsc. 2004;215(Pt 2):15661.
116. Altelaar AF, et al. Gold-enhanced biomolecular surface imaging of cells and tissue by SIMS
and MALDI mass spectrometry. Anal Chem. 2006;78(3):73442.
117. Berman ES, et al. Preparation of single cells for imaging/profiling mass spectrometry. J Am
Soc Mass Spectrom. 2008;19(8):12306.
118. Fartmann M, et al. Characterization of cell cultures with ToF-SIMS and laser-SNMS. Surf
Interface Anal. 2002;34:636.
119. Wittig A, et al. Preparation of cells cultured on silicon wafers for mass spectrometry analysis.
Microsc Res Tech. 2005;66(5):24858.
120. Ostrowski SG, et al. Mass spectrometric imaging of highly curved membranes during
Tetrahymena mating. Science. 2004;305(5680):713.
121. Piehowski PD, et al. Freeze-etching and vapor matrix deposition for ToF-SIMS imaging of
single cells. Langmuir. 2008;24(15):790611.
122. Vaidyanathan S, et al. Subsurface biomolecular imaging of Streptomyces coelicolor using
secondary ion mass spectrometry. Anal Chem. 2008;80(6):194251.
123. Parry S, Winograd N. High-resolution TOF-SIMS imaging of eukaryotic cells preserved in a
trehalose matrix. Anal Chem. 2005;77(24):79507.
124. Monroe EB, et al. Vitamin E imaging and localization in the neuronal membrane. J Am Chem
Soc. 2005;127(35):121523.
125. Sjovall P, et al. Imaging of membrane lipids in single cells by imprint-imaging time-of-flight
secondary ion mass spectrometry. Anal Chem. 2003;75(14):342934.
126. Hoetelmans RW, et al. Effects of acetone, methanol, or paraformaldehyde on cellular struc-
ture, visualized by reflection contrast microscopy and transmission and scanning electron
microscopy. Appl Immunohistochem Mol Morphol. 2001;9(4):34651.
127. Levi-Setti R, Gavrilow KL, Neilly ME. Cations in mammalian cells and chromosomes: sam-
ple preparation protocols affect elemental abundances by SIMS. Appl Surf Sci.
2005;252:67659.
128. Fardim P, et al. Extractives on fiber surfaces investigated by XPS, ToF-SIMS and AFM.
Colloid Surface A. 2005;255:91103.
129. Fardim P, Holmbom B. ToF-SIMS imaging: a valuable chemical microscopy technique for
paper and paper coatings. Appl Surf Sci. 2005;249:393407.
3 Biomolecular Analysis by Time-of-Flight Secondary Ion Mass 69

130. Mazel V, et al. Chemical imaging techniques for the analysis of complex mixtures: new appli-
cation to the characterization of ritual matters on African wooden statuettes. Anal Chim Acta.
2006;570(1):3440.
131. Aoyagi S, Kudo M. Effective monitoring of protein reaction on glass plate surfaces by
TOF-SIMS. Biosens Bioelectron. 2005;20(8):162630.
132. Hellweg S, et al. Mass spectrometric characterization of DNA microarrays as a function of
primary ion species. Appl Surf Sci. 2006;252:67425.
133. Belu AM, et al. Enhanced TOF-SIMS imaging of a micropatterned protein by stable isotope
protein labeling. Anal Chem. 2001;73(2):14350.
134. Aoyagi S, et al. TOF-SIMS imaging of protein adsorption on dialysis membrane by means of
information entropy. Surf Sci Nanotechnol. 2003;1:6771.
135. Canavan HE, et al. Surface characterization of the extracellular matrix remaining after cell
detachment from a thermoresponsive polymer. Langmuir. 2005;21(5):194955.
136. Rading D, et al. Dual beam depth profiling of organic materials: variations of analysis and
sputter beam conditions. Surf Interface Anal. 2010. doi:10.1002/sia.3422.
137. Shard AG, et al. Measurement of sputtering yields and damage in C60 SIMS depth profiling
of model organic materials. Surf Interface Anal. 2007;39:2948.
138. Sostarecz AG, et al. Depth profiling of LangmuirBlodgett films with a buckminsterfullerene
probe. Anal Chem. 2004;76(22):66518.
139. Cheng J, Winograd N. Depth profiling of peptide films with TOF-SIMS and a C60 probe. Anal
Chem. 2005;77:36519.
140. Fletcher JS, et al. TOF-SIMS 3D biomolecular imaging of Xenopus laevis oocytes using
buckminsterfullerene (C60) primary ions. Anal Chem. 2007;79:2199206.
141. Nygren H, et al. Bioimaging TOF-SIMS: high resolution 3D imaging of single cells. Microsc
Res Tech. 2007;70(11):96974.
142. Wucher A, Cheng J, Winograd N. Protocols for three-dimensional molecular imaging using
mass spectrometry. Anal Chem. 2007;79(15):552939. doi:10.1021/ac070692a.
143. Delcorte A. On the road to high-resolution 3D molecular imaging. Appl Surf Sci. 2008;255(4):
9548.
144. Arlinghaus HF. Laser-SNMS. In: Bubert H, Jenett H, editors. Surface and thin film analysis.
Principles, instrumentation, applications. Weinheim: Wiley-VCH; 2002. p. 1329.
145. Wucher A. Laser post-ionisation: fundamentals. In: Vickerman JC, editor. ToF-SIMSsurface
analysis by mass spectrometry. Manchester/Chichester: IM Publications; 2001. p. 34774.
146. Arlinghaus HF, et al. Subcellular imaging of cell cultures and tissue for boron localization
with laser-SNMS. Surf Interface Anal. 2004;36:698701.
147. Altelaar AF, et al. Direct molecular imaging of Lymnaea stagnalis nervous tissue at subcel-
lular spatial resolution by mass spectrometry. Anal Chem. 2005;77(3):73541.
148. Nygren H, Johansson BR, Malmberg P. Bioimaging TOF-SIMS of tissues by gold ion bom-
bardment of a silver-coated thin section. Microsc Res Tech. 2004;65(6):2826.
149. Delcorte A, Medard N, Bertrand P. Organic secondary ion mass spectrometry: sensitivity
enhancement by gold deposition. Anal Chem. 2002;74(19):495568.
150. Delcorte A, et al. Sample metallization for performance improvement in desorption/ioniza-
tion of kilodalton molecules: quantitative evaluation, imaging secondary ion MS, and laser
ablation. Anal Chem. 2003;75(24):687585.
151. Kim YP, et al. Gold nanoparticle-enhanced secondary ion mass spectrometry imaging of
peptides on self-assembled monolayers. Anal Chem. 2006;78(6):191320.
152. Delcorte A. Matrix-enhanced secondary ion mass spectrometry: the alchemists solution?
Appl Surf Sci. 2006;252(19):65827.
153. Graham DJ, Wagner MS, Castner DG. Information from complexity: challenges of TOF-
SIMS data interpretation. Appl Surf Sci. 2006;252(19):68608.
154. Tyler BJ. Multivariate statistical image processing for molecular specific imaging in organic
and bio-systems. Appl Surf Sci. 2006;252(19):687582.
70 D. Breitenstein et al.

155. Tyler BJ, Rayal G, Castner DG. Multivariate analysis strategies for processing ToF-SIMS
images of biomaterials. Biomaterials. 2007;28(15):241223.
156. Milillo TM, Gardella JA. Spatial statistics and interpolation methods for TOF SIMS imaging.
Appl Surf Sci. 2006;252(19):688390.
157. Smentkowski VS, et al. Multivariate statistical analysis of concatenated time-of-flight
secondary ion mass spectrometry spectral images. Complete description of the sample with
one analysis. Anal Chem. 2005;77(5):15306.
158. Kulp KS, et al. Chemical and biological differentiation of three human breast cancer cell
types using time-of-flight secondary ion mass spectrometry. Anal Chem. 2006;78(11):
36518.
159. Thompson CE, et al. ToF-SIMS studies of Bacillus using multivariate analysis with possible
identification and taxonomic applications. Appl Surf Sci. 2006;252(19):671922.
160. Tidwell CD, et al. Static time-of-flight secondary ion mass spectrometry and X-ray photo-
electron spectroscopy characterization of adsorbed albumin and fibronectin films. Surf
Interface Anal. 2001;31:72433.
161. Quong JN, et al. Molecule-specific imaging analysis of carcinogens in breast cancer cells
using time-of-flight secondary ion mass spectrometry. Appl Surf Sci. 2004;231232:4247.
162. Wagner MS, et al. Limits of detection for time of flight secondary ion mass spectrometry
(ToF-SIMS) and X-ray photoelectron spectroscopy (XPS): detection of low amounts of
adsorbed protein. J Biomater Sci Polym Ed. 2002;13(4):40728.
163. Wagner MS. Molecular depth profiling of multilayer polymer films using time-of-flight sec-
ondary ion mass spectrometry. Anal Chem. 2005;77:91122.
164. Yamada I, et al. Materials processing by gas cluster ion beams. Mat Sci Eng: R-Reports Rev
J. 2001;34(6):23195.
165. Ninomiya S, et al. Precise and fast secondary ion mass spectrometry depth profiling of poly-
mer materials with large Ar cluster ion beams. Rapid Commun Mass Spectrom. 2009;23:
16016.
166. Carado A, et al. C60 secondary ion mass spectrometry with a hybrid-quadrupole orthogonal
time-of-flight mass spectrometer. Anal Chem. 2008;80(21):79219.
167. Popov J, et al. Chemical mapping of ceramide distribution in sphingomyelin-rich domains in
monolayers. Langmuir. 2008;24(23):135028.
168. Heeren RMA, et al. Why dont biologists use SIMS? A critical evaluation of imaging MS.
Appl Surf Sci. 2007;252(19):682735.
169. Breitenstein D, et al. The chemical composition of animal cells and their intracellular com-
partments reconstructed from 3D mass spectrometry. Angew Chem Int Edit. 2007. Published
online.
Chapter 4
Cluster Secondary Ion Mass Spectrometry

Joseph Kozole and Nicholas Winograd

Abstract In principle, secondary ion mass spectrometry (SIMS) molecule-specific

imaging has vast implications in biological research where submicrometer spatial
resolution, uppermost surface layer sensitivity, and chemically unmodified sample
preparation are essential. Yet SIMS imaging using atomic projectiles has been
rather ineffective when applied to biological materials. The common pitfalls expe-
rienced during these analyses include low secondary ion yields, extensive fragmen-
tation, restricted mass ranges, and the accumulation of significant physical and
chemical damage after sample erosion beyond 1% of the surface molecules.
Collectively, these limitations considerably reduce the amount of material available
for detection and result in inadequate sensitivity for most applications. In response,
polyatomic (cluster) ions have been introduced as an alternate imaging projectile.
Cluster ion bombardment has been observed to enhance secondary ion yields,
extend the spectral mass range, and decrease the incidence of physical and chemical
damage during sample erosion. The projectiles are expected to considerably increase
the number of molecules available for analysis and to significantly improve the
overall sensitivity. Hence, the objectives of this chapter are to describe the unique
physical basis for the improvements observed during polyatomic bombardment and
to identify the emerging biological applications made practical by the introduction
of cluster projectiles to SIMS.

J. Kozole
Department of Chemistry, Penn State University, 104 Chemistry Building,
University Park, Philadelphia, PA 16802, USA
DuPont, Philadelphia, PA, USA
N. Winograd (*)
Department of Chemistry, Penn State University, 104 Chemistry Building,
University Park, Philadelphia, PA 16802, USA
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 71

DOI 10.1007/978-3-319-01360-2_4, Springer International Publishing Switzerland 2014
72 J. Kozole and N. Winograd

4.1 Introduction

Energetic (keV) atomic projectiles were first employed to desorb intact molecules
for mass-spectrometric (MS) detection nearly 40 years ago [1]. However, polyatomic
ions were not identified as potentially valuable projectiles until 20 years later [2].
Appelhans etal. observed that when compared to the atomic Cs+, the cluster SF60
increases secondary ion efficiency and decreases the occurrence of sample damage
during SIMS molecule analysis. Similar results were achieved using aromatic
hydrocarbons, massive glycerol clusters, inorganic complexes, and SF5+ as primary
sources shortly thereafter [36]. Despite the initial success, polyatomic projectiles
were not widely adopted by the SIMS community. The reluctance was credited to a
number of device performance issues, which included low beam currents, inade-
quate beam focusing, and poor source lifetimes [6, 7]. Furthermore, the amount of
evidence identifying polyatomic projectiles as better-quality sources for SIMS
experiments was not overwhelming [6]. Consequently, cluster ion sources were
considered too unreliable for implementation to everyday SIMS applications, and
the widespread use of well-established atomic projectiles continued.
The mainstream acceptance of polyatomic projectiles to SIMS did not occur
until 10 years after the preliminary observations of Appelhans etal. The break-
through was initiated by the development of commercially available liquid metal
ion (LMIG) sources to produce Au3+ and Bi3+ and by the marketing of improved
gas-phase ion sources to produce C60+ [810]. LMIG technology uses a heated,
field-emission tip coated in a eutectic metal alloy (AuGe for Aun+ or Bi for Bin+) to
extract a mixture of metal cluster ions. The ions are mass-selected and electrostati-
cally aligned and focused to generate a bright, spatially defined metal cluster ion
beam. On the other hand, gas-phase ion sources use conventional electron impact
strategies to ionize vapor-phase C60. The ions are extracted, mass-filtered, and focused
using a sophisticated series of beam-minimizing apertures and electrostatic lenses to
obtain an intense, laterally defined C60 cluster ion beam. Regardless of scheme, poly-
atomic ion beam technology has successfully developed high- performance ion
sources characterized by 1-nA maximum beam currents, ~100200-nm optimal
beam sizes, and 1,000-h source lifetimes. Consequent to the previously observed
enhancements and the aforementioned technical advances, Au3+, Bi3+, and C60+ clus-
ter projectiles were rapidly adapted to SIMS instrumentation as potential successors
to atomic counterparts.
As the number of SIMS experiments involving polyatomic projectiles increased,
so did the number of observations regarding cluster ion bombardment [6, 7]. The
most recent observations have led to the identification of several important proper-
ties influencing the application of cluster ions to SIMS. Included among the proper-
ties are increased neutral and secondary ion yields and decreased physical and
chemical sample damage. Together, the characteristics increase the number of mol-
ecules available for SIMS analysis and improve the overall sensitivity of the imag-
ing modality. Accordingly, the objectives of this chapter are to introduce the unique
mechanism for desorption during polyatomic bombardment, to identify the special
4 Cluster Secondary Ion Mass Spectrometry 73

properties of cluster ions in SIMS, and to discuss the implications of these properties
to sensitivity, lateral resolution, and depth resolution during the imaging modality.
Using these properties, we will show how the improved performance of cluster ions
in SIMS imaging can be used to characterize the chemical composition of various
biological samples. The discussion is not meant to provide an exhaustive literature
review of this very large subject, but is aimed toward illustrating the important stra-
tegic advantages associated with the cluster SIMS imaging approach.

4.2 Physics of Cluster SIMS

Molecular dynamics (MD) computer simulations have been recognized to provide

valuable insight into the energetic ion bombardment of solids [11]. Consequently,
the calculations will be used as a platform to discuss the differences between the
atomic desorption event and the polyatomic desorption event. As a note, neutral
atoms are used as the incident projectile in MD calculations, while ions are used as
the incident projectile in SIMS experiments. Therefore, for the remainder of the
chapter, projectiles used in MD simulations will be described in the neutral state,
while projectiles used in SIMS will be described in the ion state. Despite the dis-
crepancy, the physical phenomena in MD simulations are still representative of the
desorption event in SIMS experiments.
An understanding of the atomic bombardment event is essential for a complete
appreciation of the cluster desorption mechanism to be realized. Accordingly,
a representative simulation of a typical trajectory of a 15-keVGa normal incidence
impact on an Ag solid crystal is shown in Fig.4.1a [12, 13]. The atomistic motion
within the Ag crystal can be described as a complicated game of billiards.
Specifically, individual atoms collide with and transfer energy to other individual
atoms. The cascade of substrate atoms is predominantly influenced by the trajec-
tory of the projectile through the solid. The large incident energy and small cross
section of Ga cause the atom to penetrate deep into the Ag crystal. Therefore, a
large amount of the primary energy is deposited well below the solid surface. The
dissipation of significant energy into the bulk causes substantial disruption of Ag
atoms within the structure. Moreover, insufficient energy at the solid surface leads
to the ineffective desorption of material. Hence, a large amount of the projectile
energy is wasted in terms of providing useful SIMS information.
The issue of energy deposition into solids at depths efficient for material ejection
and inefficient for the accumulation of sample damage is critical to the practical
behavior of primary projectiles in SIMS. Polyatomic projectiles overcome the
energy dissipation pitfalls observed during atomic bombardment by distributing its
total incident energy over a number of constituent atoms. For example, each C atom
in a 15-keV C60 projectile has 250eV of kinetic energy [7]. The energy per C atom
is significantly greater than the CC bond strength in C60. As a result, the C60 cluster
dissociates into 60 separate 250-eV C atoms upon impact with the solid. Because
each C atom has 250eV of energy and initiates an individual cascade event, the
74 J. Kozole and N. Winograd

Fig. 4.1 Cross-sectional view of the temporal evolution of a typical collision event leading to the
ejection of atoms due to 15-keVGa (a) and 15-keV C60 (b) bombardment of an Ag surface at
normal incidence. The dimensions of the solid are 101010nm3. The Ag atoms are colored by
original layers in the solid. The projectiles are in black. The bar graphs are the relative frequency
of impacts leading to a given sputter yield for 15keVGa (top) and 15keV C60 (bottom) (This figure
is from Refs. [12, 13]. The calculations in this figure can be viewed as a movie file by visiting the
website http://www.chem.psu.edu/group/bjg/sputtering-animations)

process for the deposition of energy into the substrate is considerably different than
during 15-keVGa bombardment. Specifically, the incident energy dissipates nearer
the solid surface and over a larger surface area. Therefore, the motion of atoms
within the substrate is also expected to be significantly different. A representative
simulation of a typical trajectory of a 15-keV C60 normal-incidence bombardment
of a silver solid crystal is shown in Fig.4.1b [12, 13]. The impact event itself resem-
bles a meteor striking the earth. The motion of substrate material consequent to the
event is similar to an organized expansion of a super-heated and super-dense gas
from a pressurized nozzle. The cascade results in the formation of a significant cra-
ter within the solid and the occurrence of limited sample disruption outside this
immediate region. Furthermore, the impact event is characterized by the large-scale
4 Cluster Secondary Ion Mass Spectrometry 75

Fig. 4.2 Cross-sectional view of the temporal evolution of a typical collision event leading to the
ejection of atoms due to 5-keV Au3 (a) and 5-keV C60 (b) bombardment of an amorphous water
surface at normal incidence. The dimensions of the solid are 291616 nm3. The gray atoms
represent the water molecules. The snapshot at 1ps contains a time-lapse overlay of the incident
projectile motion within the solid. The snapshot at 20ps displays the ejected atoms in red by origi-
nal position in the solid (This figure is from Ref. [14]. The calculations in this figure can be viewed
as a movie file by visiting the website http://www.chem.psu.edu/group/bjg/sputtering-animations)

ejection of material. In fact, the MD simulations illustrate a nonlinear enhancement

of the number of Ag atoms removed for each incident C60. That is, the yield for a
15-keV C60 impact is three times larger than the collective yield from 60 separate
250-eV C impacts. Thus, the polyatomic desorption event abandons linear cascade
principles and enters the realm of the mesoscopic domain.
The unique mechanism for mesoscopic desorption is credited to the near-surface
trajectories of the individual atoms of the cluster within the solid [14]. Specifically,
the trajectories of the constituent atoms are influenced by two factors: the energy
and mass per atom. Together, the factors determine the momentum of the constitu-
ent atoms and the mass-match of the constituent atoms with the substrate atoms. To
-illustrate this point, MD simulations of 5-keV Au3 (Fig.4.2a) and 5-keV C60
(Fig. 4.2b) normal-incident trajectories on amorphous water are considered [14].
The platform is significant since Au3 and C60 are two commonly employed poly-
atomic projectiles. Moreover, the water sample consists of low-mass, weakly bound
atoms comparable to organic materials. In general, the results of the Au3 and C60
impact events on the water sample are similar. The motion of substrate material has
76 J. Kozole and N. Winograd

a mesoscale character, and a significant crater is formed within the solid. However,
the trajectories of the incident projectiles within the solid giving rise to desorption
are different. The motion of the constituent atoms in the Au3 and C60 collision events
are displayed in the time-lapse color portion of Fig.4.2a, b, respectively. The differ-
ent trajectories are attributed to each C atom in the 5-keV C60 particle having 83eV
of energy and a mass of 12amu, while each Au atom in the 5-keV Au3 particle has
1.7keV of energy and a mass of 197amu. Therefore, the Au atoms have a momen-
tum 11 times larger than the C atoms and a mass that is much greater than the sub-
strate atoms (MW=18amu). Consequently, the C atoms are promptly deflected by
the substrate atoms upon impact, and the incident motion and excitation energy is
confined to the near-surface region, whereas the Au atoms are slowly deflected by
the much lighter substrate atoms, and the incident motion and excitation energy
penetrate well beyond the near-surface region. Thus, the trajectories of the C atoms
in the C60 collision event are more efficient for the near-surface deposition of energy
into the solid. Thus, it is no surprise that the number of ice molecules removed in
the C60 experiment (Y=1,644) is larger than the number removed in the Au3 experi-
ment (Y=998). Despite this modest difference, the constituent atoms in both the Au3
and C60 clusters have significantly less momentum than a 15-keVGa atomic projec-
tile. Hence, polyatomic projectiles deposit their incident energy much more effi-
ciently for providing useful SIMS information.

4.3 Properties of Cluster SIMS

The unique mechanism for desorption during polyatomic bombardment has led to
an improved performance in SIMS. Experimental observations have identified sev-
eral important properties when cluster ions are applied to SIMS. Specifically, when
compared to atomic projectiles, polyatomic projectiles have been demonstrated to
enhance secondary neutral and secondary ion yieldsparticularly in the case of
large moleculesenhance surface sensitivity, reduce sample topography and inter-
layer mixing, and make practical molecular depth profile and three-dimensional
imaging experiments [6, 7]. Collectively, the behavior presents a unique opportunity
for increased sensitivity, lateral resolution, and depth resolution during SIMS analy-
sis. Therefore, this section will review the important characteristics of polyatomic
projectiles in SIMS and comment on their implications to the imaging modality.

4.3.1 Enhanced Yields

The MD simulations shown in Fig.4.1a, b indicate the yield of a 15-keV C60 impact
(Y=331) on a silver sample is 15 times larger than the yield for a 15-keVGa impact
(Y=21) [12, 13]. The calculations are supported by the measurement of a similar
yield increase for 15keV C60+ over 15keVGa+ on a polycrystalline Ag substrate
using a quartz crystal microbalance (QCM) to determine the mass removed for each
4 Cluster Secondary Ion Mass Spectrometry 77

impact [15]. Sputter-yield enhancements have also been observed for the
polyatomic bombardment of organic-like materials. Specifically, QCM measure-
ments have determined the number of molecular equivalents sputtered from an
amorphous waterice film using 20keV C60+ (Y=1,800) to be 18 times larger than
the number sputtered using 25keV Au+ (Y=100) [16]. Moreover, when 40keV
C603+ is employed, the absolute number of water molecules removed for each impact
is increased to 10,000 [17]. Additional organic materials determined to experience
large yield enhancements include polymethyl methacrylate (PMMA), polylactic
acid (PLA), arachadic acid, benzene, phenylalanine, gramicidin S, and trehalose, to
name a few [1824]. Regardless of the sample, the amount of material removed for
each cluster ion impact is significantly larger than the material removed for each
atomic ion impact.
Similar to secondary neutral yields, secondary ion yields are enhanced during
cluster projectile impact events [810, 20, 25, 26]. However, the increase observed
for ions is not as straightforward to interpret as the increase observed for neutrals [6,
7]. Particularly, the extent of the secondary ion improvement is specific to the mass
of the molecule being analyzed. For the case of molecules that weigh no more than
500amu, polyatomic projectiles typically increase ion yields by a factor of 10100
over atomic projectiles at comparable energies [20]. The observed ion enhancement
is similar in magnitude to the observed neutral enhancement. Thus, the increase in
ion yield is attributed to a corresponding increase in the neutral yield, and not an
increase in ionization efficiency. This idea is supported by a series of experiments
performed on a barium arachidate (MW=449amu) multilayer structure prepared
using LangmuirBlodgett techniques [20]. The experiments use the known thick-
ness of the structure to measure the number of incident projectiles required to
remove the film using both Ga+ and C60+. From the information, the neutral yield and
ion yield for Ga+ and C60+ can be compared. The values indicate that the enhance-
ment for C60+ over Ga+ is a factor of 100 for both neutrals and ions. Therefore, the
increase in secondary ion yield can be explained by an equivalent increase in the
secondary neutral yield.
On the other hand, molecules weighing at least 500amu typically experience a
1001,000 fold increase in secondary ion yield when polyatomic projectiles are
used [8, 25]. For some examples (Fig.4.3a, b), such as the polymer PS-2000
(MW1,800amu) and the peptide Gramicidin D (MW=1,880amu), a parent ion
signal can only be observed if cluster projectiles are employed [8, 10]. The reason
for larger enhancements at higher masses is not completely understood. However, a
leading conjecture involves a larger propensity for polyatomic projectiles to lift
off large, intact molecules while minimizing fragmentation. This proposition is
supported by the MD simulations in Fig.4.1a, b, which indicate that C60 has a
higher probability of forming larger Ag clusters than Ga does [12, 13]. In addition
to decreased fragmentation, another possible reason for the enhancement includes
an improved environment for ionization under cluster bombardment [6, 7].
Experiments on amorphous ice films suggest that polyatomic projectiles are prodi-
gious producers of protons in the impact region [27]. Thus, the ionization efficiency
could potentially be increased through proton-attachment, chemical ionization of
heavier, slower-moving molecules. Regardless of the scheme, polyatomic
78 J. Kozole and N. Winograd

Fig. 4.3(a) Positive SIMS spectrum of PS-200 using 10keVGa (bottom) and 10keV C60 (top).
(b) Negative SIMS spectrum of Gramicidin D using 15keVGa (bottom) and 15keV C60 (top). All
spectra acquired using the same primary ion fluence (From Refs. [8, 10])

projectiles have been demonstrated to considerably increase the sensitivity and

extend the mass range of molecules in SIMS. These properties make available a
new class of molecules, which includes polymers, peptides, proteins, and lipids, for
detection duringSIMS.
4 Cluster Secondary Ion Mass Spectrometry 79

4.3.2 Reduced Physical Damage

The MD simulations in Fig.4.1a, b provide valuable insight into the incidence of

physical damage during SIMS [12, 13]. Specifically, the calculations indicate that
C60 disrupts the sample bulk to a lesser extent than Ga does. That is, C60 appears to
be more sensitive to the surface material than Ga does. In fact, surface sensitivity
during projectile bombardment has been measured by observing the SIMS response
of an Ag substrate that is covered with waterice overlayers of various thicknesses
and bombarded with 25keV Au1-3+ and 20keV C60+ [16]. The experiments indicate
the amount of silver signal attenuation is greatest when the waterice overlayer is
interrogated with C60+. Therefore, C60+ transfers the largest amount of its incident
energy into the waterice overlayer and has the highest surface sensitivity of the
projectiles studied. The observation of enhanced surface sensitivity has the potential
to reduce ion-beaminduced topography and interlayer mixing.
The incidence of reduced physical damage during polyatomic bombardment has
significant implications to depth-profile experiments. Depth-profiling measurements
are achieved by alternating between erosion cycles and SIMS acquisition cycles
using a single ion beam. That is, the ion beam is operated in direct current (DC)
mode to systematically etch material from a sample layer by layer and operated in
pulsed mode to characterize the composition of the uncovered surfaces [6, 7]. For the
individual layers to be resolved in the analysis, the ion beam must erode the sample
without the occurrence of significant physical damage. Atomic projectiles do not
meet these requirements unless considerable changes to the analytical strategy are
made [28, 29]. These changes, which include low-energy primary ion bombardment
(100500eV), glancing incident angles, and sample rotation, add substantial com-
plexity to the measurements and eliminate the ability to acquire images [28, 29].
Conversely, polyatomic projectiles have been demonstrated in a number of instances
to meet the depth-profile requirements without method modification.
The different behaviors of 15keVGa+ and 15keV C60+ during the controlled
erosion of alternating nickel/chromium (Ni/Cr) layers in a multilayer structure are
shown in Fig.4.4 [28, 29]. The extent of interlayer mixing during the depth-profile
experiment was determined by observing the response of the metal signal as a func-
tion of sample depth and calculating the interface distance between the alternating
layers within the structure. In addition to the interface width, atomic force micro-
scope (AFM) images of the eroded sample area were acquired to evaluate the ion-
beaminduced sample topography. The measurements in Fig.4.4 provide important
insight into the depth resolution achieved during the Ga+ and C60+ experiments. The
depth profile using Ga + (bottom) does not resolve the individual Ni/Cr layers of the
structure; the AFM measurement of the bombarded surface region (far right) indi-
cates a 100-nm root-mean-square (rms) surface roughness. In contrast, the depth
profile using C60+ (top) resolves the individual Ni/Cr layers with interface widths of
10nm and the AFM measurement (left) determines a 2.5-nm rms surface roughness.
Therefore, C60+ controllably erodes the Ni/Cr multilayer structure with a depth reso-
lution of 10nm, a value that is approaching the best resolution attained during low-
energy, glancing incident, atomic projectile depth profiling. Overall, the improved
80 J. Kozole and N. Winograd

Fig. 4.4 The integrated metal atom signal response as a function of total sputter time for depth-
profile experiments on a nine-layer Ni/Cr multilayer structure. The top panel shows the behavior
of 15keV C60 and the bottom panel shows the behavior of 15keVGa. The offset graphs are AFM
images of the eroded region of the Ni/Cr structure using 15keV C60 (left) and 15keVGa (right)
(From Refs. [28, 29])

behavior of C60+ over Ga+ during the Ni/Cr depth-profile experiments is attributed to
reduced interlayer mixing and the absence of ion-beaminduced topography as a
result of enhanced surface sensitivity.
Polyatomic projectiles have been employed to successfully depth-profile a num-
ber of multilayer structures [6, 7]. The ability to probe various materials without the
incidence of significant physical damage has important implications to molecular
depth profiling [6, 7]. To determine the potential for these experiments, the observa-
tion of reduced topography and interlayer mixing during cluster bombardment must
be extended from atomic materials to organic materials. A platform for investigat-
ing the physical damage of organic materials during sample erosion is illustrated in
Fig.4.5 [22, 30, 31]. The platform involves a spin-coated, GGYR peptide-doped,
sugar trehalose film on a silicon substrate. Figure4.5a is an AFM image of the
unbombarded trehalose surface, and Fig.4.5b is an AFM image of the same surface
after 11014cm2 20-keV C60+ bombardment. The AFM measurements indicate the
rms surface roughness of the film before bombardment is 2.2nm, while that after
bombardment is 0.5nm. Therefore, C60+ bombardment does not alter the surface
topography of the trehalose film; however, this observation is not consistent for all
materials [32]. In addition to limited topographical effects, the depth profile in
Fig.4.5c suggests the interface region of the trehalose molecules with the silicon
substrate atoms during sample erosion is 10nm. Thus, the physical structure of the
trehalose film is maintained as a function of sample depth.
4 Cluster Secondary Ion Mass Spectrometry 81

Fig. 4.5 AFM images of (a) an unbombarded sugar trehalose film surface and (b) a 11014-cm2
C60+ bombarded sugar trehalose film surface. The field of view is 20m20 m. The labeled
roughness corresponds to the rms values determined from statistical analysis of the entire image.
(c) Depth-profile plot of secondary ion intensities of trehalose m/z 325 (red circle), GGYR peptide
m/z 452 (green triangle), and silicon (black square) as a function of C60+ ion fluence. The trehalose/
GGYR film is 350nm thick (From Ref. [22])

4.3.3 Molecular Depth Profiling

SIMS experiments of organic materials using atomic projectiles have traditionally

been limited by the accumulation of chemical damage at the solid surface [6, 7, 11].
The chemical damage is created by the ion-beaminduced fragmentation of mole-
cules. After extended bombardment, fragmented molecules collect at the solid
82 J. Kozole and N. Winograd

surface and cover undamaged molecules. To avoid the loss of information, ion beam
erosion of the solid is often limited to 1% of the surface molecules. This restriction,
known as the static limit, considerably reduces the amount of material available for
analysis. Consequently, the sensitivity and lateral resolution of the SIMS imaging
modality are often inadequate for most organic and biological applications when
atomic projectiles are used.
Polyatomic projectiles have the potential to overcome the static limit require-
ment. Collectively, large yields, enhanced surface sensitivity, low topography, and
reduced interlayer mixing open the door to molecular depth-profile experiments.
Molecular depth-profile experiments aim to systematically remove material from an
organic solid layer by layer and expose a molecularly intact surface for SIMS char-
acterization [6, 7]. The success of the experiment relies heavily on the ability to etch
the sample without the accumulation of chemical damage. The large yields and high
surface sensitivities characteristic of cluster bombardment severely confine the
deposition of incident energy to the solid surface. Therefore, the bulk of the ion-
beam fragmented molecules are sputtered during the impact event. Moreover, resid-
ual molecule damage remains near the surface and is easily removed by subsequent
bombardment. Thus, polyatomic projectiles erode organic material at a rate that
prevents the accumulation of significant chemical damage at the sample surface. In
addition to reduced chemical damage, decreased topography and interlayer mixing
allow for the organic material to be removed without a significant physical change
to the underlying solid. Thus, the pieces are in place for successful molecular depth-
profile experiments.
The advantages of cluster projectiles have been used to successfully depth-profile
a number of molecules. Gillen etal. have routinely demonstrated the feasibility of
the experiments using 5keV SF5+ as a sputter projectile [3, 19, 33, 34]. An example
from the research involves the controlled erosion of a polylactic acid (PLA) poly-
mer film dosed with 5% drug molecule [19]. The depth-profile plots of secondary
ion intensity versus SF5+ primary ion fluence for the PLA polymer, the drug mole-
cules, and the silicon substrate are illustrated in Fig.4.6c, d. The measurements are
characterized by three distinct regions: an initial period of molecule signal fluctua-
tion often referred to as the transient region of a depth-profile measurement; an
extended steady-sputter-state region, where the ion intensities do not vary as a func-
tion of erosion time; and the complete disappearance of molecular signal at the silicon
interface. Most notably, the steady sputter state indicates SF5+ can erode the material
without the accumulation of significant chemical damage.
The idea of molecular stability during sample erosion has been extended to addi-
tional organic materials using C60+ as a sputter projectile by Winograd etal. [18, 20,
22, 30, 31, 3537]. An example from the research uses a spin-coated, GGYR
peptide-doped, sugar trehalose film as a platform [22, 30, 31]. A depth-profile plot
of secondary ion intensity versus C60+ primary ion fluence for the trehalose mole-
cule, the peptide molecule, and the silicon substrate is illustrated in Fig.4.5c.
Similar to the PLA polymer films, the trehalose signal and the peptide signal reach
a steady sputter state before the silicon interface is reached. Furthermore, if the
same film is eroded using 20-keV Au+, all molecular signals immediately disappear.
4 Cluster Secondary Ion Mass Spectrometry 83

Fig. 4.6(a) Optical micrograph of a typical sputter crater formed after bombarding a 560-nm-
thick PLA film with 5keV SF5+. Scale bar represents 200m. (b) Si+ molecule-specific SIMS
image of a sputter crater formed after bombardment of a PLA film with SF5+. (c) Secondary ion
intensities as a function of increasing SF5+ primary ion dose for PLA films doped with 20%
4-acetamidophenol: () m/z 152, 4-acetamidophenol (M+H)+; () m/z 109, 4-acetamidophenol
(M+HCOCH3)+; () m/z 145, PLA fragment (2n+H)+; () m/z 128, PLA fragment (2nO)+;
and () m/z 28, Si+. (d) Secondary ion intensities as a function of increasing SF5+ primary ion dose
for PLA films doped with 5% theophylline: () m/z 128, PLA fragment (2nO)+; () m/z 145,
PLA fragment (2n+H)+; () m/z 165, theophylline (M+HO) +; () m/z 181, theophylline
(M+H)+; and () m/z 28, Si+ (From Ref. [19])

Consequent to the observations, a simple analytical model was developed to explain

the basic response of molecular ion intensity as a function of primary ion fluence
during the erosion process [22]. The molecular depth-profile model considers a
number of parameters, including the molecule sputter yield, the damage cross sec-
tion of the surface molecules, and the thickness of the surface layer altered by the
projectile. Specifically, the model is described by the equation

dcs Y tot cb Y tot cs

= s D cs
df nd nd

where cs is the surface concentration of intact molecules, cb is the bulk concentration
of intact molecules, f is the primary ion fluence, Ytot is the total molecule sputtering
yield, n is the molecular density, d is the altered layer depth beneath the surface, and
d is the area of damage on the surface [22]. The first term in the equation describes
the supply of undamaged molecules from the bulk to the surface during the erosion
process, the second term describes the loss of intact molecules from the surface
84 J. Kozole and N. Winograd

during sputtering, and the third term describes the damage of intact molecules
remaining at the surface. Collectively, the equation indicates favorable conditions
for molecular depth profiling exist when the total sputter yield is large relative to the
damage cross section and altered layer thickness. The conditions have repeatedly
been observed in molecular depth-profile experiments using Au3+, Bi3+, and C60+.
Molecules that have been successfully eroded include trehalose, arachadic acid,
PLA polymer, PMMA polymer, amino acids in ice, cholesterol, and phospholipids,
to name a few [1820, 22, 27, 31, 33, 3742]. Currently, the experiments are limited
to a depth resolution of 10nm. Moreover, several molecules have been identified
that do not respond well during the erosion process [32]. To overcome the chal-
lenges and better generalize the strategy, the experimental variables need to be opti-
mized to best fit the conditions of the depth-profile model. Among the variables that
can be varied, the incident energy and incident angle of the cluster projectile seem
most promising. Accordingly, molecular depth-profile studies investigating the
effect of the projectile incident energy and projectile incident angle are currently
ongoing [4244]. The experiments are expected to provide valuable information
about the optimal incident projectile energy and the optimal incident projectile
angle for the deposition of primary energy into a solidan important factor in the
determination of sputter yield and damage volume [45]. Thus, the research should
identify the best conditions for molecular depth profiling and extend the usefulness
of the analysis.

4.3.4 Implications to SIMS Imaging

For a number of molecules, depth-profile experiments are feasible. The ability to

eliminate the static limit requirement has major implications for the SIMS imaging
of organic and biological materials [6, 7, 11]. SIMS images are acquired by raster-
ing a focused ion beam across a sample surface and by collecting a mass spectrum
at a sequence of surface positions. Software can be used to construct a two-
dimensional (2D) image that displays ion intensity as a function of position (pixel)
for a particular mass or a particular set of masses. In principle, the lateral resolution
of the experiment is limited by the size of the ion beam, typically 100nm. However,
when atomic projectiles are employed, a fundamental flaw exists that causes the spa-
tial resolution to be larger than the 100-nm beam size. The pitfall involves the accu-
mulation of chemical damage after bombardment beyond the static limit [6, 7, 11].
Since only 1% of the surface molecules are available for detection, the sensitivity
of the SIMS measurement becomes inadequate as the pixel size approaches
100nm100nm. For instance, a molecular solid (11022 molecules/cm3) has
approximately 5104 surface molecules per 100-nm100-nm pixel. Taking into
consideration the 1% static restriction, only 5102 of the molecules are available
for analysis. With a typical ionization efficiency of 1104, less than 1 molecule
would be detected per 100-nm100-nm pixel [31]. In fact, the pixel size would
4 Cluster Secondary Ion Mass Spectrometry 85

have to be increased to 1m1 m before 1 ion count is registered. Therefore, the

lateral resolution of a static SIMS imaging experiment using atomic projectiles is
fundamentally limited by sensitivity, and not by the size of the primary ion beam.
Polyatomic projectiles increase the number of molecules available for detection
for samples more than one atomic layer thick. An increase in signal is attained by
interrogating a pixel beyond the surface molecules and into the bulk of the solid
[6,7]. Considering the same molecular solid used in the previous example and the
fact that cluster projectiles eliminate the static limit requirement, an increase in
molecules over the same surface area can be achieved by changing the 100-nm100-
nm pixel to a 100-nm100-nm10-nm-deep voxel. The proposed three-dimen-
sional voxel contains 1106 molecules, each of which is available for analysis.
With an ionization efficiency of 1104, 100 molecules would be detected within
the 100-nm100-nm10-nm voxel. Furthermore, 104 molecules would be detected
from a 1-m1-m10-nm voxel. Thus, SIMS imaging using cluster projectiles
has a sensitivity many orders of magnitude greater than traditional static SIMS
imaging using atomic projectiles. Moreover, the effective lateral resolution is con-
siderably better when polyatomic projectiles are used.
In addition to an improved sensitivity, molecular depth-profile experiments have
the potential to construct a three-dimensional (3D) chemical map of a multicompo-
nent solid. A 3D SIMS image of a solid is assembled by acquiring retrospective
images between erosion cycles during molecular depth-profile analysis. Following
data acquisition, software is used to assemble the image series in a manner in which
the ion intensity of a particular mass or a particular set of masses is displayed as a
function of the lateral and depth distributions within the solid. Wucher etal. have
developed a protocol for 3D SIMS image reconstruction [46]. The procedure was
developed using a peptide-dosed, trehalose film patterned by bombardment with a
focused 15-keVGa+ ion beam as a model. A high-resolution, image-series depth
profile of this system was obtained using a focused, 40-keV C60+ ion beam. In addi-
tion to the SIMS images, complementary AFM images of the system were acquired
before and after the depth profile. Together, the measurements were used to calibrate
a depth scale for the construction of the 3D image. However, the calibration is com-
plicated by highly nonuniform erosion rates within different regions of the solid.
The dissimilar erosion rates were attributed to the heterogeneous distribution of
various materials throughout the system. Therefore, the protocol mandates that an
individual depth calibration must be performed for each pixel of the imaged area for
a true 3D representation of the solid to be constructed. A depth scale for each pixel
was calculated using the SIMS and AFM measurements, and a new sequence of 2D
images was assembled to contain the correct mass spectra for a specific depth for a
specific pixel. The new images were stacked in an array to produce a 3D image of
the multicomponent solid. The 3D image results of the experiment before and after
depth calibration are illustrated in Fig.4.7a, b, respectively. The images clearly dem-
onstrate the necessity of the depth-calibration protocol for the accurate composition
of the solid to be visualized. Moreover, the images provide an exciting insight into
the possibly vast information that may be available during 3-D SIMS imaging.
86 J. Kozole and N. Winograd

Fig. 4.7(a) Uncalibrated 3D

representation of a stack of
sequential SIMS images with
equidistant vertical spacing
during a depth-profile
experiment. Red, Ga+
(m/z 69) signal from the
initial Ga+ bombardment;
blue, M+H+ (m/z 452)
molecular ion signal of
GGYR peptide in the
trehalose overlayer; green,
Si+ (m/z 28) from the Si
substrate. (b) Depth-
calibrated 3D representation
of solid composition as
constructed from the
AFM-SIMS measurements.
Color representation is the
same as in (a). The field of
view is 200m280 m and
the total eroded depth is
280nm (From Ref. [46])

4.3.5 Comparison of Different Strategies

Before we introduce the various applications of cluster SIMS, a brief discussion for
rationally selecting the optimal projectile for a specific experiment is appropriate [6, 7].
The assessment will emphasize the difference between Au3+/Bi3+ and C60+ projec-
tiles in the SIMS. Other cluster projectiles, such as SF5+, are not yet amendable to
SIMS imaging due to the poor focusing characteristics and are omitted in the
4 Cluster Secondary Ion Mass Spectrometry 87

discussion for simplicity. In general, C60+ provides enhanced yields and improved
spectral quality as compared to Au3+/Bi3+ [8, 14, 17]. Furthermore, C60+ has a more
surface-sensitive sampling depth and is more effective in molecular depth profiling
[16, 22, 41]. The reason for these observations is attributed to each particle in the
C60+ cluster having less energy than each particle in the Au3+/Bi3+ cluster. On the
other hand, the imaging properties of Au3+/Bi3+ are currently better than C60+ [6, 7].
Specifically, Au3+/Bi3+ provides a brighter, more laterally defined beam size,
although technical advances in C60+ focusing optics are rapidly narrowing the differ-
ence between these projectiles. In addition to smaller beam sizes, Au3+/Bi3+ implants
metal atoms into the solid during the erosion process [41]. Heavy-metal implanta-
tion changes the chemical nature of the solid and may improve the ionization effi-
ciency during SIMS [47]. The ramification of this occurrence has yet to be
determined. Overall, the larger C60+ cluster is better for the acquisition of mass spec-
tra and for molecular erosion, while the smaller Au3+/Bi3+cluster is better for high-
lateral-resolution imaging applications. Perhaps a compromise for optimal 3D
SIMS imaging lies in dual-beam depth-profile analysis: C60+ is used for sample
erosion and spectra acquisition, and Au3+/Bi3+ is used for imaging acquisition.
In addition to an assessment of projectile type, comparing cluster SIMS and its
principal MS imaging complementmatrix-assisted laser desorption ionization
(MALDI)is useful when optimizing an MS strategy for a particular application [48].
Briefly, MALDI uses a matrix solution to segregate analyte molecules from a complex
sample. The analyte-doped matrix crystals are ablated using a UV laser and the
desorbed material is directed into a mass spectrometer. In general, MALDI provides
high-quality mass spectra of large-molecular-weight molecules. The spectra are char-
acterized by little fragmentation and a nearly unlimited mass range (MW 106amu).
In addition, if a focused laser beam is used to acquire the spectra, an image of the
sample can be constructed. MALDI imaging has been especially effective in the assay
of large biological molecules, namely, peptides and proteinsan area SIMS imaging
has been lacking. However, MALDI does not completely depict the vast range of mol-
ecules present within the sample. Specifically, the analysis is limited by little low-mass
information (1,000), a chemical background attributed to the matrix, poor surface
sensitivity (100nm), and a lateral resolution restricted by the 10100-m laser beam
size. The properties of cluster SIMS are an excellent complement to the MALDI
pitfalls. Cluster SIMS excels in research where 01,000-amu target molecule mass
ranges, chemically unmodified sample preparation, uppermost surface layer sensitiv-
ity, and submicrometer spatial resolution are important [6, 7, 11]. Thus, the applica-
tions of cluster SIMS emphasize research where these characteristics are essential.

4.4 Applications of Cluster SIMS

The special properties of cluster projectiles make the SIMS technique an exciting
option for a number of applications in a number of research fields, including
semiconductors, polymers, organic, combinatorial, chemistry, nanotechnology, and
88 J. Kozole and N. Winograd

biology. Perhaps the most intriguing application is the use of cluster SIMS as a
chemical microscope in the discovery of new biology. Therefore, the remainder of
this chapter will discuss the biological analyses made practical by the introduction
of cluster projectiles to SIMS imaging, with particular attention paid to lipid mole-
cules within biological tissue and biological single cells.

4.4.1 Biological Tissue

Recently, the distribution of lipid molecules in cellular membranes has become an

increasingly important subject in the field of biology [49]. The interest has been
prompted by the identification of lipids as key contributors in a number of cell pro-
cesses, including signaling pathways, exocytosis, and endocytosis. The molecules
have also been shown to play an important role in a variety of clinical diseases,
namely Alzheimers and Parkinsons. Thus, additional insight into lipid activity
within cell membranes may lead to new findings in the areas of physiology, neuro-
biology, medicine, and pharmaceutical development.
A useful platform for investigating the function of lipids in intercellular activity
is biological tissue [38, 50, 51]. Presently, strategies for in vivo analysis of these
samples are problematic and provide little molecule specificity (i.e., magnetic reso-
nance imaging, X-ray imaging, and microdialysis). However, procedures for dis-
secting, cryosecting, and preserving tissue while maintaining biological integrity
for ex vivo analysis are better established. Unfortunately, common methods for ex
vivo analysis offer an incomplete chemical representation of the samples. The strat-
egies are flawed by the use of chemical labeling (i.e., fluorescence microscopy),
insufficient molecule specificity (i.e., electron microscopy, Raman spectroscopy),
and inadequate spatial resolution (i.e., MALDI imaging) [11, 50, 51]. In principle,
SIMS is an ideal approach for overcoming these pitfalls. However, SIMS imaging
using atomic projectiles of tissue samples has been limited by inadequate mass
ranges (500amu), poor sensitivities, and the identification of only a few lipid
molecules, namely, the phosphatidylcholine headgroup (MW=184amu) [50, 51].
Collectively, these pitfalls considerably reduce the amount of biologically relevant
information that can be learned.
By replacing atomic projectiles with cluster projectiles, one can considerably
improve the effectiveness of the SIMS imaging modality [6, 7]. The broad reper-
toire of lipid molecules that is detected during cluster SIMS imaging of tissue is
illustrated in Fig.4.8a [50]. The figure is a sequence of negative SIMS molecule-
specific images of a mouse brain section acquired using a focused 25-keV Au3+ ion
beam and by scanning the sample target over a 9-mm9-mm area. Most notably,
the images identify numerous lipid species ranging in mass from 01,000amu,
including phosphate lipid headgroups, cholesterol, palmitate, oleate, stearate, phos-
phosulfatide, phosphatidylinositol, and phosphatidylcholine. Moreover, the images
show a distinctive lateral distribution for each lipid present within the tissue. In fact,
these distributions are so unambiguous that anatomic structures of the brain are
4 Cluster Secondary Ion Mass Spectrometry 89

Fig. 4.8(a) Negative SIMS molecule-specific images of a freeze-dried, coronal-sliced, mouse

brain section acquired using 25keV Au3+. The specific ions, which include cholesterol (m/z 385),
phosphosulfatide (ST, m/z 806, 822, 888, 890), phosphatidylinositol (PI, m/z 965), and unknowns,
mapped in each image are indicated below the image itself. The field of view is 9mm9mm.
(b)Positive and negative SIMS images of the distribution of cholesterol and phosphocholine in a
mouse brain section at different fields of view. The first column of images was a positive SIMS
image at a 9-mm9-mm field of view, the second column of images was a positive SIMS image at
a 500-m500-m field of view, and the third column of images was a negative SIMS image at a
100-m100-m field of view. The effective lateral resolution in the images is 300nm. The mag-
nified images were obtained from the areas indicated by the black squares in the phosphocholine
images (From Ref. [50])
90 J. Kozole and N. Winograd

Fig. 4.9 Positive SIMS molecule-specific images of brain section taken from a rat treated in vivo
with the drug raclopride acquired using 40keV C60+. The specific ions mapped include (a) phos-
phatidylcholine (m/z=184), (b) cholesterol (m/z=369), and (c) raclopride (m/z 247). In each
image, the green intensity represents the substrate. The field of view is 1.6mm8mm. (d) Optical
image of the raclopride-treated, rat brain section. The shaded region illustrates the SIMS interro-
gated area. (e) Mass spectra illustrating the absence of the raclopride molecule in the control sec-
tion and the presence of the raclopride molecule in the drug-treated section (From Ref. [39])

recognized. For instance, the images show a complementary localization between

cholesterol and phosphocholine within the tissue. The heterogeneous distributions
identify the cholesterol-enriched regions as white brain matter and the phosphocholine-
enriched regions as gray matter. In addition, this occurrence of lipid heterogeneity is
observed down to the micrometer scale (Fig.4.8b).
An example of the improved detection levels attainable when cluster SIMS imag-
ing is applied to tissue is shown in Fig.4.9ae [39]. The research uses a focused
40-keV C60+ ion beam to probe the drug raclopridea dopamine uptake inhibitor
within a brain tissue section taken from a rat dosed in vivo. The positive SIMS mole-
cule-specific images of the rat brain section are shown in Fig.4.9ae and are
constructed by stitching together a number of analyses acquired at smaller fields of
view. Figure4.9a illustrates the distribution of phosphocholine (MW=184amu) to
identify the gray-matter region of the brain, Fig.4.9b illustrates the distribution of
cholesterol (MW=369amu) to identify the white-matter region of the brain, and
Fig. 4.9c illustrates the distribution of the raclopride drug (MW=247amu). The
images indicate that raclopride can be identified from within the treated brain section
and that the drug is distributed within the white-matter region of the brain (Fig.4.9e).
Unfortunately, the spatial distribution of the drug within the sample does not com-
pletely agree with the known gray-matter location of the dopamine receptors.
Although the basis for the inconsistency is unknown, potential explanations include
4 Cluster Secondary Ion Mass Spectrometry 91

the redistribution of the drug after tissue preparation and the incidence of different
ionization environments in the different regions of the tissue. Regardless of the issues,
the experiment is encouraging since the raclopride-specific image demonstrates a
two-part-per-million (ppm) detection level, a value that will improve even further
once molecular depth profiling is applieda feat that has already been deemed fea-
sible on tissue homogenate when using 40keV C60+ as erosion projectile [39].
Broad ranges of lipid molecules, distinct lipid localizations, micrometer lateral
resolution, 1-ppm sensitivities, and molecular depth profiling make cluster SIMS
imaging a powerful technique for the analysis of tissue. A real-life biological appli-
cation utilizing the strategy is illustrated in Fig.4.11 [52]. The application involves
the interrogation of human atherosclerotic plaque to determine the role of lipids in
the development of cardiovascular disease. In order to analyze the unhealthy artery,
SIMS spectra of healthy rat aorta were acquired as a control using a Bi3+ ion beam
(Fig.4.10a, b). The spectra indicate that in a healthy artery, the lamellar tissue in the
intima region (the innermost layer) is enriched in cholesterol, oxysterol, and diacyl-
glycerols and that the smooth muscle tissue in the media region (the middle layer)
is enriched in phosphocholine. Interestingly, the localization of lipids in the human
atherosclerotic plaque shows a different distribution. Specifically, the SIMS
molecule-specific image taken from the unhealthy artery shows an irregularly
shaped distribution of cholesterol in the intima region and an irregularly located
distribution of diacylglycerols in the media region. The unique lipid distributions
suggest cholesterol and diacylglycerols play an important role in the development
of cardiovascular disease. It is hypothesized that cholesterol may be an important
ingredient in the apoptotic process leading to plaque formation and that diacylglyc-
erol may be a mediator in the activation of this process. These hypotheses may be
further developed by characterizing the role of these lipids in the formation of
plaque at the cellular level.

4.4.2 Biological Single Cells

A useful platform for investigating lipids at the cellular level is biological single
cells [6, 7]. For reasons similar to the analysis of tissue, SIMS is an excellent strat-
egy for analyzing these samples. An example of the biological findings that can be
learned from this partnership is illustrated in Fig.4.11 [53]. The research involves
the examination of highly curved lipids during cell conjugation. Specifically, the
junction region between two conjoined cells contains a large number of fusion
pores. The fusion pores, which are important in a number of cellular events, includ-
ing sexual reproduction, exocytosis, and endocytosis, are hypothesized to be formed
through the heterogeneous redistribution of lamellar and nonlamellar lipids through-
out the membrane. Particularly, increased levels of nonlamellar or high-curvature
lipids and decreased levels of lamellar or rigid lipids at the junction region are
expected to provide the membranes with the necessary elasticity to form the highly
curved intermediate structures required for conjugation (Fig.4.11d). To examine
92 J. Kozole and N. Winograd

Fig. 4.10(a) Positive and

(b) negative SIMS spectra
from a high-pressure,
fresh-frozen, freeze-fractured,
and freeze-dried rat aorta
acquired using a Bi3+ ion
beam. The spectra were used
as reference in the analysis of
human atherosclerotic plaque.
(c) An overlay of positive
SIMS molecule-specific
images of human
atherosclerotic plaque. The
field of view is
500m500 m. The red
signal in the image represents
cholesterol, the green signal
represents phosphocholine,
and the blue signal represents
diacylglycerol. The section M
indicates the media region of
the tissue and the section I
indicates the tunica intima
region (From Ref. [52])

this hypothesis, SIMS molecule-specific images of mating Tetrahymena cells were

acquired using a focused 15-keV In+ ion beam. The images, which are illustrated in
Fig.4.11a, b, show a heterogeneous distribution of 2-aminoethylphosphonolipid
(2-AEP; MW =126amu), a nonlamellar lipid, and phosphocholine (PC;
MW=184amu), a lamellar lipid. Most notably, the fusion site between the two cells
contains an elevated amount of the cone-shaped lipid 2-AEP and a depleted amount
of cylindrical-shaped lipid PC (Fig.4.11c). This observation supports the idea that
4 Cluster Secondary Ion Mass Spectrometry 93

Fig. 4.11(ab) Positive SIMS molecule-specific images of mating, freeze-fractured, Tetrahymena

cells acquired using a focused 15-keV In+ ion beam. The specific ions are (a) 2-aminoethylphos-
phonolipid (2-AEP; MW=126amu), a nonlamellar lipid. The field of view is 100m100 m,
and (b) phosphocholine (PC; MW=184amu), a lamellar lipid. (c) Mass spectrum from the pixels
along the conjugation junction (top spectrum) and from the cell body (bottom spectrum). (d) A
schematic of the membrane fusion intermediate structure. The wavy lines depict the acyl tailgroups
of the membrane phospholipids. The green circles are PC and red circles are 2-AEP, a cylinder-
shaped lamellar lipid. The black circles represent the headgroup of 2-AEP, a cone-shaped nonla-
mellar lipid. Membrane fusion sites contain a large amount of cone-shaped lipids since these lipids
fit well into contoured intermediate structures (From Ref. [53])

membrane fusion sites contain an increased concentration of nonlamellar lipids

since cone-shaped lipids fit well into highly curved intermediate structures.
Future experiments using mating Tetrahymena aim to characterize the biologi-
cal response that redistributes the lipids in preparation for the conjugation event.
Unfortunately, single-cell SIMS imaging using atomic projectiles is limited to the
identification of only a few lipid species. Moreover, these analyses are restricted to
only the lipid molecules present within the uppermost layers of the exposed s urface
[54]. Therefore, a complete picture of the chemistry involved in the conjugation
94 J. Kozole and N. Winograd

Fig. 4.12(ab) Optical micrograph of a Xenopus laevis oocyte cell mounted on copper tape for
40-keV C60+ SIMS analysis (a) before etching and (b) after etching. (c) Positive and (d) negative
SIMS spectra from the oocyte cell after a 11015 C60+/cm2 etch. (ef) SIMS molecule-specific
images of the oocyte cell after a 11015 C60+/cm2 etch. The specific ions are (e) cholesterol
(MW=369, positive SIMS) and oleic acid (MW=281, negative SIMS). The field of view is
1mm1mm (From Ref. [54])

event cannot be drawn using the atomic SIMS imaging modality. Nevertheless, the
desired information can be attained when using cluster projectiles as both an imag-
ing projectile and an erosion projectile in 3D SIMS imaging. An experiment dem-
onstrating the effectiveness of this strategy is shown in Figs.4.12 and 4.13 [54].
The research involves the 3D molecule-specific SIMS imaging of a freeze-dried
Xenopus laevis oocyte using a 40-keV C60+ ion beam (Fig.4.12a, b). The positive
and negative SIMS spectra shown in Fig.4.12c, d demonstrate the wide array of
lipid molecules that can be detected from the oocyte cell. Specifically, the lipid
species identified in the spectra include phosphocholine (MW=58, 86, 166,
184amu), cholesterol (MW=369amu), lipid fatty acid side chains (MW: 540
720amu), and glycosphingolipids (MW: 8001,000amu). In addition to lipid
identification, the strategy is exciting with respect to lipid distribution as a function
of the lateral and depth positions. Figure4.13ad display a 3D chemical represen-
tation of the oocyte cell. Of particular interest, the distribution of lipids within the
image varies as a function of depthwith some species having a maximum inten-
sity well below the sample surface. This observation is attributed to the removal of
some molecules and the uncovering of others, suggesting that the 3D chemical
integrity of the sample is maintained during erosion and the change in signal is
representative of the molecular composition within the cell. Thus, 3D SIMS imag-
ing using C60+ is an effective means for probing the chemistry of single cells in
three dimensions.
4 Cluster Secondary Ion Mass Spectrometry 95

Fig. 4.13(ad) Three-dimensional positive SIMS image of a oocyte cell for (a) phosphocholine
(MW=58, 86, 166, and 184), (b) signal summed over m/z range 540650, (c) signal summed over
the m/z range 815960, and (d) cholesterol peak (MW=369) acquired using 40keV C60+. The 3-D
representation is cut along two dimensions to display the third dimension. Color scale normalized
for total counts per pixel for each mass. The field of view is 1mm1mm and the depth is 100m
(From Ref. [54])

4.5 Future Directions of Cluster SIMS

The special properties of cluster projectiles have greatly improved the usefulness of
the SIMS imaging modality [6, 7]. Therefore, the future use of the technique in the
characterization of real-life biological samples, such as mating Tetrahymena, may
lead to the discovery of new and exciting biology. While the prospects for these
types of experiments are promising, several issues still remain that could possibly
limit the technique, including sample preparation, instrument duty cycle, and ion-
ization probability. While these challenges are not specific to the ion beam used,
cluster projectiles may be useful in overcoming them.
Sample preparation is essential when acquiring meaningful SIMS images of single
cells [5557]. In order to prepare cells for in-vacuum analysis, the 3D integrity of the
sample must be preserved in the solid state with micrometer precision. Common
strategies for cell preservation include freeze-drying, freeze-etching, freeze-fractur-
ing, chemical fixation, and sugar vitrification [5559]. Although each of the strategies
96 J. Kozole and N. Winograd

has its advantages and its disadvantages, the gold standard is freeze-fracturing.
Specifically, freeze-fracturing involves quenching hydrated cells for in-vacuum, cryo-
genic fracture and analysis. Unfortunately, the usefulness of the approach is limited
by the absorption of an ice contaminant overlayer. Nevertheless, the overlayer can be
removed with no damage to the underlying molecules using cluster ion bombardment
[59]. Therefore, a good strategy for maintaining the 3D structure of biological sam-
ples while exposing an intact, meaningful surface for SIMS imaging involves com-
bining freeze-fracture technology and cluster ion beams.
In addition to sample preparation, SIMS imaging of biological samples is also
restricted by the instrumental duty cycle of axial time-of-flight (ToF) MS [60]. For
example, to analyze each of the 108 molecules in a 1-m1-m10-nm voxel
using a 10-pA ion beam with a 100-ns pulse width and a 10-kHz repetition rate, 10s
of instrumental time is required (assuming a sputter yield of 100). Therefore, it
would take 182h to acquire a 256-voxel256-voxel image. However, by eliminat-
ing the pulsed nature of the ion beam, the same image can be acquired in only
11min, or 103 less time. Specifically, a direct current (DC) ion beam is utilized by
switching from an axial ToF geometry to an orthogonal ToF (oToF) geometry. In
addition to rapid sampling, the proposed design allows for tandem MS/MS, colli-
sional focusing, collisional cooling, and ion trapping when a series of quadrupoles
(q) is placed between the sampling region and the ToF region. Therefore, the instru-
mental scheme extends the usefulness of the SIMS imaging modalitymaking the
quadrupole-orthogonal ToF design with a cluster ion beam a possible SIMS instru-
ment of the future.
Perhaps the most critical of the remaining issues is the ionization probability of
the SIMS technique [61, 62]. Typically, only 1 in 104 desorbed molecules is an ion,
meaning the majority of the signal is lost in the neutral fraction. Postionization of
the neutral molecules using light sources is a common strategy for recovering the
lost signal. To date, the postionization analysis has been limited by extensive frag-
mentation [61, 62]. The observation is attributed to photofragmentation and photo-
dissociation of molecules sputtered with internally excited electronic states. The
pitfall may be overcome by using cluster ion bombardment to desorb electronically
cooled molecules [15, 63, 64]. Specifically, molecule cooling is achieved through
large-scale collision events in the dense sputter plume. Therefore, by combining
cluster projectiles with light sources, such as vacuum ultraviolet (VUV) radiation,
mid-infrared (IR) light sources, and femtosecond visible light sources, postioniza-
tion of neutral molecules may be possible without extensive fragmentation, improv-
ing the ionization probability and overall sensitivity of the experiments.

4.6 Summary

Recently, cluster projectiles have been introduced as replacements to atomic projec-

tiles in the SIMS imaging modality. Since cluster ions have less energy per atom
than atomic ions, a unique mechanism for desorption is utilized during bombard-
ment. The mechanism, which is mesoscopic in nature, has led to the observation of
4 Cluster Secondary Ion Mass Spectrometry 97

several important properties when cluster projectiles are applied to SIMS, including
enhanced yields, reduced damage, and the feasibility of molecular depth profiling.
Together, the properties considerably extend the usefulness of the SIMS imaging
modalityespecially when used to interrogate lipid molecules in biological tissue
and biological single cells. Specifically, the analyses are characterized by extended
mass ranges, improved detection levels, submicrometer lateral resolution, and 3D
molecule-specific imaging. Moreover, the usefulness of the analyses can be
improved further by combing cluster ion beams with freeze-fracture technology,
qq-oToF instrumental design, and postionization using light sources.

Acknowledgments The authors acknowledge the National Institutes of Health under grant num-
ber EB002016-13, the National Science Foundation under grant number CHE-0555314, and the
Department of Energy under grant number DE-FG02-06ER15803 for financial support.

References

1. Benninghoven A. Zeitschrift for Physik AAtoms Nuclei. 1970;230:40317.

2. Appelhans AD, Delmore JE. Anal Chem. 1989;61:108792.
3. Gillen G, Simons DS, Williams P. Anal Chem. 1990;62:212230.
4. Cornett DS, Lee TD, Mahoney JF. Rapid Commun Mass Spectrom. 1994;8:9961000.
5. McMahon JM, Dookeran NN, Todd PJ. J Am Soc Mass Spectrom. 1995;6:104758.
6. Winograd N, Postawa Z, Cheng J, Szakal C, Kozole J, Garrison BJ. Appl Surf Sci.
2006;252:683643.
7. Winograd N. Anal Chem. 2005;77:142a9.
8. Weibel D, Wong S, Lockyer N, Blenkinsopp P, Hill R, Vickerman JC. Anal Chem. 2003;
75:175464.
9. Davies N, Weibel DE, Blenkinsopp P, Lockyer N, Hill R, Vickerman JC. Appl Surf Sci. 2003;
203:2237.
10. Wong SCC, Hill R, Blenkinsopp P, Lockyer NP, Weibel DE, Vickerman JC. Appl Surf Sci.
2003;203:21922.
11. Vickerman JC, Briggs D, editors. ToF-SIMS: surface analysis by mass spectrometry.

Manchester: SurfaceSpectra; 2001.
12. Postawa Z, Czerwinski B, Szewczyk M, Smiley EJ, Winograd N, Garrison BJ. J Phys Chem B.
2004;108:78318.
13. Postawa Z, Czerwinski B, Szewczyk M, Smiley EJ, Winograd N, Garrison BJ. Anal Chem.
2003;75:44027.
14. Russo MF, Garrison BJ. Anal Chem. 2006;78:720610.
15. Sun SX, Szakal C, Winograd N, Wucher A. J Am Soc Mass Spectrom. 2005;16:167786.
16. Szakal C, Kozole J, Russo MF, Garrison BJ, Winograd N. Phys Rev Lett. 2006;96:216104.
17. Russo MF, Szakal C, Kozole J, Winograd N, Garrison BJ. Anal Chem. 2007;79:44938.
18. Szakal C, Sun S, Wucher A, Winograd N. Appl Surf Sci. 2004;2312:1835.
19. Mahoney CM, Roberson SV, Gillen G. Anal Chem. 2004;76:3199207.
20. Sostarecz AG, McQuaw CM, Wucher A, Winograd N. Anal Chem. 2004;76:66518.
21. Smiley EJ, Winograd N, Garrison BJ. Anal Chem. 2007;79:4949.
22. Cheng J, Wucher A, Winograd N. J Phys Chem B. 2006;110:832936.
23. Van Stipdonk MJ, Harris RD, Schweikert EA. Rapid Commun Mass Spectrom. 1996;

10:198791.
24. Tempez A, Schultz JA, Della-Negra S, Depauw J, Jacquet D, Novikov A, Lebeyec Y, Pautrat
M, Caroff M, Ugarov M, Bensaoula H, Gonin M, Fuhrer K, Woods A. Rapid Commun Mass
Spectrom. 2004;18:3716.
98 J. Kozole and N. Winograd

25. Ostrowski SG, Szakal C, Kozole J, Roddy TP, Xu JY, Ewing AG, Winograd N. Anal Chem.
2006;78:973.
26. Xu JY, Szakal CW, Martin SE, Peterson BR, Wucher A, Winograd N. J Am Chem Soc.
2004;126:39029.
27. Conlan XA, Lockyer NP, Vickerman JC. Rapid Commun Mass Spectrom. 2006;20:132734.
28. Sun S, Szakal C, Roll T, Mazarov P, Wucher A, Winograd N. Surf Interface Anal.

2004;36:136772.
29. Sun S, Wucher A, Szakal C, Winograd N. Appl Phys Lett. 2004;84:51779.
30. Cheng J, Winograd N. Appl Surf Sci. 2006;252:6498501.
31. Cheng J, Winograd N. Anal Chem. 2005;77:36519.
32. Shard AG, Brewer PJ, Green FM, Gilmore IS. Surf Interface Anal. 2007;39:2948.
33. Mahoney CM, Fahey AJ, Gillen G. Anal Chem. 2007;79:82836.
34. Wagner MS, Lenghaus K, Gillen G, Tarlov MJ. Appl Surf Sci. 2006;253:260310.
35. Wucher A, Sun S, Szakal C, Winograd N. Appl Surf Sci. 2004;2312:6871.
36. Sostarecz AG, Sun S, Szakal C, Wucher A, Winograd N. Appl Surf Sci. 2004;2312:17982.
37. Wucher A, Sun SX, Szakal C, Winograd N. Anal Chem. 2004;76:723442.
38. Touboul D, Kollmer F, Niehuis E, Brunelle A, Laprevote O. J Am Soc Mass Spectrom.
2005;16:160818.
39. Jones EA, Lockyer NP, Vickerman JC. Int J Mass Spectrom. 2007;260:14657.
40. Kozole J, Szakal C, Kurczy M, Winograd N. Appl Surf Sci. 2006;252:678992.
41. Cheng J, Kozole J, Hengstebeck R, Winograd N. J Am Soc Mass Spectrom. 2007;18:40612.
42. Fletcher JS, Conlan XA, Jones EA, Biddulph G, Lockyer NP, Vickerman JC. Anal Chem.
2006;78:182731.
43. Kozole J, Willingham D, Winograd N. Appl Surf Sci. 2007, submitted.
44. Cheng J, Wucher A, Winograd N. Appl Surf Sci. 2007, submitted.
45. Ryan KE, Smiley EJ, Garrison BJ. Appl Surf Sci. 2007, submitted.
46. Wucher A, Cheng J, Winograd N. Anal Chem. 2007;79:552939.
47. Marcus A, Winograd N. Anal Chem. 2006;78:1418.
48. Karas M, Bahr U, Ingendoh A, Nordhoff E, Stahl B, Strupat K, Hillenkamp F. Anal Chim
Acta. 1990;241:17585.
49. Mukherjee S, Maxfield FR. Annu Rev Cell Dev Biol. 2004;20:83966.
50. Sjovall P, Lausmaa J, Johansson B. Anal Chem. 2004;76:42718.
51. Sjovall P, Johansson B, Lausmaa J. Appl Surf Sci. 2006;252:696674.
52. Malmberg P, Borner K, Chen Y, Friberg P, Hagenhoff B, Mansson JE, Nygren H. Biochim
Biophys Acta-Mol Cell Biol L. 2007;1771:18595.
53. Ostrowski SG, Van Bell CT, Winograd N, Ewing AG. Science. 2004;305:713.
54. Fletcher JS, Lockyer NP, Vaidyanathan S, Vickerman JC. Anal Chem. 2007;79:2199206.
55. Colliver TL, Brummel CL, Pacholski ML, Swanek FD, Ewing AG, Winograd N. Anal Chem.
1997;69:222531.
56. Pacholski ML, Donald M, Cannon J, Ewing AG, Winograd NJ. Am Chem Soc.

1999;121:47167.
57. Roddy TP, Donald M, Cannon J, Meserole CA, Winograd N, Ewing AG. Anal Chem.
2002;74:40119.
58. Parry S, Winograd N. Anal Chem. 2005;77:79507.
59. Kurczy ME, Piehowski P, Parry S, Ewing AG, Winograd N. Appl Surf Sci. 2007, submitted.
60. Carado A, Kozole J, Passerelli M, Winograd N, Loboda A, Wingate J. Appl Surf Sci. 2007,
submitted.
61. Vorsa V, Kono T, Willey KF, Winograd N. J Phys Chem B. 1999;103:788995.
62. Vorsa V, Willey KF, Winograd N. Anal Chem. 1999;71:57481.
63. Wojciechowski IA, Kutliev U, Sun SX, Szakal C, Winograd N, Garrison BJ. Appl Surf Sci.
2004;2312:727.
64. Wojciechowski IA, Sun SX, Szakal C, Winograd N, Garrison BJ. J Phys Chem A.

2004;108:29938.
Chapter 5
Biological Tissue Imaging at Different Levels:
MALDI and SIMS Imaging Combined

J. Stauber and Ron M.A. Heeren

Abstract Mass spectrometry has been employed to analyze the composition and
structure of biologically relevant molecules in solution. Advances in methodology
and instrumentation developments now allow the application of mass spectrometry
for local biomolecular analysis directly on biological tissue surfaces; this technique
is called imaging mass spectrometry (IMS). IMS is an innovative discovery tool for
the biomedical sciences. This chapter describes the two main approaches relevant
for molecular tissue imaging studies in the life sciences: secondary ion mass spec-
trometry (SIMS) imaging and matrix-assisted laser desorption and ionization
(MALDI)based imaging techniques. The benefits of imaging mass spectrometry
for the fields of drug metabolism, lipidomics, and proteomics are discussed.
Integrated MS imaging and proteomics protocols as well as tandem-MS imaging
strategies, which are key to the identification of larger-molecular-weight compounds,
are also reviewed.

5.1 Introduction

Imaging mass spectrometry (IMS) [1] is a very sensitive molecular imaging tech-
nique that provides combined molecular resolution and spatial resolution. It allows
the identification and localization of the molecular content directly from tissue sec-
tions, single cells, and many other surfaces. The key features for biological studies
are the sensitivity provided by modern mass spectrometers, the label-free nature of
the technique, the ability to image posttranslational modifications, and the spatial
resolution, which ranges from the organisms level (hundreds of micrometers) to the

J. Stauber R.M.A. Heeren (*)

FOM-Institute for Atomic and Molecular Physics, Science park 104,
1098 XG Amsterdam, The Netherlands
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 99

DOI 10.1007/978-3-319-01360-2_5, Springer International Publishing Switzerland 2014
100 J. Stauber and R.M.A. Heeren

350

300

Number of publications 250

200

150

100

50

0
1985 1990 1995 2000 2005 2010
Year

Fig. 5.1 The number of publications resulting from a web-of-science search using the keywords
MALDI imaging (open circles) and imaging mass spectrometry (solid circles). The results
were obtained from the December 31, 2009, database

cellular level (tens of nanometers). IMS allows the simultaneous detection and
imaging of thousands of species images in a single experiment. As such, it consti-
tutes an efficient, multicomponent molecular imaging technique. The large number
of new studies that include imaging mass spectrometry is evidence of the rapid
growth of the field (Fig. 5.1).
IMS can be used to study different compositions and structures of surfaces in the
context of biological studies. The IMS protocols and instruments have been devel-
oped to study the biodistribution of endogenous compounds, such as lipids or pro-
teins, and exogenous compounds, such as polymers or drugs designed for tissue
treatment. IMS helps to understand biological processes from subcellular to multi-
cellular levels of spatial resolution. Moreover, it detects many different types of
compounds in order to cover the large number of molecules present in the animal
and vegetable kingdoms.
Compared to other molecular imaging approaches, such as magnetic resonance
imaging (MRI), positron emission tomography (PET), or fluorescent immunochem-
istry, IMS provides unique information: IMS identifies molecules without the label-
ing of compounds, which allows the discovery of new localized compounds; no
other technique can do this. Even though IMS is an Ex vivo technique and not an in
vivo diagnosis technique such as MRI or PET, it can be coupled with them to vali-
date molecular repartition, examine degradation of biomarkers, and/or study drug
delivery, as demonstrated in recent publications [2].
Biological surface analysis with mass spectrometry has evolved around two
main desorption and ionization methods: secondary ion mass spectrometry (SIMS)
and matrix-assisted laser desorption ionization (MALDI). Both techniques are used
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 101

in imaging mass spectrometry and combined with different mass analyzers and
detectors in order to increase the sensitivity and spatial resolution. SIMS has been
in development since the early 1970s and applied to many different biological sur-
faces [2]. Although SIMS was initially limited to elements and small molecules,
the detection and imaging of higher-molecular-weight species were realized
through different surface modifications, such as metal-assisted SIMS (MetA-SIMS)
and matrix-enhanced SIMS (ME-SIMS) [35], as well as by the use of cluster ion
beam as the projectile species, as outlined in Chap. 4 of this book. MALDI imaging
was developed in the late 1990s to image larger, intact molecules such as peptides
and proteins. The SIMS and MALDI techniques both possess complementary char-
acteristics in mass range, sensitivity, and spatial resolution, as will be demonstrated
in this chapter.
IMS is crucially dependent on suitable and adequate sample-preparation tech-
niques, as is any molecular analytical technique. Sample preparations, including
solvent treatment, sample storage, and matrix deposition, are important for classical
proteomics and lipidomics experiments directly on tissue. Reproducible sample
preparations allow sensitive and high-resolution image analysis with ME-SIMS,
MetA-SIMS, and MALDI imaging. Due to the different sensitivities and mass
ranges accessible by SIMS and MALDI, different classes of molecules can be
detected by combining these techniques [6]. Peptides and proteins, as well as lipids,
metals, drugs, and metabolites, can be monitored and localized in different biologi-
cal samples. Combinations of specific (bio-)markers that characterize a disease state
can be identified, as IMS provides simultaneous images of different compounds
found on pathological tissue sections.
In this chapter, we present the three biomolecular IMS techniques (ME-SIMS,
MetA-SIMS, and MALDI), discuss different tissue-preparation methods, and sum-
marize representative applications in different fields of research, such as lipidomics,
plant studies, pharmaceuticals, proteomics, and clinical proteomics. The chapter
ends with an outlook on future developments and the related applications of surface-
enhanced SIMS and MALDI imaging. These high-end mass-spectrometricbased
imaging methodologies are continuously under development to improve the sensi-
tivity, resolution (both spatial and mass-resolving power), and data workup proto-
col, in order to increase the usage of IMS in clinical medicine.

5.2 Imaging Mass Spectrometry Modes

5.2.1 Imaging Modes

Many different ionization methods and mass analyzers have been used in imaging
mass spectrometry (IMS) experiments. SIMS and MALDI have emerged as the two
dominant methodologies in the generation of mass-resolved molecular images of
surfaces. Both display different but complementary molecular imaging capabilities
102 J. Stauber and R.M.A. Heeren

and are employed in a wide variety of (biological) applications. Traditionally, SIMS

targeted elements and small to medium-sized biomolecules (under 600 Da), whereas
MALDI extended molecular imaging capabilities to a larger-molecular-weight
domain (100 Da80 kDa).
The two techniques combine an ion source, a mass analyzer, an ion detector, and
data acquisition and processing software. They display complementary spatial reso-
lution, sensitivity, and molecular weight ranges. Two different types of imaging
modes, irrespective of the ionization method, are used: microprobe mode imaging
and microscope mode imaging. The next section briefly explains the two different
imaging approaches

5.2.2 Microprobe Mode

Microprobe (Fig. 5.2) is the most common imaging mode in mass spectrometry. It
is a relatively straightforward technique to image a small, localized region by focus-
ing a desorption/ionization beam on the sample. The ionization beam rasters, or
scans, a selected region of the sample, and a mass spectrum is recorded for each
beam shot. The mass spectra are stored along with the coordinates of the analyzed
spot, as defined by the focus and position of the ionization beam. Molecular images
of different ions are retrospectively reconstructed by dedicated software. This
microprobe scanning mode is usually used with SIMS and MALDI imaging mass
spectrometry. The spatial resolution of the ionization beam can be as high as 50 nm
in SIMS and is conventionally limited by the laser spots size (typically 50 m) in
MALDI. For that reason, the microscope imaging mode was developed and applied
for MALDI studies in order to increase the spatial resolution [7].

5.2.3 Microscope Mode

The microscope imaging (Fig. 5.2) mode does not require position rastering with
focused desorption beams. Instead, ion-optical elements in the time-of-flight mass
spectrometer are used to retain the spatial organization of the ions after desorption
and ionization between the ion source and the detector [7]. This results in a mass-
resolved projection of the spatial origin of the ions generated at the sample surface
by the defocused ionization beam and allows direct mass-resolved molecular image
observation when combined with a position-sensitive detector. The magnification of
the microscope elements, the quality of the ion optics, and the resolution of the
position-sensitive detector determine the spatial resolution. The best spatial resolu-
tion demonstrated to date is a pixel size of 600 nm [7, 8]. The microscope mode
directly analyzes a larger field-of-view (FOV) than the microprobe mode: It allows
for the analysis of up to a 400-m-diameter area. The microscope mode is approxi-
mately 10,000 times faster than the microprobe mode.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 103

Fig. 5.2 Schematics illustrating the two approaches in molecular imaging mass spectrometry: (1)
Microprobe-mode imaging records the mass spectra in spot-by-spot (x, y) position on the sample
surface; (2) microscope-mode imaging records magnified images of mass-resolved ion distribu-
tions using a 2D detector

5.3 MALDI and SIMS: Two Sources of Ionization

5.3.1 MALDI Source

Developed by Karas and Hillenkamp in 1987 [9, 10], MALDI allows the simultane-
ous ionization/desorption of sample molecules, provides high sensitivity, and allows
the analysis of large (up to 200 kDa) molecules. MALDI imaging was introduced in
1997 with advanced software capabilities [11, 12]. A large number of applications
were presented in the last decade, with many biological applications. Most studies
104 J. Stauber and R.M.A. Heeren

use the MALDI imaging technique in a microprobe mode, by combining laser irra-
diation of the sample and advanced software capabilities to reconstruct images. More
recently, commercial instruments with well-optimized system control and a laser
spot size down to approximately a 50-m diameter have been developed. Important
MALDI considerations include the size of the laser beam, the wavelength, the pulse
width, and the matrices used. MALDI imaging traditionally uses either N2 (337-nm)
or Nd:YAG (335-nm) UV lasers, with the latter becoming more routinely used over
the last two years, resulting in increased laser performance and stability. The laser
pulse length (width) was found to have very little to no influence on MALDI mass
spectra, at least up to pulse lengths of tens of nanoseconds [13]. This suggests that
the desorption/ionization process is determined by the energy density supplied to the
sample by the laser pulse (fluence or energy density, 100 J/m2) rather than by the rate
of energy flow (irradiance, W/cm2). The fluence or energy flow required on the target
is dependent on the spot size and pulse length (typically 35 ns).
(a) Desorption/ionization process
Since 1976, laser irradiation has been employed to ionize peptides from solid
samples deposited on a surface [14]. Studies over the first decade of its use showed
it was not efficient for large peptides. Therefore, the use of a matrix as an energy-
transfer medium presented advantages in the desorption/ionization laser process.
The technique of MALDI imaging uses a pulsed beam of laser in the UV (or
infrared laser source) to desorb and ionize a mixture of co-crystallized matrix and
tissue. The matrix minimizes the sample degradation caused by energy absorption
of the laser beam. The laser-transmitted energy is absorbed by the matrix, which
acts as a resonant absorber for the photons and causes a phase explosion due to
overheating of the surface. The sample molecules expand, are ejected into the gas
phase, ionized, and then detected.
Partially ejected molecules are ionized by proton transfer, in the solid phase
either before desorption or after desorption by collision with the excited matrix and
other molecules. These processes generate different singly or multiply charged ions
[M + nH]n+, but the majority are singly charged ions.
No single mechanism can explain all the ions observed in a MALDI experiment.
The large range of samples, matrices, and experimental parameters hinders the elu-
cidation of ion-formation mechanisms. However, ion-formation mechanisms are
described by two categories: primary and secondary ion formation [15]. The times-
cale for ion formation is of utmost interest, where the laser pulse typically lasts
35 ns, but the time required for expansion is much longer, many microseconds
[16]. Primary ionization is caused by the laser pulse directly or by ion-molecule
reactions within the excited matrix plume. Basic thermodynamics of ion formation,
multiphoton ionization [17], energy pooling [18], and excited-state proton transfer
[9] describe primary ionization. Although many primary ionization pathways have
been identified, it has been hypothesized that secondary processes that occur within
the expanding plume of sputtered species are the most important factor for the pro-
duction of ions [19, 20]. In the expanding plume, reactions between ions and neu-
trals continue as long as there are collisions. Electron and proton transfer from the
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 105

matrix to the analyte is probably the most important secondary reaction and causes
the predominantly detected protonated form of peptides and proteins. The gas-phase
reaction of these two products is exoergic (protonation of peptide and deprotonation
of matrix) since the proton affinity of the matrix is lower than the gas-phase basicity
(GB) of the peptides or proteins, while the deprotonation of peptides and proteins is
endoergic [20].
(b) The matrix: The key role in a MALDI imaging experiment
As explained above, the thermodynamics of the reaction depends on the type of
matrix employed. Each matrixanalyte system exhibits different proton-transfer
energetics. Therefore, different matrices are used for different types of analytes.
Figure 5.3 shows MALDI matrices and their structures. Three matrices are
commonly used: 3,5-dimethoxy-4-hydroxycinnamic acid (sinapinic acid, SA),
-cyano-4-hydroxycinnamic acid (HCCA), and 2,5-dihydroxybenzoic acid (DHB).
SA is specially used as a more energetic matrix for the detection of proteins between
2 and 75 kDa. However, SA generates many derived ions at a low mass range, which
results in a large amount of chemical noise [20]. Thus, for peptide ion detection,
HCCA or DHB is often used. These cold matrices do not generate as much derived ion
signal at low mass and allow easy mass spectral analysis and image reconstructions..
Recently, ionic matrices have been applied [21], to improve the signal intensities
and crystal size homogeneity. These ionic matrices are obtained by a reaction
between an acid matrix and a base (aniline). The comparison of images between
classical matrix and ionic matrix HCCA/ANI positions them as an alternative to
classical matrices (Fig. 5.4).

5.3.2 SIMS Source

Secondary ion mass spectrometry has been used as a high-spatial-resolution analyti-

cal microscopy technique since its inception as an imaging technique. Developed by
Castaing and Slozdian in 1962 [2], a broad beam (0.5-mm diameter of Ar+ primary
ion beam) was used to desorb and ionize sample in a secondary ion and create an
image resulting in a 1-m spatial resolution. In the microscope mode, the ultimate
spatial resolution of mass-resolved images is comparable to that of the best optical
microscopes (0.5 m). This limitation is imposed by the energy spread of the sec-
ondary ions, causing imaging chromatic aberrations. The microprobe scanning
mode, which was developed to be an alternative SIMS imaging mode [22], and the
introduction of high-brightness liquid ion sources (LMIS) revolutionized SIMS
imaging mass spectrometry in the 1980s. These ion sources, with suitable apertures
and focusing ion optics, can be focused to small spot sizes (in the best case 20 nm)
while retaining sufficient ion current for high-spatial-resolution microprobe experi-
ments [23]. SIMS coupled with a time-of-Flight (ToF) analyzer provides a tech-
nique with high spatial resolution and high mass-resolving power and allows for the
simultaneous detection of all species.
106 J. Stauber and R.M.A. Heeren

Fig. 5.3 Structures, chemical names, trivial names, and abbreviations of frequently used MALDI
matrices (From Ref. [15])

(a) Enhancing molecular ion yield in SIMS

SIMS is one dominant technique for surface analysis and imaging by mass spec-
trometry of small molecules [24]. Here, we discuss strategies that can enhance the
applicability of SIMS to biological tissue surfaces. These innovative strategies,
which include matrix-enhanced SIMS and metal-assisted SIMS, complement the
MALDI-based imaging approaches.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 107

Fig. 5.4 Comparison of HCCA classical matrix (bottom) and HCCA/ANI ionic matrix (top). The
HCCA/ANI matrix shows higher signal intensities than the classical HCCA matrix [21]

The sputtering process in SIMS, generated by high-energy primary ions, results

in fragmentation of almost all labile components on the surface. Increased detection
sensitivity for these labile intact molecular ions has been a major research topic in
organic SIMS for more than 10 years. Several strategies have been developed to
minimize the internal energy deposition during desorption and ionization in SIMS.
The use of polyatomic primary ion sources, such as C60+, SF6-, Aun, and Binm+,
results in an increased secondary ion yield [25, 26]. Here, a beam of energetic clus-
ter ions (up to several thousand electron volts) is directed at the surface. The cluster
ions are believed to dissociate at the moment of surface impact, which results in the
redistribution of the initial kinetic energy over the atoms present in the cluster [26,
27]. ME-SIMS (matrix-enhanced SIMS [28]) and MetA-SIMS (metal-assisted
SIMS [3, 29]) have also been developed to increase the dynamic mass range, as
outlined below.
(b) Matrix-enhanced SIMS
Odom and Wu reported the first ME-SIMS results in 1996. In their study, the
sample was prepared in a solid organic matrix similar to sample preparations used
for MALDI [28]. Over the past decade, ME-SIMS has been applied to a variety of
biological studies [3, 5, 30]. This technique requires specific sample-preparation
steps in order to optimize the signal intensity and minimize the redistribution or
modification of the sample analytes.
The idea of diluting analyte molecules in a solid matrix preceded the invention
of MALDI. Michl and coworkers analyzed small organic molecules by SIMS with
frozen rare gas (Ar) as a matrix [31]. Other groups investigated the use of carbon as
a matrix, which is particularly helpful in the analysis of polycyclic aromatic com-
pounds [32, 33]. Barber and coworkers also popularized the use of glycerol matrices
for ME-SIMS. In comparison with solid matrices, the liquid matrix refreshes the
sample by evaporation and/or macroscopic flow under ion beam bombardment con-
ditions [34]. Gillen et al. reported enhanced secondary ion signals for small mole-
cules embedded in frozen glycerol matrix [35] and also studied gelatin matrices as
a model of biological samples for secondary ion emission [36]. In addition, Cooks
and coworkers embedded samples in ammonium chloride [37]. Similarly, Wu and
108 J. Stauber and R.M.A. Heeren

Odom applied a solid organic matrix similar to sample preparations used in MALDI
for high-mass detection [28]. The co-crystallization of the matrix on the sample
surface permits the observation of molecular ions using static secondary ion mass
spectrometry (SSIMS). Matrix-enhanced SIMS yields substantial increases in the
ionization efficiency of peptides, proteins, and oligonucleotides, enabling the detec-
tion of species with masses up to ~10,000 Da.

5.3.3 Characteristic of Ionization, Differences Between SIMS

and Matrix-Enhanced SIMS

Cooks and coworkers report that the use of a matrix such as NH4Cl results in the
ejection of ions with a lower internal energy since fewer fragmentation products are
observed in the mass spectra [37, 38]. The authors proposed that analyte and matrix
molecules are sputtered from the surface in the form of clusters. After emission,
these clusters cool via evaporation, releasing analyte molecules with a low internal
energy.
Other authors [28, 39] have compared a series of MALDI matrices and biological
analytes with various masses (MW: 1,75914,000 Da) in order to assess positive/
negative matrix and the ion analyte yields. It is not clear what proportion of the effi-
ciency enhancement provided by the matrix is attributable to the dynamics of desorp-
tion and to the chemistry of the molecular ion-emission ionization steps. It is proposed
that the matrix plays an essential role as a strong proton donor, such as those used as
MALDI matrices [37]. In the ME-SIMS process, the matrix should efficiently pro-
mote the cationization of neutral molecules sputtered from a solid mixture by an
energetic ion beam, without photochemical reactions initiated by UV laser photons.
In addition, a large number of molecules are sputtered per ion impact, and the analyte
appears to be naturally entrained by the sputtered matrix flux. Moreover, the use of a
matrix/analyte mixture yields less fragmentation, which is attributed to the excess
matrix in the environment, which provides a much softer yet efficient desorption and
ionization of the compound [28]. Finally, the matrix/analyte mixture has a sufficiently
high concentration of analytes in the top monolayers of the surface.

5.3.4 Characteristic of Metal-Assisted SIMS

Surface metallization (MetA-SIMS) was also developed to enhance the desorption/

ionization of higher-molecular-mass species [40]. For MetA-SIMS, a submonolayer
coverage of a noble metal such as Au or Ag is applied to the surface of a material.
Silver and gold metallization demonstrated an enhancement of nearly two orders of
magnitude compared with traditional SIMS methods for the detection of polysty-
rene, a common organic polymer [29]. Metallization is especially useful for increas-
ing the sensitivity for SIMS imaging of thick organic samples. Nevertheless,
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 109

ME-SIMS is still superior for the analysis of lipids and peptides. Several explanations
for the sputtering mechanism of MetA-SIMS have been proposed. Metal cluster and
nanoparticle formation may enhance the molecular ion yield and signal detection,
by a number of proposed processes. First, the metallization of sample may eliminate
sample charging due to the conducting paths created at the surface by the gold pat-
tern. The elimination of charging effects induces a better-quality image [41].
Second, the metal clusters sitting on the surface of the sample can constitute a
matrix that enhances desorption/ionization yield. Gold embedded in the sample acts
as a cationing agent and may improve ion formation. Third, the metal evaporation
by the primary ion beam may induce migration of mobile analytes onto metal
nano-islands [29].

5.4 Analyzers to Improve IMS Capabilities

The mass spectrometry analytical technique combines sources and analyzers, which
are interlocked and developed to increase the sensitivity and the number of parallel
ion-detection events. Table 5.1 describes the different mass analyzer characteristics
used for SIMS and MALDI imaging mass spectrometry. Different geometries of
time-of-flight analyzers (linear, orthogonal, delayed extraction) and innovative
mass analyzer instruments, such as quadrupole-TOF, ion mobility-quadrupole-TOF,
and Fourier transform ion cyclotron resonance (FT-ICR), supply new perspectives
and applications in imaging mass spectrometry [4245].
In the early development of SIMS, the dynamic mode, which utilizes high-
primary-ion current densities on the sample, was used due to limited sensitivity. The
characteristic effect of the dynamic SIMS mode is typically increased erosion rates
equivalent to the removal of up to several hundred monolayers of sample per sec-
ond. Therefore, the instruments offered high detection efficiencies. Nevertheless,
the dynamic mode is not suitable for the surface analysis of small areas and for long
experiments. Thus, the static mode was developed by reducing the primary ion cur-
rent density. At the first stage of static SIMS development, quadrupole and sector
analyzers were combined to the SIMS source [46, 47]. These analyzers have a low
transmission ratio (ratio of ions leaving a region of a mass spectrometer to the num-
ber of ions entering that region) and operate in a scanning mode. These properties
resulted in limited sensitivity and unique m/z ion detection (low mass-resolving

Table 5.1 Characteristic performances of different mass analyzers

Mass-resolving
Analyzer power m/z Range Transmission Detection Sensitivity Rep. rate
Quadrupole 102103 103 0.010.1 Sequential 1 <1 Hz
Sector 104 >104 0.10.5 Sequential 10 <1 Hz
Time-of-flight 103104 105 0.51.0 Parallel 104 >10 kHz
FT-ICR 104106 106 0.20.9 Sequential Zmol <1 Hz
110 J. Stauber and R.M.A. Heeren

power), which results in the loss of important information for imaging MS

experiments. The increased use of static SIMS in the 1980s [48] resulted in the
replacement of quadrupole and sector mass analyzers with time-of-flight analyzers
(ToF). TOF analyzers offer two main advantages: a theoretically unlimited mass
range and a parallel mass registration, which allows for the collection of a mass
spectrum at every image pixel (rapid data collection) [4951].

5.4.1 Time-of-Flight

Introduced by Stephens in 1946 [52], the time-of-flight analyzer has been combined
with a SIMS ion source since the 1980s and with MALDI sources since its develop-
ment in the early 1990s. ToF offers a good transmission ratio (50100%), sensitiv-
ity, dynamic mass range, and repetition rate. The first high-mass-resolving power
imaging with a ToF-SIMS instrument in a scanning microprobe mode was pub-
lished by Schwieters group in 1991 [53]. They reported secondary-ion images of a
polymer with an average molecular weight of 1,400 m/z.
Rapidly, the efficiency and the utility of ToF-SIMS imaging were demonstrated
by many experts who consider mass spectrometry the most powerful technique for
chemical analysis. Since the 1990s, the SIMS and MALDI sources have been com-
bined with different configurations of time-of-flight detectors.
Time-of-flight analyzers allow the separation of ionized accelerated molecules
according to their molecular masses. Generated by the ionization beam in the
source, ions characteristic of the surface species are accelerated by an electric field
between the conductive support (and sample) and the extraction grid (1025 kV) to
the same kinetic energy. Therefore, the ions arrive at the detector with different
speeds, which are inversely proportional to their mass over charge values. Three
main ToF analyzer geometries are defined in order to increase the sensitivity, mass
accuracy, and mass-resolving power. These entry geometries are linear, reflectron,
and orthogonal.
The linear geometry is commonly used in ToF imaging mass spectrometry and
provides the highest sensitivity. As shown in Fig. 5.5, reflector mode and a delayed
extraction time can be performed to provide an increase in sensitivity, mass-
resolving power, and mass accuracy. The reflectron is an electrostatic ion mirror
placed at the end of the drift tube. The ion mirror reflects the ions and corrects the
different velocities by focusing the ions with the same m/z value and improves the
mass-resolving power. The delayed extraction time or pulsed ion extraction corre-
sponds to a short period of time delay between firing the laser and extracting the
ions, which improves the sensitivity and mass-resolving power of a MALDI experi-
ment [54, 55]. The linear geometry with reflector mode and delayed extraction is
widely used in imaging mass spectrometry [55]. Nevertheless, delayed extraction
must be tuned for each mass and is less effective at high mass. Therefore, a second
time-of-flight geometry, orthogonal acceleration, was presented by Guilhaus and
coworkers [56]. This geometry, shown in Fig. 5.5, places the ion source orthogonal
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 111

Fig. 5.5 Mass spectrometer combined with an ion-mobility cell allows the separation of com-
pounds according to their structure and conformation. In this example, an orthogonal time-of-flight
(O) and electrostatic mirrors (R) as a reflector are used [43]

to the time-of-flight analyzer. This decoupling of the source from the mass analyzer
results in an independent mass determination. The orthogonal mass analyzer allows
a better mass accuracy in MALDI imaging as subtle height differences on the sam-
ple no longer influence the time-of-flight measured.
Previously used alone with a linear or orthogonal geometry, TOF mass analyzers
are now combined with other mass analyzers in order to improve functionalities and
capabilities. The quadrupole mass filter, developed in the mid-1950s by Wolfgang
Paul, corresponds to four parallel hyperbolic rods that apply static or oscillating
electrical fields to select ions. The quadrupole is well suited as an ion guide and
collision cell to fragment parent ions and therefore allow molecule identification.
The MALDI-Q-ToF mass spectrometer improves the capabilities of the MALDI
imaging experiment by the localization, identification, and validation of molecules
in tissue [44, 57].

5.4.2 New Instruments for Imaging Mass Spectrometry

The goal of developing imaging mass spectrometry is not to replace existing molec-
ular imaging technologies but to offer an alternative detection method and the pos-
sibility of molecule identification directly on the tissue. Mass spectrometry is the
instrument of choice for simultaneous localization and identification. To this end,
new instruments with different mass analyzers have been developed to improve the
selectivity, sensitivity, and identification of detected molecules. Fourier transform
112 J. Stauber and R.M.A. Heeren

ion cyclotron resonance (FT-ICR), microscope MALDI, and ion-mobility mass

spectrometers are presented and discussed as new instruments for MALDI imaging
mass spectrometry. FT-ICR and ion-mobility IMS increase the number of identified
molecules in tissue samples [43, 58], whereas the microscope MALDI increases
MALDI image resolution, down to 4 m [7].
The MALDIion-mobility (IM)orthogonal-time-of-flight mass spectrometer
(oTOFMS) provides a gas-phase separation before the MS experiment. IM, also
called plasma chromatography, was developed in the 1970s as a technique for ion
separation [59]. IM applications range from protein interaction studies to conformer
differentiation. Ion-mobility mass spectrometry (IMMS) consists of an applied
electric field in a buffer gas to separate individual components according to their
mobility differences (collisional cross section) followed by mass-to-charge ratio
discrimination in a time-of-flight instrument. Flight times in the mass spectrometer
are much shorter than residence times in the drift tube; thus, it is possible to record
mass-resolved ion mobilities for all ions simultaneously [60]. The combination of
IM and imaging mass spectrometry is especially well suited to discriminate lipids
and peptides with the same mass-to-charge ratio [45]. Therefore, this separation
allows simultaneous localization and identification [61].
The ultrahigh-mass-resolving power of MALDI FT-ICR reveals novel
molecular distributions that remain hidden with low-resolution mass-spectrometric
techniques [58].
The FT-ICR geometry was used in the early 1930s [62] for nuclear physics
experiments and 40 years later for applications [63, 64]. The FT-ICR mass spec-
trometer is composed of an ion cyclotron resonance cell placed in a homogeneous
high magnetic field. A Lorenz force perpendicular to the magnetic field causes the
ions to rotate with a frequency that is inversely proportional to their m/z ratio. This
frequency of rotation is measured as the induced charge on detection electrodes
located on the ICR cell. The time-domain signal is measured, amplified, and Fourier-
transformed to yield frequency-domain data and then converted to m/z [65, 66].
A longer time-domain signal yields a higher mass-resolving power and a more pre-
cise frequency determination, and thus a higher mass accuracy (Table 5.1). This
increased mass-resolving power for MALDI imaging allows the identification and
localization of more peptides and lipids in tissue. The main difference between
FT-ICR and ToF is the delay time of mass analysis. In FT-MS, this period is much
longer (ms to seconds) than in the ToF-MS (s). Tissue digestion combined with ion
mobility or FT-ICR mass spectrometry identifies a large number of peptides from
different tissue types (formalin-fixed paraffin-embedded).
The microscope mode uses ion-optical microscope elements to project the spa-
tial origin of the ions generated at the sample surface onto a two-dimensional
position-sensitive detector [7, 67]. This approach of MALDI microscope mode
imaging allows a 200-m-diameter area analysis with a high spatial resolution
(4 m). In this case, the spatial resolution is independent of the spot diameter of the
ionizing beam but depends on the quality of the ion optics and the detector. A single
analysis over a large area now results in a higher efficiency of ion production as well
as an increase in the temporal resolution. This mass spectrometer design allows for
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 113

the detection of compounds up to 4 kDa, such as lipids, peptides, and small proteins,
which are of interest in lipidomics and proteomics. These applications are detailed
later in Sect. 5.6 on the applications of surface-enhanced IMS.
Another technique describes a way to decrease the spatial resolution without
modifying the ion optics. Microstructures of silicon wafer masks with small aper-
tures were designed to decrease the laser diameter and the irradiated area. By appli-
cation of these design masks close to the sample, 30-m irradiated surfaces are
easily obtained and can reach 10 m without the loss of signal intensities [67, 68].

5.5 Sample Preparation and Protocols

The new fields of lipidomics and proteomics were born in the early 1980s when
mass spectrometry (analytical chemistry) was applied to biological samples. The
MS approach revolutionized molecular biology and biochemistry with its high sen-
sitivity and high-throughput analysis. However, MS approaches involve stringent
and precise sample preparation to avoid contaminants and false-positive detection
of biomarkers. Due to the high sensitivity and the need for reproducibility, many
congresses and initiatives are focused on standardizing the techniques [69, 70].
Sample preparation is one of the most important steps for sensitive, rational, and
coherent analysis. The conditioning of the tissue and the timescale of the analysis
are critical and were neglected in the past. Tissue samples are very sensitive to
enzyme activity and surface contamination. After dissection, enzymes are still
active and play a role in the degradation of the proteome and the transcriptome. New
methods such as fast microwave irradiation and homogeneous heating have been
used to prevent the degradation of proteins and peptides [71, 72].
Therefore, high reproducibility and standardization are essential prerequisites
for any analytical approach. Moreover, maintaining the biochemical, molecular, and
structural sample integrity is a crucial supplementary aspect for the field of IMS.
These points are discussed here, from the tissue collection to the final analysis.

5.5.1 Tissue Collection

The first step of the IMS experiment is tissue collection, which is basically the
application of the knowledge of anatomopathologists. The analysis of fresh tissue
sections is not as difficult as that of preserved tissues. Frozen tissue and formalin-
fixed and paraffin-embedded (FFPE) tissue are the two most common preparations
used for the preservation of tissue integrity while avoiding molecular degradation.
The analysis of archived specimens is widely restricted due to methylene cross-
linkage between peptides and proteins [73]. However, recent FFPE tissue prepara-
tions have been presented and applied to proteomics studies [74, 75]. These new
possibilities open the proteomics and imaging experiments to a large number of
sample bibliotheca.
114 J. Stauber and R.M.A. Heeren

5.5.1.1 Fast Frozen Tissues

Generally, after an organ has been removed from a sacrificed animal or a biopsy has
been taken, the sample is snap-frozen in isopentane (at 50 C), depending on the
volume of the tissue. Isopentane is preferred over liquid nitrogen (LN2) due to its
larger heat capacity, which prevents the formation of ice crystals in the tissue.
Cooling down larger tissue volumes (such as entire organs) is more cumbersome
due to the introduction of large temperature gradients. In those cases, LN2 is often
used. The formation of ice crystals can destroy the cellular morphology by disrupt-
ing the cell membranes. Different reports show the necessity to heat or to use micro-
wave to inhibit the enzymatic activities and avoid protein degradation just after the
animal sacrifice. In the past 10 years, an additional step of heating or microwave
irradiation has been used to inactivate peptide and protein degradation [71, 72].

5.5.1.2 FFPE Tissue

Formalin-fixed and paraffin-embedded tissues are difficult to analyze by classical

proteomics studies due to the cross-linkage between the amino and thiol groups of
amino acids. A protocol with enzyme cleavage has been used to extract, detect, and
localize protein fragments [76]. A dewaxing step with xylene and rehydration with
graded water/ethanol baths are followed by applying enzyme on the sample surface
using a microspotting technique. Enzyme spots with a 5080-m diameter allow a
reproducible and controlled amount of deposited enzyme [74, 75].

5.5.1.3 Tissue Section

The biological tissue sample is cut at a temperature of 18/20 C in a cryo-

microtome following one of the previous sample-preparation steps. Samples embed-
ding in embedding materials containing synthetic polymers such as polyethylene
glycol (PEG) and polydimethylsiloxanes (PDMS) should preferably be avoided.
These materials have a tendency to smear a thin layer of polymer over the sample
surface during the sectioning procedure. This surface pollution often leads to pref-
erential ionization and ion-suppression effects, which will reduce the resulting
image quality. Clean tissue sections with a thickness of 1015 m are immediately
deposited onto a stainless steel plate and directly analyzed or kept at 80 C. The
best extraction conditions and consequently the highest sensitivity and mass-
resolving power are obtained when using stainless steel plates. However, other
sample holders, such as a glass plate coated by a thin conductive layer, which allows
staining optical image records, can be used. Just before analysis, the plates are
warmed to room temperature and dried under vacuum. There is no delocalization of
observed compounds (on the scale of a few hundred nanometers), regardless of their
mass. Compounds such as phospholipids are unlikely to delocalize, but it is gener-
ally feared that light elements, such as Na+ or K+, can be rapidly delocalized on a
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 115

nanometer scale [1]. Nevertheless, no such movement of alkaline or organic ions

has been observed at the micrometer scale under the conditions described above. In
addition, this sample-preparation method is robust, reproducible, and easy and does
not limit the lateral resolution.

5.5.2 Surface Modifications

5.5.2.1 Organic Solvent Treatment

Organic treatment consists of a step of fast solution application (chloroform, cold

ethanol, water) on the tissue surface to remove a particular class of compounds
(lipids or peptides) according to their solubilities in the solution used. The aim of
this approach is to decrease the complexity of the sample and to avoid ion-
suppression effects during the MS analysis. Many different sample preparations
have been developed to increase sensitivity, such as washes with chloroform, meth-
anol, water, or ethanol [77, 78]. These washes must be performed with caution to
avoid any delocalization of hydrophilic peptides. Apolar solvents, such as chloro-
form or xylene, are used particularly to remove lipids and deconstruct the lipids
layers of the cell membrane to increase the detection of high-mass proteins [74, 79].
Each sample is a unique model, and parameters must be optimized sample to sam-
ple. As shown in Fig. 5.6, analyses must be validated with immunochemistry to
demonstrate that there is no delocalization of compounds after tissue preparation
and treatment [79].

5.5.2.2 Matrix Deposition

Matrix-enhanced SIMS and MALDI techniques need a matrix deposited on the sur-
face after sample sectioning and surface treatment. A large number of matrices and
deposition techniques have been tested [28]. One of the key issues associated with
ME-SIMS and MALDI is the selection of appropriate matrix material for different
analytes according to the type of biological sample. Table 5.2 shows different matri-
ces and their associated applications.
Generally, matrices are freshly prepared by dissolution in 50 % H2O and 50 %
acetonitrile to a concentration of 0.5 M. The goal of matrix application is to produce
a very thin coating (several monolayers thick) on the tissue surface. Many tech-
niques have been developed for matrix deposition without molecular delocalization,
including spin coating, spraying, acoustic deposition, microspotting, and pipette
depositing [74, 75, 85]. Pipette deposition and manual spraying were used for rapid
analysis but are not suited for reproducible and accurate imaging experiments.
However, the other techniques cited above (which have now been commercialized;
the ImagePrep vibrational sprayer by Bruker Daltonics, the Portrait 630 acoustic
spotter by Labcyte and the ChIP piezo spotting system by Shimadzu are a few
116 J. Stauber and R.M.A. Heeren

Fig. 5.6 Images obtained after immunochemistry of oxytocin peptides before (a) and after chlo-
roform treatment (b) showing no delocalization of the peptide [79]

examples of available matrix-deposition systems) deposit matrix reproducibly with

a well-known quantity and spot size (between 3080-m spot diameter). Ultimately,
when a matrix deposition system is being used to enhance the molecular signal qual-
ity, the size of the matrix crystals will determine the obtainable spatial resolution
and the sensitivity. Figure 5.7 provides an example where electrospray deposition
was utilized to produce extremely small matrix crystals that are compatible with the
spatial resolution of SIMS. Crystal sizes were smaller than 1 m, as shown in
Fig. 5.7c, allowing for the visualization of subcellular detail in these neuronal cells.
In this case of ME-SIMS, the dried droplet deposition should not be used, in order
to avoid any peptide and/or biomolecule delocalization. Matrix droplets deposited
on cryo-microtome cut tissue sections spread over the surface and crystallize. This
leads to very large matrix crystal sizes (10 m) and lateral analyte diffusion after
wetting the tissue. Matrix solutions can be sprayed or deposited with a very small
droplet diameter (90 m) to avoid diffusion of molecules over the sample.
 5

Table 5.2 Common matrices used for matrix-enhanced mass spectrometry techniques and their associated applications
Compound Name Structure Solvents Applications
DHB 2,5- Hydroxybenzoc acid Acetonitrile, water, methanol, Peptides, proteins, lipids,
acetone, chloroform glycoproteinsglycanes [30, 80]

SA 3,5-Dimethoxy-4- Acetonitrile, water, methanol, Proteins [78, 81], glycoproteins

hydroxycinnamic acid acetone, chloroform

3-HPA 3-Hydroxypicolinic Ethanol Alcalins, oligonucleotides, proteins

acid [82, 83]

HCCA a-Cyano-4-hydroxycinnamic Acetonitrile, water, ethanol, Lipids, peptides [39], proteins [84]
acid acetone

THAP 2,4,6-Trihydroxy acetophenone Ethanol Oligonucleotides, peptides

Biological Tissue Imaging at Different Levels: MALDI and SIMS
117
118 J. Stauber and R.M.A. Heeren

Fig. 5.7 Scanning electron

microscopy images of
Lymnaea stagnalis nervous
tissue (a) prior to matrix
application (scale bar 10 m)
and (b) after electrospray
deposition of 2,5-DHB
(scale bar 10 m);
(c) higher-magnification
image showing submicron
crystal dimensions
(scale bar 1 m) [86]

5.5.2.3 Metal Deposition

Metal deposition consists of a unique surface modification (MetA-SIMS) or

sometimes an additional step for MALDI ionization or ME-SIMS ionization
enhancement and will be discussed in detail in the next section.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 119

Metal deposition is realized using metal-coating techniques similar to those used

in electron microscopy. This provides a reproducible and accurate deposition of
metal to the sample surface. Usually, a thickness of 25 nm is deposited and
increases the ionization efficiency up to 3 kDa [40, 41]. The metal thickness is
determined during the deposition process using a quartz crystal microbalance
integrated in the metal-deposition system. Studies using atomic force microscopy
indicated that during 12-nm gold deposition small islands are formed on the tissue
surface. The surface becomes conductive although it is difficult to exactly assess the
conductivity of the ITOtissueAu sandwich structure. If the thickness exceeds
5 nm, many gold clusters are formed and the molecular signal is suppressed
completely. This is believed to indicate the formation of a solid, closed film that
minimizes energy deposition directly in the tissue.

5.5.3 Comparison of Metallization

and Matrix-Deposition Procedures

Metallization and matrix deposition enhance ionization and desorption, but these
processes do possess drawbacks. Here, we will recapitulate a series of criteria,
which are initiated from the personal interpretation of Delcorte [87] and applied to
MALDI and SIMS imaging experiments. The interpretation and coherence of the
imaging results are correlated to different criteria of efficiency, reproducibility,
versatility, and ease of interpretation, as summarized in Table 5.3.
Efficiency is characterized by an enhanced ionization yield, which is com-
paratively the same in MetA-SIMS and ME-SIMS. However, the dynamic mass
range of the MetA-SIMS procedure is more limited (up to ~ 3 kDa) than that of
ME-SIMS [40].
Reproducibility is the major factor for imaging techniques to obtain coherent
observations. The reproducibility of imaging MS experiments is characterized by
two main aspects: the detection and localization of identical intensity signals. The
reproducibility of matrix deposition is less effective than metal deposition due to the
difficulty to reproduce the same deposition conditions and matrix crystal size/struc-
ture [86, 88, 89]. Nevertheless, many different matrix-deposition methods have
been developed to reproduce the same quality of coverage and allow more

Table 5.3 A comparison Criterion Matrix deposition Metal deposition

between surface metallization
Efficiency ++ ++
and matrix deposition for
SIMS-based tissue imaging Reproducibility +
mass spectrometry Versatility +/ ++
Ease of interpretation +
Equipment/cost +
++ very good, + good, +/ satisfactory, limiting
120 J. Stauber and R.M.A. Heeren

reproducible images. Moreover, the crystal size (around 15 m) does not have the
same effect in SIMS and MALDI imaging, with regard to the beam diameter. In the
case of MALDI, the laser beam could reach 20 m, and the image does not have
enough spatial resolution to be affected by the crystal size. However, in SIMS a
smaller beam (500 nm) is used and quality of the images is restrained to the crystal
size. In addition, matrix deposition induces an increase in low-mass-range matrix
signals, which interfere with sample signals, unlike metal deposition.
Versatility defines the number of possible applications of SIMS and MALDI
imaging with metal or matrix deposition. Metal deposition allows a large number of
applications because of its accurate deposition for any type of sample surface with-
out any delocalization [30]. Moreover, the interpretation of their spectra is easier
because only peaks from the metal are detected [90]. In the case of a matrix-
enhanced SIMS experiment, the matrix produces interference in the spectra [91].

5.6 Applications of MALDI and Enhanced SIMS Imaging

Chemical imaging mass spectrometry provides both chemical information and the
spatial organization of each component on a surface. The development of imaging
mass spectrometry is revolutionizing the field of biological analysis [92]. SIMS and
MALDI are the two most common techniques for the characterization and localiza-
tion of compounds and provide a broad spectrum of applications.
The first analyses of biological tissue used a SIMS mass spectrometer for metal,
lipids, and peptides localizations [93] with a high spatial resolution of 0.5 m.
A high spatial resolution combined with a high sensitivity for high-mass molecules
remains the goal of chemical imaging mass spectrometry. Both MALDI- and SIMS-
based approaches have been developed to achieve these goals, as discussed above.
Three large families of molecules are implicated in the organization (structure
and activity) of cells: proteins, lipids, and carbohydrates. These biomolecules are
observed at different concentrations (106 of magnitude) in the tissue, have different
biochemical properties (molecular weight, pH iso-electric, proton affinity), and
result in a different degree of desorption and ionization efficiency. The key to suc-
cess is to apply appropriate imaging techniques (IMS and traditional imaging tech-
niques) to answer a precise question.
In this section, we present different applications of molecular imaging mass
spectrometry in the detection of molecules [such as pharmaceuticals, metabolites,
atoms (metals), lipids, peptides, and proteins], that play a role in disease or specific
conditions from plants to animal tissue. The choice of the complementary IMS tech-
nique depends on two parameters: the nature of the detected compounds (and the
correlated mass) and the spatial resolution. The surface enhanced IMS helps to
understand the role of each molecule in many different fields of research (lipido-
mics, proteomics, clinical proteomics, and pharmacology), showing a large spectra
of applications.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 121

5.6.1 Single-Cell Imaging of Peptides and Lipids

for Fundamental Biological Studies

Surface-enhanced SIMS is the suitable instrument for the detection of low-mass

molecules such as lipids and small peptides. Moreover, the detection of molecules
in ganglia samples of small animals requires a high spatial resolution; thus, SIMS
could play an important role for high spatial localization of biomolecules without
chemical labeling (as is required in immunochemistry or targeted approaches).
Recent studies have combined surface-enhanced SIMS and MALDI to detect
and localize compounds in biological samples. Heerens group developed and
applied surface analysis using ME-SIMS and MetA-SIMS for the detection of neu-
ropeptides and lipids in Lymnaea stagnalis [86] and neuroblastoma [3].
Matrix (2,5-dihydrobenzoic acid in these examples) was deposited by an
electrospray-deposition (ESD) system in order to ensure small matrix crystals com-
patible with the high spatial resolution of SIMS (Fig. 5.7). After tissue sectioning
and pickup, the Lymnaea stagnalis tissue was sectioned with a cryo-microtome,
placed on an indium-tin-oxide (ITO) coated glass slide, and coated with DHB
matrix by ESD. The tissues were analyzed on a TRIFT-II time-of-flight SIMS (ToF-
SIMS) equipped with an 115In+ liquid metal ion gun (LMIG), and on a MALDI
equipped with 337 nm nitrogen laser VSL-337i in reflectron mode.
Different peptide and lipid profile images produced from SIMS and MALDI
mass spectrometry were compared to scanning electron microscopy (SEM) images
[86]. SEM was used to determine the size of the matrix crystals, thus validating
ESD for ME-SIMS matrix deposition. The crystal size was ~1 m, which allows
imaging at subcellular resolution. ME-SIMS and MALDI spectra show peaks under
3,000 m/z (Fig. 5.8), which reveal the presence of the neuropeptide APGWamide in
the anterior lobe of the right cerebral ganglia (Fig. 5.8) in the associated images
(Fig. 5.9). This peptide regulates the male copulation behavior of Lymnaea stagna-
lis [94]. This study corroborates previous studies of matrix-enhanced SIMS and
shows the power of ME-SIMS to detect and image both lipids and localized neuro-
peptides with a high spatial resolution (1.5 m).
Surface metallization was used as an alternative to matrix deposition with SIMS
and added (in addition to the matrix) for MALDI analysis. High-resolution images
of lipids and peptides were obtained for neuroblastoma cells and rat brain tissues
with MetA-SIMS, ME-SIMS, and a homemade MALDI-BioTRIFT instrument
with a higher spatial resolution than commercially available MALDI instruments
[7, 30].
ME-SIMS increases the signal intensities for phospholipids, phosphatidyl cho-
line (PC), sphingomyelin (SM), cholesterol, ceramide, and di-acyl glycerol (DAG)
(Fig. 5.8) compared to regular SIMS analysis. MetA-SIMS further increases the
same signals (~50) and also increases the yield of other molecular signals, as
shown in Fig. 5.10.
Lateral resolution analyte diffusion is minimized in MetA-SIMS, where only
migration on gold islands at the nanometer scale may occur, which produces
122 J. Stauber and R.M.A. Heeren

Fig. 5.8 ME-SIMS vs. MALDI: (a) ME-SIMS and (b) MALDI spectra of a commissure extract
of the pond snail Lymnaea stagnalis. Both spectra were obtained using standard dried droplet
sample preparation with 2,5-DHB as matrix. The peptides identified are R-caudorsal cell peptide
(R, protonated and cationized), peptide (), HFFYGPYDVFQRDVamide (m/z 1,790), Calfluxin
(CaFl), carboxyl terminal peptide (CTP), and caudodorsal cell hormone (CDCH). The ME-SIMS
measurements used indium primary ions (total ion dose 8.9 1011 ions/cm2) and the MALDI exper-
iments 250 shots of the 4-ns 337-nm nitrogen laser [80]

better-quality images than ME-SIMS. Moreover, higher mass signals are observed
(peptides) up to 1,400 Da. The combination of the spatial resolution and the high
sensitivity for cholesterol and other ions allows the description of subcellular
organization and the differentiation of cell clusters of the neuroblastoma. The
membrane, nucleus, and intracellular compounds are observed to delineate the
neuroblastoma.
Moreover, except for the lipids signals, the spectrum shows no other molecular
signals in ME-SIMS, while signals up to 4,000 and 1,500 Da are observed from the
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 123

Fig. 5.9 Direct molecular imaging of Lymnaea stagnalis nervous tissue by ME-SIMS:: (a) optical
image of the Lymnaea cerebral ganglia; inset shows high-magnification image of neurons in the
anterior lobe (solid box); arrows indicate nuclei. Different regions in the section are right and left
cerebral ganglia (Cgr and Cgl), anterior lobe (Al), commissure (Cm), and dorsal bodies (Db); (b)
ME-SIMS image of APGWamide (429.0433.2 m/z; green) distribution; (c) ME-SIMS image of
cholesterol (368.2371.3 m/z; blue) distribution. Scale bar: 200 m; scale bar inset: 10 m.
Molecular images (b and c) are presented as colored overlays on top of the gray-scale TIC (total
ion count) image (mass range: 1.05,000 m/z) [3]
124 J. Stauber and R.M.A. Heeren

Fig. 5.10 Cellular localization of MetA-SIMS selected ion count signals from neuroblastoma
cells. Cells were imaged after deposition of 1-nm gold: (a) m/z 369 (cholesterol [MOH]+, 014
counts) and 607 (DAG, 06 counts); (b) m/z 970 (cholesterol [2 M + Au]+, 04 counts) and 1,080
(01 counts); (c) m/z 369 (cholesterol [MOH]+, 012 counts) and 895 (02 counts) [3]

same section by stigmatic MALDI and MetA-SIMS. These results are explained by
the major difference between the two approaches, which relates to the amount of
sample consumed. In MALDI, each laser shot consumes approximately 100 nm1 m
of sample, while the penetration depth of the SIMS is typically 10 nm. This is
explained by the crystallization and segregation of lipids on the crystal surface of the
matrix, while the peptides are mainly incorporated into the matrix [95].
Sample preparations of tissue with chemical modifications used with the
MALDI-BioTRIFT technique show high-mass signals up to 4,000 Da, with a higher
spatial resolution (approximately 2 m). Gold deposition results in an enhanced
image quality and signal intensity, not only for SIMS but also for MALDI.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 125

Fig. 5.11 MALDI stigmatic imaging of a rat brain tissue section: (a) a line scan summed mass
spectrum showing gold cluster peaks and several peptide peaks, with the vasopressin mass at m/z
1,085 (*); (b) TIC image (gray scale) of an HCCA-coated rat brain tissue section overlaid with the
selected ion image for vasopressin (green). Scale bar: 1 mm [3]

Figure 5.11 shows images of a localized vasopressin (1,085-Da) protein obtained

from matrix-enhanced MALDI.
ME- and MetA-SIMS describe localized low-mass signals from lipids and pep-
tides by the assistance of chemical modifications. The matrix deposition for
ME-SIMS is the prerequisite for MALDI analysis as well. MALDI allows the mul-
timolecular imaging of peptides and proteins up to 75 kDa. The large dynamic mass
range allows for the analysis of complex mixtures of compounds in tissues and has
many biological applications. Applications include imaging of organs or whole ani-
mal body slices as well as vegetable models [11, 96], with the detection and local-
ization of drugs, metabolites, peptides, proteins, and carbohydrates.
126 J. Stauber and R.M.A. Heeren

5.6.2 Peptide and Protein Detection: MALDI Imaging

Applications

The first direct image construction by MALDI was initiated by Capriolis group
[11]. At this juncture, in 1997, MALDI was described as a technique to analyze a
biological solution with proteins and peptides. Different groups have applied this
technique to a direct analysis of tissue by matrix droplet deposition followed by
MALDI-ToF MS analysis [9799]. Different average spectra for regions of interest
(ROI) are compared and allow differentiation of the proteome and lipidome in cer-
tain regions.
In the past decade, many developments for sample preparation, instrumentation
optimization, and software have enabled MALDI imaging at a subcellular resolu-
tion with good sensitivity [80].

5.6.3 Proteomics and Clinical Proteomics

No specific sample-preparation methods were used for the first MALDI imaging
experiments. The matrix was applied directly with a pipette or spayed onto the tis-
sue in order to make a reproducible procedure to detect some standard peptides and
proteins [11, 100, 101]. The choice of a spraying system or microspot deposition of
matrix is crucial for the spatial resolution and the reproducibility of experiments.
The mass range detectable with MALDI imaging is approximately 080 kDa. The
first application of MALDI on a biological surface was to detect large molecules for
proteomics studies (Fig. 5.12) [102]. In this case, human glioblastoma cells were
implanted into a hind limb of a nude mouse model and specific markers (T.4 pro-
tein, m/z = 43,965) were observed in the tumor proliferation area (Fig. 5.12). These
first publications illustrate the interest in MALDI imaging for proteomics and clini-
cal proteomics research: detecting proteins implicated in disease pathologies, that
discriminate different cell or organ states.
The work of the Sweedler group illustrates the usefulness of MALDI imaging in
the neuroscience field [103105]. The Sweedler group has developed methodolo-
gies and sample preparations for invertebrate ganglia cell imaging. The application
of matrix onto the Aplysia californica exocrine gland and neuronal tissue shows the
ability to spatially image neuropeptides and proteins. Two different approaches
have been used. First, a microspotted matrix deposition on different ganglia slides
(every 30 m) demonstrated the ability to profile neurons without removing cells
from the ganglion matrix from the ventral to the dorsal side. Another approach uses
total laser ablation of the matrix and moves the plate a distance (25 m) smaller than
the laser spot size, a so-called oversampling approach. In this manner, even with a
large laser spot, a small analytical area can be studied, improving the spatial resolu-
tion of the resulting image. The use of an adequate bioinformatics tool to recon-
struct the image allows the imaging of molecules with a resolution of 50 m.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 127

Fig. 5.12 Selected protein images from a glioblastoma section: (a) human glioblastoma slice
mounted on a metal plate, coated with matrix (the lines are from ablation of matrix with the laser);
(bd) mass-spectrometric images of proteins showing a high concentration in the proliferating
area of the tumor (d) and other proteins present specifically in the ischemic and necrotic areas
(b and c) [102]

At the same time, other groups have applied MALDI imaging to different
diseases by integrating sample comparisons [78, 81, 106, 107]. As illustrated in
these papers, MALDI mass-spectrometric imaging on tissues is used in endogen
biomarker discovery by determining under- and overexpressed peptides/proteins of
a disease state versus a healthy control (AD). MALDI mass-spectrometric imaging
has also been applied to the study of amyloid beta peptide distribution in brain
sections from mice and shows features reminiscent of Alzheimers disease.

5.6.4 Drug Distribution and Quantification

IMS has also been used to observe the elimination and repartition of drug distribu-
tions on tissue sections [44, 78, 96, 108111]. The drug profile and ADME (adsorp-
tion, distribution metabolism, and excretion) studies appear to be particularly well
128 J. Stauber and R.M.A. Heeren

Fig. 5.13 Drug distribution of erlotinib in rat liver (a, b) and spleen (c, d) tissue section. (a) and
(c) represent the optical image, while (b) and (d) represent the MALDI image of erlotinib at
m/z = 278.1. (e) represents the dilution series of erlotinib on target plate as a calibration (the HCCA
matrix was deposited by hand) [78]

suited with IMS [57, 109, 112]. In these articles, different drugs are injected into the
animals, and monitored by single-MS or tandem-MS imaging to visualize the drug
incorporation. Signor et al. compared the quantification of erlotinib by autoradiogra-
phy, LC-MS, and MALDI imaging experiments. As shown in Fig. 5.13, at a concen-
tration of 3.76 ng/mg, erlotinib and its metabolites are observed in the rat liver and
spleen sections. The comparison of quantification techniques (autoradiography and
LC-MS/MS) was also similar with MALDI analysis as a quantitative approach [78].
However, signal intensities may vary even in the same sample; thus, MALDI
MS is not the method of choice for absolute quantification. In this study [112], the
drug distributions and the intensity ratios could be reproduced within a range of
20 % (if the sample-preparation procedure as well as instrument settings and laser
intensity are kept constant). An internal calibration and reproducible deposition of
matrix enable relative quantification with the comparison of peak areas instead of
height peaks.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 129

Several groups have shown that this method allows a quantification error
variation of about 23 % [96, 113, 114]. New methods such as multiple reaction
monitoring (MRM)based imaging on triple quad systems provide a rapid, more
sensitive targeted analysis of drug distributions in whole body tissue sections [115].
Although these methods are relevant to quantify sample-to-sample variations, vali-
dation processes are still used with established methods such as Western blot and
immunochemistry for determining accurate concentrations. It is necessary to com-
bine IMS with HE staining and immunohistochemistry to validate IMS as an
approach for the identification of molecules from tissue [116]. Thus, a validation
method was developed with a tag-mass, called specific mass spectrometry imaging.
This approach allows the simultaneous detection and validation of molecules of
interest using IMS [81, 83, 107].

5.6.5 Specific Mass Spectrometry Imaging

Specific mass spectrometry imaging is used to image different types of molecules,

such as proteins and nucleotides, with a combination of mass spectrometry imaging
and immunohistochemistry. A tag-mass is used, which is a tagged antibody or
aptamer that can be observed in the IMS experiment. This concept was developed
to detect compounds such as nucleotides and large proteins, which are difficult to
ionize by classical IMS experiments. A specific mass spectrometry imaging experi-
ment is shown in Fig. 5.14 along with the different steps of tag-mass multiplexing.

5.6.6 Application with Carbohydrates and Metabolites

in Plants

IMS applications in the vegetable kingdom, including MALDI imaging of oligosac-

charides and primary metabolites in a plant system, have recently been presented
[117]. MALDI imaging was used to identify metabolites, namely, glucose-6-
phosphate, in potato tubers [118] and to determine agrochemical compounds in
soybeans [119]. In these examples, the sensitivity of MALDI imaging and the semi-
quantitative evaluation allow the detection of hundreds of metabolites with a con-
centration of around 1 g/g. Other approaches, including electrospray mass
spectrometry and enzyme-linked assay, could be combined with IMS, but the sensi-
tivity of IMS offers the potential of simultaneous metabolite assays. Nevertheless,
there are over 100,000 plant metabolites, and MS/MS is necessary for accurate
identification. Combined IMS and MS/MS approaches have been developed for fro-
zen, fixed, and embedded tissues [120]. These methods will be discussed in the
section entitled Perspectives ahead.
Recently, Thomas-Oates published a study on the localization of carbohydrate
metabolites as an indicator of grain yield. Cross and longitudinal sections from the
130 J. Stauber and R.M.A. Heeren

Fig. 5.14 Specific mass spectrometry imaging approach uses a tagged reporter, which allows the
multiplex detection of a large number of molecules, including nucleotides and proteins. In this
example, the tag-mass is a complementary oligonucleotidic sequence, which has a high affinity to
the complementary sequence of interest in the tissue. This affinity and the IMS experiment allow
the localization of the interest sequence of DNA, RNA, or proteins in a more general aspect [83]

wheat stems of Triticum aestivum were used to localize water-soluble carbohydrates

[121]. They demonstrated the sensitivity of IMS and verified the results using com-
plementary techniques, including LC-MS. An advantage of IMS over the other
techniques included the ability to determine in situ localization.

5.6.7 MALDI Imaging of Lipids

Lipid distributions are usually detected and mapped by SIMS mass spectrometry,
especially because of the sensitivity to SIMS in the low-mass range. Lipids can also
be detected and localized by MALDI with special matrices to minimize interfer-
ence from matrix peaks. In this context, the MALDI ion-mobility mass spectrom-
etry imaging technique was developed. MALDI-Q-ToF, coupled with an
ion-mobility cell, developed by the groups of Woods and Schultz, allows for the
separation of lipids, matrix, and peptide species [43, 45, 122]. This separation
allows a differentiation between species, with a higher signal-to-noise ratio, as
shown in Fig. 5.15.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 131

Fig. 5.15 Ion-mobility diagram obtained by MALDI-IM imaging of a rat brain section. The y-axis
is the time separation in the drift cell, and the x-axis is mass-to-charge ratio. Different species
(here, DHB matrix, peptides, and lipids) fall on unique trend lines [45]

5.6.8 Element/Metal Detection

One other dedicated application of IMS is the localization of contrast agents used
for MRI imaging [123125]. Contrast agents, used to increase the contrast of images
during in vivo imaging by MRI or PET instruments, were studied by IMS to accu-
rately localize their distributions in organs and cells with a better sensitivity and
spatial resolution than classical MRI. In these examples, IMS allowed the detection
of contrast agents in their native or degraded form [126].

5.7 Perspectives

As presented here, surfaced-enhanced SIMS and MALDI imaging are two recent
approaches to detect and image macromolecular compounds directly in biological
samples. Important recent developments include sample-preparation protocols and,
in particular, instrument combinations of SIMS and/or MALDI ionization sources
with different mass analyzers. These developments are correlated to the same devel-
opments observed in classical proteomics studies many years ago. In that sense, it is
easy to foresee the inclusion of IMS into biological studies.
132 J. Stauber and R.M.A. Heeren

In a general view, IMS is able to visualize different compounds with a spatial

resolution of 50 nm100 m with good sensitivity. Nevertheless, the accurate iden-
tification and validation of these compounds is still difficult. In this way, many dif-
ferent approaches, such as immunochemistry and autoradiography, validate the
presence of biomolecules. In the past decade, different tandem mass spectrometry
(MS/MS) approaches were developed with instruments with new geometries to
directly identify molecules in tissue [127]. These new geometries allow a better
separation and possible in situ identification of compounds such as lipids and
digested proteins. Thus, two groups have developed in situ enzymatic digestion to
map large proteins in frozen and formalin-fixed and paraffin-embedded (FFPE) tis-
sues [74, 75, 128].
Trypsin digestion (from classical proteomics experiments) on FFPE tissues
allows proteomics studies on archived tissues from hospital libraries [76]. This
method of digestion requires a micro-deposition instrument to avoid any delocal-
ization of digested peptides, where trypsin deposition comes before matrix deposi-
tion. Digestion on tissue allows the identification and localization of thousands of
digested peptides. Nevertheless, it is difficult to identify all of the fragments because
of the high concentration of peptides and lipids in the mass range of 350950 m/z.
In this perspective, digested FFPE and frozen tissues were analyzed with an ion-
mobility mass spectrometer [128]. This approach separates molecules (e.g., pep-
tides, drugs, and proteins) according to their conformation (cross section) and
offers better identification. Moreover, MALDI ion-mobility imaging allows the
separation of drugs from lipids, saccharides, nucleotides, and peptides. The use of
standard protocols, coupled with sensitive instruments that can separate ions in the
gas phase prior to mass analysis, could be the way to improve the field of imaging
mass spectrometry.

5.8 Conclusion

As shown in this chapter, surface-enhanced imaging mass spectrometry and IMS

are technologies based on fundamental innovations in physics and chemistry.
Their applicability to biological and biomedical studies has been shown. SIMS
and MALDI combined offer complementary surface-analysis approaches to study
a range of species, from elements to proteins (Fig. 5.16). Applications include
biomarker discovery or drug assessment (ADMET: absorption, distribution,
metabolism, elimination, toxicology) on biopsy surfaces. As a result, this method
can be involved in each step of drug discovery: for target identification, lead opti-
mization, assessment of drug-delivery systems, and profiles of the biodistribution
and metabolism.
These young techniques are under development to increase their sensitivity,
speed of analysis, reproducibility, and spatial resolution. The numbers of papers
published recently demonstrate the power and potential of these approaches to over-
come classical imaging problems. Initially developed to localize lipids, peptides,
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 133

Fig. 5.16 Description of the different enhanced-surface IMS techniques according to their spatial
resolution, mass-range detection, and sensitivity

and proteins, MALDI imaging and surface-enhanced SIMS allow the detection of
biomarkers related to specific tissue conditions.
Imaging mass spectrometry is a label-free imaging technique for molecular
pathology that complements existing, less specific imaging modalities in biomedi-
cal research. This innovative MS-based technique, together with such techniques as
MRI, CT, PET, or ICC imaging, will visualize the fundamentals of diseases in future
biomedical imaging research.

References

1. McDonnell LA, Heeren RMA. Imaging mass spectrometry. Mass Spectrom Rev. 2007;26:
60643.
2. Castaing R, Slodzian G. Microanalyse par emission ionique secondaire. J Microsc. 1962;1:
395410.
3. Altelaar AFM, Klinkert I, Jalink K, De Lange RPJ, Adan RAH, Heeren RMA, Piersma SR.
Gold-enhanced biomolecular surface imaging of cells and tissue by SIMS and MALDI mass
spectrometry. Anal Chem. 2006;78:73442.
4. McDonnell LA, Heeren RMA, de Lange RPJ, Fletcher IW. Higher sensitivity secondary ion
mass spectrometry of biological molecules for high resolution, chemically specific imaging.
J Am Soc Mass Spectrom. 2006;17:1195202.
5. Altelaar AFM, Luxembourg SL, McDonnell LA, Piersma SR, Heeren RMA. Imaging mass
spectrometry at cellular length scales. Nat Protoc. 2007;2:118596.
6. Heeren RMA, McDonnell LA, Amstalden ER, Altelaar AFM, Piersma SR. Why dont biolo-
gists use SIMS; a critical evaluation of imaging MS. Appl Surf Sci. 2006;252:682735.
7. Luxembourg SL, Mize TH, McDonnell LA, Heeren RMA. High-spatial resolution mass
spectrometric imaging of peptide and protein distributions on a surface. Anal Chem. 2004;76:
533944.
134 J. Stauber and R.M.A. Heeren

8. Klerk LA, Lockyer NP, Kharchenko A, MacAleese LP, Dankers PYW, Vickerman JC, Heeren
RMA. C60+ secondary ion microscopy using a delay line detector. Anal Chem. 2010;82:
8017.
9. Karas MI, Bachmann D, Bahr U, Hillenkamp F. Matrix-assisted ultraviolet laser desorption
of non-volatile compounds. Int J Mass Spectrom. 1987;78:5368.
10. Hillenkamp F, Karas M. Matrix-assisted laser desorption/ionisation, an experience. Int
J Mass Spectrom. 2000;200:717.
11. Caprioli RM, Farmer TB, Gile J. Molecular imaging of biological samples: localization of
peptides and proteins using MALDI-TOF MS. Anal Chem. 1997;69:475160.
12. Stoeckli M, Farmer TB, Caprioli RM. Automated mass spectrometry imaging with a matrix-
assisted laser desorption ionization time-of-flight instrument. J Am Soc Mass Spectrom.
1999;10:6771.
13. Dreisewerd K, Schrenberg M, Karas M, Hillenkamp F. Influence of the laser intensity and
spot size on the desorption of molecules and ions in matrix-assisted laser desorption/ioniza-
tion with a uniform beam profile. Int J Mass Spectrom. 1995;141:12748.
14. Posthumus MA, Kistemaker PG, Meuzelaar HLC, Ten Noever de Brauw MC. Laser
desorption-mass spectrometry of polar nonvolatile bio-organic molecules. Anal Chem. 1978;
50:98591.
15. Zenobi R, Knochenmuss R. Ion formation in MALDI mass spectrometry. Mass Spectom Rev.
1998;17:33766.
16. Zhigilei LV, Garrison BJ. Molecular dynamics simulation study of the fluence dependence of
particle yield and plume composition in laser desorption and ablation of organic solids. Appl
Phys Lett. 1999;74:13413.
17. Ehring H, Karas M, Hillenkamp F. Role of photoionization and photochemistry in ionization
processes of organic molecules and relevance for matrix-assisted laser desorption ionization
mass spectrometry. Org Mass Spectrom. 1992;27:47280.
18. Preston-Schaffter LM, Kinsel GR, Russell DH. Effects of heavy-atom substituents on matri-
ces used for matrix-assisted laser desorption-ionization mass spectrometry. J Am Soc Mass
Spectrom. 1994;5:8006.
19. Knochenmuss R, Zenobi R. MALDI ionization: the role of in-plume processes. Chem Rev.
2003;103:44152.
20. Breuker K, Knochenmuss R, Zhang J, Stortelder A, Zenobi R. Thermodynamic control of
final ion distributions in MALDI: in-plume proton transfer reactions. Int J Mass Spectrom.
2003;226:21122.
21. Lemaire R, Tabet JC, Ducoroy P, Hendra JB, Salzet M, Fournier I. Solid Ionic matrixes for
direct tissue analysis and MALDI imaging. Anal Chem. 2006;78:80919.
22. Liebl H. Ion microprobe mass analyzer. J Appl Phys. 1967;38:527783.
23. Chabala JM, Soni KK, Li J, Gavrilov KL, Levi-Setti R. High-resolution chemical imaging
with scanning ion probe SIMS. Int J Mass Spectrom. 1995;143:191212.
24. Benninghoven A. Die Analyse monomolekularer Festkrperoberflchenschichten mit Hilfe
der Sekundrionenemission. Z Physik. 1970;230:40317.
25. Appelhans AD, Delmore JE. Comparison of polyatomic and atomic primary beams for sec-
ondary ion mass spectrometry of organics. Anal Chem. 1989;61:108793.
26. Winograd N. The magic of cluster SIMS. Anal Chem. 2005;77:142A9.
27. Gillen G, Fahey A. Secondary ion mass spectrometry using cluster ion beams. Appl Surf Sci.
2002;203:20913.
28. Wu KJ, Odom RW. Matrix-enhanced secondary ion mass spectrometry: a method for molec-
ular analysis of solid surfaces. Anal Chem. 1996;68:83782.
29. Delcorte A, Bour J, Aubriet F, Muller JF, Bertrand P. Sample metallization for performance
improvement in desorption/ionization of kilodalton molecules: quantitative evaluation, imag-
ing secondary ion MS, and laser ablation. Anal Chem. 2003;75:687585.
30. McDonnell LA, Piersma SR, Altelaar AFM, Mize TH, Luxembourg SL, Verhaert PDEM, van
Minnen J, Heeren RMA. Subcellular imaging mass spectrometry of brain tissue. J Mass
Spectrom. 2005;40:1608.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 135

31. Jonkman HT, Michl J, King RN, Andrade JD. Low-temperature secondary positive ion mass
spectrometry of neat and argon-diluted organic solids. Anal Chem. 1978;50:207882.
32. Ross MM, Colton RJ. Summary abstract: secondary ion mass spectrometry of organic adsor-
bates on carbon particles and liquid metal surfaces. J Vac Sci Technol A. 1983;1:4412.
33. Ross MM, Colton RJ. Carbon as a sample substrate in secondary ion mass spectrometry. Anal
Chem. 1983;55:1503.
34. Barber M, Bordoli RS, Sedgwick RD, Tyler AN. J Chem Soc Chem Comm. 1981;325327.
35. Gillen G, Christiansen JW, Tsong IST, Kimball B, Williams P, Cooks RG. Sputter yields of
ammonium chloride and solid glycerol. Rapid Commun Mass Spectrom. 1988;2:678.
36. Bennett J, Gillen G. Formation and emission of tetraalkylammonium salt molecular ions
sputtered from a gelatin matrix. J Am Soc Mass Spectrom. 1993;4:9307.
37. Busch KL, Hsu BH, Xie YX, Cooks RG. Matrix effects in secondary ion mass spectrometry.
Anal Chem. 1983;55:115760.
38. Liu LK, Busch KL, Cooks RG. Matrix-assisted secondary ion mass spectra of biological
compounds. Anal Chem. 1981;53:10913.
39. Wittmaack K, Szymczak W, Hoheisel G, Tuszynski W. Time-of-flight secondary ion mass
spectrometry of matrix-diluted oligon- and polypeptides bombarded with slow and fast pro-
jectiles: positive and negative matrix and analyte ion yields, background signals, and sample
aging. J Am Soc Mass Spectrom. 2000;11:55363.
40. Delcorte A, Medard N, Bertrand P. Organic secondary ion mass spectrometry: sensitivity
enhancement by gold deposition. Anal Chem. 2002;74:495568.
41. Delcorte A, Bertrand P. Interest of silver and gold metallization for molecular SIMS and
SIMS imaging. Appl Surf Sci. 2004;2312:2505.
42. Taban IM, Altelaar AFM, Fuchser J, van der Burgt YEM, McDonnell LA, Baykut G, Heeren
RMA. Imaging of peptides in the rat brain using MALDI-FTICR mass spectrometry. J Am
Soc Mass Spectrom. 2007;18:14551.
43. Tempez A, Ugarov M, Egan T, Schultz JA, Novikov A, Della-Negra S, Lebeyec Y, Pautrat M,
Caroff M, Smentkowski VS, Wang H-YJ, Jackson SN, Woods AS. Matrix implanted laser
desorption ionization (MILDI) combined with ion mobility-mass spectrometry for bio-
surface analysis. J Proteome Res. 2005;4:5405.
44. Bunch J, Clench MR, Richards DS. Determination of pharmaceutical compounds in skin by
imaging matrix-assisted laser desorption/ionisation mass spectrometry. Rapid Commun Mass
Spectrom. 2004;18:305160.
45. Jackson SN, Ugarov M, Egan T, Post JD, Langlais D, Schultz JA, Woods AS. MALDIion
mobilityTOFMS imaging of lipids in rat brain tissue. J Mass Spectrom. 2007;42:10938.
46. Mller A, Benninghoven A. Investigation of surface reactions by the static method of second-
ary ion mass spectrometry: III. The oxidation of vanadium, niobium and tantalum in the
monolayer range. Surf Sci. 1973;39:42736.
47. Plog C, Wiedmann L, Benninghoven A. Empirical formula for the calculation of secondary
ion yields from oxidized metal surfaces and metal oxides. Surf Sci. 1977;67:56580.
48. Standing KG. Timing the flight of biomolecules: a personal perspective. Int J Mass Spectrom.
2000;200:597610.
49. Wirth A, Thompson G, Gregory SP, editors. In: Benninghoven A, editor. Secondary ion mass
spectrometry SIMS VI. New York: Wiley;. 1987. 639 pp.
50. Waugh AR, Kingham DR, Hearn MJ, Briggs DA, editors. In: Benninghoven A, editor.
Secondary ion mass spectrometry SIMS VI. New York: Wiley; 1987. 231 pp.
51. Mullock SJ, Reich DF, Dingle T, editors. In: Secondary ion mass spectrometry SIMS VII.
New York: Wiley; 1989. 847 pp.
52. Stephens WE, Serin B, Meyerhof WE. A method for measuring effective contact e.m.f.
between a metal and a semi-conductor. Phys Rev. 1946;69:42.
53. Schwieters J, Cramer HG, Heller T, Jurgens U, Niehuis E, Zehnpfenning J, Benninghoven A.
High mass resolution surface imaging with a time-of-flight secondary ion mass-spectroscopy
scanning microprobe. J Vac Sci Technol A. 1991;9:286471.
136 J. Stauber and R.M.A. Heeren

54. Brown RS, Lennon JJ. Mass resolution improvement by incorporation of pulsed ion
extraction in a matrix-assisted laser desorption/ionization linear time-of-flight mass spec-
trometer. Anal Chem. 1998;1995:67.
55. Vestal ML, Juhasz P, Martin SA. Delayed extraction matrix-assisted laser desorption time-of-
flight mass spectrometry. Rapid Commun Mass Spectrom. 1995;9:1044.
56. Laiko VV, Dodonov AF. Resolution and spectral-line shapes in the reflecting time-of-flight
mass-spectrometer with orthogonally injected ions. Rapid Commun Mass Spectrom. 1994;8:
7206.
57. Hsieh Y, Casale R, Fukuda E, Chen J, Knemeyer I, Wingate J, Morrison R, Korfmacher W.
Matrix-assisted laser desorption/ionization imaging mass spectrometry for direct measure-
ment of clozapine in rat brain tissue. Rapid Commun Mass Spectrom. 2006;20:96572.
58. Taban IM, Altelaar AFM, van der Burgt YEM, McDonnell LA, Baykut G, Heeren RMA.
Imaging of peptides in the rat brain using MALDI-FTICR mass spectrometry. J Am Soc
Mass Spectrom. 2007;18:14551.
59. Kanu AB, Dwivedi P, Tam M, Matz L, Hill Jr HH. Ion mobility-mass spectrometry. J Mass
Spectrom. 2008;43:122.
60. Gillig KJ, Ruotolo B, Stone EG, Russell DH, Fuhrer K, Gonin M, Schultz AJ. Coupling high-
pressure MALDI with ion mobility/orthogonal time-of-flight mass spectrometry. Anal Chem.
2000;72:396571.
61. Stauber J, Lemaire R, Wisztorski M, At-Menguellet S, Lucot JP, Vinatier D, Desmond A,
Deschamps M, Proess G, Rudlof I, Salzet M, Fournier I. New developments in MALDI imag-
ing mass spectrometry for pathological proteomic studies; introduction to a novel concept,
the specific MALDI imaging. Mol Cell Proteomics. 2006;5:S247.
62. Lawrence EO, Livingston MS. The production of high speed light ions without the use of
high voltages. Phys Rev. 1932;40:19.
63. Comisarow MB, Marshall AG. Fourier transform ion cyclotron resonance spectroscopy.
Chem Phys Lett. 1974;25:2823.
64. Marshall AG, Hendrickson CL, Jackson GS. Fourier transform ion cyclotron resonance mass
spectrometry: a primer. Mass Spectrom Rev. 1998;17:135.
65. Ledford Jr EB, Rempel DL, Gross ML. Space charge effects in Fourier transform mass spec-
trometry. Mass calibration. Anal Chem. 1984;56:27448.
66. Shi SDH, Drader JJ, Freitas MA, Hendrickson CL, Marshall AG. Comparison and
interconversion of the two most common frequency-to-mass calibration functions for Fourier
transform ion cyclotron resonance mass spectrometry. Int J Mass Spectrom. 2000;195196:
5918.
67. Wisztorski M, Verplanck N, Thomy V, Stauber J, Camart JC, Salzet M, Fournier I. Use of
masks in MALDI-MSI: an easy tool for increasing spatial resolution of images by decreasing
irradiated area. Indianapolis: American Society of Mass Spectrometry; 2007.
68. Wisztorski M, Croix D, Macagno E, Fournier I, Salzet M. Molecular MALDI imaging: an
emerging technology for neuroscience studies. Dev Neurobiol. 2008;68:84558.
69. Taylor CF, Paton NW, Lilley KS, Binz P-A, Julian RK, Jones AR, Zhu W, Apweiler R,
Aebersold R, Deutsch EW, Dunn MJ, Heck AJR, Leitner A, Macht M, Mann M, Martens L,
Neubert TA, Patterson SD, Ping P, Seymour SL, Souda P, Tsugita A, Vandekerckhove J,
Vondriska TM, Whitelegge JP, Wilkins MR, Xenarios I, Yates JR, Hermjakob H. The mini-
mum information about a proteomics experiment (MIAPE). Nat Biotechnol. 2007;25:
88793.
70. Orchard S, Hermjakob H. The HUPO proteomics standards initiativeeasing communica-
tion and minimizing data loss in a changing world. Brief Bioinform. 2008;9:16673.
71. Skld K, Svensson M, Norrman M, Sjgren B, Svenningsson P, Andrn PE. The significance
of biochemical and molecular sample integrity in brain proteomics and peptidomics: Stathmin
220 and peptides as sample quality indicators. Proteomics. 2007;7:444556.
72. Theodorsson E, Stenfors C, Math AA. Microwave irradiation increases recovery of neuro-
peptides from brain tissues. Peptides. 1990;11:11917.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 137

73. Metz B, Kersten GFA, Hoogerhout P, Brugghe HF, Timmermans HAM, de Jong A, Meiring
H, ten Hove J, Hennink WE, Crommelin DJA, Jiskoot W. Identification of formaldehyde-
induced modifications in proteins: reactions with model peptides. J Biol Chem. 2004;279:
623543.
74. Stauber J, Lemaire R, Franck J, Bonnel D, Croix D, Day R, Wisztorski M, Fournier I, Salzet
M. MALDI imaging of formalin-fixed paraffin-embedded tissues: application to model ani-
mals of Parkinson disease for biomarker hunting. J Proteome Res. 2008;7:969.
75. Lemaire R, Desmons A, Tabet JC, Day R, Salzet M, Fournier I. Direct analysis and MALDI
imaging of formalin-fixed, paraffin-embedded tissue sections. J Proteome Res. 2007;6:
1295305.
76. Crockett DK, Lin Z, Vaughn CP, Lim MS, Elenitoba-Johnson KSJ. Identification of proteins
from formalin-fixed paraffin-embedded cells by LC-MS//MS. Lab Invest. 2005;85:140515.
77. Chaurand P, Caprioli RM. Direct profiling and imaging of peptides and proteins from mam-
malian cells and tissue sections by mass spectrometry. Electrophoresis. 2002;23:312535.
78. Stoeckli M, Staab D, Staufenbiel M, Wiederhold KH, Signor L. Molecular imaging of amy-
loid beta peptides in mouse brain sections using mass spectrometry. Anal Biochem. 2002;
311:339.
79. Lemaire R, Wisztorski M, Desmons A, Tabet JC, Salzet M, Fournier I. MALDI-MS direct
tissue analysis of proteins: improving signal sensitivity using organic treatments. Anal Chem.
2006;78:714553.
80. Altelaar AFM, van Minnen J, Jimnez CR, Heeren RMA, Piersma SR. Direct molecular
imaging of Lymnaea stagnalis nervous tissue at subcellular spatial resolution by mass spec-
trometry. Anal Chem. 2005;77:73541.
81. Lemaire R, At-Menguellet S, Stauber J, Marchaudon V, Lucot JP, Collinet P, Farine MO,
Vinatier D, Day R, Ducoroy P, Salzet M, Fournier I. Specific MALDI imaging and profiling
for biomarker hunting and validation: fragment of the 11S proteasome activator complex, Reg
alpha fragment, is a new potential ovary cancer biomarker. J Proteome Res. 2007;6:412734.
82. Armin Holle AH, Kayser M, Hhndorf J. Optimizing UV laser focus profiles for improved
MALDI performance. J Mass Spectrom. 2006;41:70516.
83. Lemaire R, Stauber J, Wisztorski M, Van Camp C, Desmons A, Deschamps M, Proess G,
Rudlof I, Woods AS, Day R, Salzet M, Fournier I. Tag-mass: specific molecular imaging of
transcriptome and proteome by mass spectrometry based on photocleavable tag. J Proteome
Res. 2007;6:205767.
84. Groseclose MR, Andersson M, Hardesty WM, Caprioli RM. Identification of proteins
directly from tissue: in situ tryptic digestions coupled with imaging mass spectrometry.
J Mass Spectrom. 2007;42:25462.
85. Chaurand P, Rahman MA, Hunt T, Mobley JA, Gu G, Latham JC, Caprioli RM, Kasper S.
Monitoring mouse prostate development by profiling and imaging mass spectrometry. Mol
Cell Proteomics. 2008;7:41123.
86. Luxembourg SL, McDonnell LA, Duursma M, Guo X, Heeren RMA. Effect of local matrix
crystal variations in matrix-assisted ionization techniques for mass spectrometry. Anal Chem.
2003;75:233341.
87. Delcorte A. Matrix-enhanced secondary ion mass spectrometry: the alchemists solution?
Appl Surf Sci. 2006;252:65827.
88. Wilfried Szymczak KW. Effect of water treatment on analyte and matrix ion yields in matrix-
assisted time-of-flight secondary ion mass spectrometry: the case of insulin in and on
hydroxycinnamic acid. Rapid Commun Mass Spectrom. 2002;16:202533.
89. McArthur SL, Vendettuoli MC, Ratner BD, Castner DG. Methods for generating protein
molecular ions in ToF-SIMS. Langmuir. 2004;20:37049.
90. Adriaensen L, Vangaever F, Gijbels R. Metal-assisted secondary ion mass spectrometry:
influence of Ag and Au deposition on molecular ion yields. Anal Chem. 2004;76:677785.
91. Adriaensen L, Vangaever F, Lenaerts J, Gijbels R. Matrix-enhanced secondary ion mass spec-
trometry: the influence of MALDI matrices on molecular ion yields of thin organic films.
Rapid Commun Mass Spectrom. 2005;19:101724.
138 J. Stauber and R.M.A. Heeren

92. Aebersold R, Mann M. Mass spectrometry-based proteomics. Nat Insights. 2003;422:

198207.
93. Garrison BJ, Winograd N. Ion beam spectroscopy of solids and surfaces. Science. 1982;216:
80512.
94. Li KW, Smit AB, Geraerts WPM. Structural and functional characterization of neuropeptides
involved in the control of male mating behavior of Lymnaea stagnalis. Peptides. 1992;13:
6338.
95. Hanton SD, Clark PAC, Owens KG. Investigations of matrix-assisted laser desorption/ioniza-
tion sample preparation by time-of-flight secondary ion mass spectrometry. J Am Soc Mass
Spectrom. 1999;10:10411.
96. Stoeckli M, Staab D, Schweitzer A, Gardiner J, Seebach D. Imaging of a [beta]-peptide dis-
tribution in whole-body mice sections by MALDI mass spectrometry. J Am Soc Mass
Spectrom. 2007;18:19214.
97. van Veelen P, Jimenez C, Li K, Wildering W, Geraerts W, Tjaden U, van der Greef J. Direct
peptide profiling of single neurons by matrix-assisted laser-desorption ionization mass-
spectrometry. Org Mass Spectrom. 1993;28:15426.
98. Garden RW, Moroz LL, Moroz TP, Shippy SA, Sweedler JV. Excess salt removal with matrix
rinsing: direct peptide profiling of neurons from marine invertebrates using matrix-assisted
laser desorption/ionization time-of-flight mass spectrometry. J Mass Spectrom. 1996;31:
112630.
99. Redeker V, Toullec J-Y, Vinh J, Rossier J, Soyez D. Combination of peptide profiling by
matrix-assisted laser desorption/ionization time-of-flight mass spectrometry and immunode-
tection on single glands or cells. Anal Chem. 1998;70:180511.
100. Stoeckli M, Chaurand P, Caprioli RM. Direct profiling of proteins in biological tissue sec-
tions by MALDI mass spectrometry. Anal Chem. 1999;71:526370.
101. Garden RW, Sweedler JV. Heterogeneity within MALDI samples as revealed by mass spec-
trometric imaging. Anal Chem. 2000;72:306.
102. Stoeckli M, Chaurand P, Hallahan DE, Caprioli RM. Imaging mass spectrometry: a new
technology for the analysis of protein expression in mammalian tissues. Nat Med. 2001;7:
4936.
103. Rubakhin SS, Greenough WT, Sweedler JV. Spatial profiling with MALDI MS: distribution
of neuropeptides within single neurons. Anal Chem. 2003;75:537480.
104. Kruse R, Sweedler JV. Spatial profiling invertebrate ganglia using MALDI MS. J Am Soc
Mass Spectrom. 2003;14:7529.
105. Monroe EB, Jurchen JC, Lee J, Rubakhin SS, Sweedler JV. Vitamin E imaging and localiza-
tion in the neuronal membrane. J Am Chem Soc. 2005;127:12152.
106. Hintersteiner M, Enz A, Frey P, Jaton A-L, Kinzy W, Kneuer R, Neumann U, Rudin M,
Staufenbiel M, Stoeckli M, Wiederhold K-H, Gremlich H-U. In vivo detection of amyloid-
beta deposits by near-infrared imaging using an oxazine-derivative probe. Nat Biotechnol.
2005;23:57783.
107. Stauber J, Lemaire R, Wisztorski M, At-Menguellet S, Lucot JP, Vinatier D, Desmond A,
Deschamps M, Proess G, Rudlof I, Salzet M, Fournier I. New developments in MALDI imag-
ing mass spectrometry for pathological proteomic studies; introduction to a novel concept,
the specific MALDI imaging. Mol Cell Proteom. 2006;5:S247.
108. Baluya DL, Garrett TJ, Yost RA. Automated MALDI matrix deposition method with Inkjet
printing for imaging mass spectrometry. Anal Chem. 2007;79:68627.
109. Hsieh Y, Chen J, Korfmacher WA. Mapping pharmaceuticals in tissues using MALDI imag-
ing mass spectrometry. J Pharmacol Toxicol Meth. 2007;55:193200.
110. Atkinson SJ, Loadman PM, Sutton C, Patterson LH, Clench MR. Examination of the distri-
bution of the bioreductive drug AQ4N and its active metabolite AQ4 in solid tumours by
imaging matrix-assisted laser desorption/ionisation mass spectrometry. Rapid Commun Mass
Spectrom. 2007;21:12716.
111. Signor L, Varesio E, Staack RF, Starke V, Richter WF, Hopfgartner G. Analysis of erlotinib
and its metabolites in rat tissue sections by MALDI quadrupole time-of-flight mass spectrom-
etry. J Mass Spectrom. 2007;42:9009.
5 Biological Tissue Imaging at Different Levels: MALDI and SIMS 139

112. Rohner TC, Staab D, Stoeckli M. MALDI mass spectrometric imaging biological tissue
sections. Mech Ageing Dev. 2004;126:17785.
113. Gusev AI, Wilkinson WR, Proctor A, Hercules DM. Direct quantitative analysis of peptides
using matrix assisted laser desorption ionization. Fresenius J Anal Chem. 1996;354:45563.
114. Wilkinson WR, Gusev AI, Proctor A, Houalla M, Hercules DM. Selection of internal stan-
dards for quantitative analysis by matrix-assisted laser desorption-ionization (MALDI) time-
of-flight mass spectrometry. Anal Bioanal Chem. 1997;357:2418.
115. Hopfgartner G, Varesio E, Stoeckli M. Matrix-assisted laser desorption/ionization mass spec-
trometric imaging of complete rat sections using a triple quadrupole linear ion trap. Rapid
Commun Mass Spectrom. 2009;23:7336.
116. Amann JM, Chaurand P, Gonzalez A, Mobley JA, Massion PP, Carbone DP, Caprioli RM.
Selective profiling of proteins in lung cancer cells from fine-needle aspirates by matrix-
assisted laser desorption ionization time-of-flight mass spectrometry. Clin Cancer Res.
2006;12:514250.
117. Burrell MM, Earnshaw CJ, Clench MR. Imaging matrix assisted laser desorption Ionization
mass spectrometry: a technique to map plant metabolites within tissues at high spatial resolu-
tion. J Exp Bot. 2007;58:75763.
118. Bunch J, Burrell MM, Clench MR. MALDI imaging to reveal metabolite profiles in potato
tubers. Abstract/Comp Biochem Physiol A. 2004;137:14760.
119. Mullen AK, Clench MR, Crosland S, Sharples KR. Determination of agrochemical com-
pounds in soya plants by imaging matrix-assisted laser desorption/ionization mass spectrom-
etry. Rapid Commun Mass Spectrom. 2005;19:250716.
120. Jonathan Stauber, Luke MacAleese J, Franck, Marten Snell, Emmanuelle Claude, basak
Kkrer Kaletas, Ingrid van der Wiel, Maxensce Wisztorski, Isabelle Fournier and Ron M.A.
Heeren JASMS. On-Tissue Protein Identification and Imaging by MALDI-Ion Mobility Mass
Spectrometry. 2010;21:33847.
121. Robinson S, Warburton K, Seymour M, Clench M, Thomas-Oates J. Localization of water-
soluble carbohydrates in wheat stems using imaging matrix-assisted laser desorption/ioniza-
tion mass spectrometry. New Phytol. 2006;173:42844.
122. Woods AS, Wang HY, Jackson SN. A snapshot of tissue glycerolipids. Curr Pharm Des.
2007;13:334456.
123. Kahn E, Tessier C, Lizard G, Petiet A, Brau F, Clement O, Frouin F, Jourdain JR,
Guiraud-Vitaux F, Colas-Linhart N, Siauve N, Cuenod CA, Frija G, Todd-Pokropek A.
Distribution of injected MRI contrast agents in mouse livers studied by confocal and SIMS
microscopy. Anal Quant Cytol Histol. 2002;24:295302.
124. Langstrom B, Andren PE, Lindhe O, Svedberg M, Hall H. In vitro imaging techniques in
neurodegenerative diseases. Mol Imaging Biol. 2007;9:16175.
125. Kahn E, Tessier C, Lizard G, Petiet A, Bernengo GC, Coulaud D, Fourr C, Frouin F, Clment
O, Jourdain JR, Delain E, Guiraud-Vitaux F, Colas-Linhart N, Siauve N, Cuenod CA, Frija
G, Todd-Pokropek A. Analysis of the distribution of MRI contrast agents in the livers of
small animals by means of complementary microscopies. Cytometry. 2003;51A:97106.
126. Acquadro E, Cabella C, Ghiani S, Miragoli L, Bucci EM, Corpillo D. Matrix-assisted laser
desorption ionization imaging mass spectrometry detection of a magnetic resonance imaging
contrast agent in mouse liver. Anal Chem. 2009;81:277984.
127. Shimma S, Sugiura Y, Hayasaka T, Zaima N, Matsumoto M, Setou M. Mass imaging and
identification of biomolecules with MALDI-QIT-TOF-based system. Anal Chem. 2008;80:
87885.
128. Stauber J, Kkrer Kaletas B, van der Wiel IM, Snel MF, Claude E, Heeren RMA. Ion mobil-
ity imaging mass spectrometry: a new tool for in situ identification proceedings, ASMS
Denver. 2008.
Chapter 6
Molecular Structure and Identification
Through G-SIMS and SMILES

F.M. Green, I.S. Gilmore, and M.P. Seah

Abstract In this chapter, we discuss the use of G-SIMS (gentle secondary ion mass
spectrometry) and SMILES (simplified molecular input line entry specification) for
analyzing biologically relevant materials and molecules. G-SIMS is an easy-to-use
method that considerably simplifies complex static secondary ion mass spectrome-
try (SSIMS) spectra into spectra with only those ions that are highly characteristic
of the surface. G-SIMS provides information about the molecular structure that is
not directly available from the mass spectrum, allowing the identification of
unknown materials without the need for experimental library spectra. For complex
molecules such as biomolecules, identification of the most characteristic fragment
ions alone may be insufficient to uniquely identify a molecule because of the com-
binatorial chemical possibilities available within the achievable mass accuracy. The
molecular structure can be reassembled by following the fragmentation pathways
(by varying the G-SIMS surface plasma temperature); this technique is known as
G-SIMS-FPM (G-SIMS fragmentation pathway mapping). This provides powerful
capability analogous to MS/MS experiments traditionally used in mass spectrome-
try. A simple method, using SMILES, is used to simulate the fragmentation path-
ways that occur in G-SIMS. These pathways are found to have good agreement with
the G-SIMS fragmentation pathways. The simulated pathways help analysts deduce
the molecular structure, leading to refined identification. A rapid method to estab-
lish a foundational database of simulated pathways using the SIMS community and
a web-based system is being developed.

F.M. Green (*) I.S. Gilmore M.P. Seah

Analytical Science Division, National Physical Laboratory,
Teddington, Middlesex TW11 0LW, UK
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 141

DOI 10.1007/978-3-319-01360-2_6, Springer International Publishing Switzerland 2014
142 F.M. Green et al.

6.1 Introduction

The strength of SSIMS, and also one of its drawbacks, is that the mass spectra are
extremely rich, with hundreds of mass peaks, many of which clearly contain aspects
of the molecular information. As we shall see later, many of these peaks are degraded
fragment ions that, unfortunately, do not relate directly to parent molecules. These
ions confuse the spectral interpretation. Figure6.1 illustrates the issue with a mass
spectrum of an industrial antioxidant, Irganox 1010, used in the polymer industry,
on a silver substrate. It is clear that there are many peaks, and by progressively
zooming in on smaller and smaller regions, we see the complexity of the spectrum.
Traditionally, to identify a material, the SSIMS spectrum is used as a fingerprint
for comparison with library spectra. Such libraries [13] are excellent andthrough
the pioneering work of Vickerman and Briggsare of great benefit to the whole
community and widely used. However, the growth and coverage of experimental
libraries are, by necessity, limited. Twenty years after the publication of the first

Fig. 6.1 The positive-ion static SIMS spectrum of a surface layer of Irganox 1010 molecules on
silver (inset), illustrating the complexity of the mass spectrum. Note, for clarity, the inset only
shows one molecule on the surface, whereas the spectrum is for a monolayer of molecules
6 Molecular Structure and Identification Through G-SIMS and SMILES 143

SIMS data

Classification Identification Localisation Quantification

/Correlation
No library data Library data Library data No library data
required exist exist required

PC-DFA Highlights G-SIMS Isolates PCA Dimensionality G-tip Molecules PCR Quantitatively
differences between molecular peaks reduction, noise localised relate two sets of
predefined groups. filtering, highlights measurements
Supervised variation using variance of
classification predictor
MCR Recovery of
HCA Unsupervised G-SIMS FPM spectral components G-SIMS Isolates PLS Quantitatively
partitioning of data Identifies and contributions molecular peak relate two sets of
into groups using molecular structure using constraints from local areas measurements
their distances using their
SMILES FPM MAF Dimensionality covariance
ANNs Supervised Finds a structure reduction, noise
classification into match using filtering, highlights
predefined groups simulated image variation
using networks of fragmentation - Require relevant
neurons pathways. experimental/library data

Fig. 6.2 The complementarity of G-SIMS with multivariate methods. Note: PC-DFA principal
component discrimination analysis, HCA hierarchal cluster analysis, ANNs artificial neural net-
works, PCA principal component analysis, MCR multivariate curve resolution, MAF maximum
autocorrelation factors, PCR principal component regression, PLS partial least squares

library in 1989, the combined content contains spectra for 500 materials only.
In comparison to the range of industrially relevant materials, this is a tiny fraction.
For new and rapidly growing industrial sectors such as pharmaceuticals, biomateri-
als, and biotechnology, the problems are acute. A library-independent method was
required that provides direct interpretation and accesses the full range of required
materials. Hence, G-SIMS was developed.
Multivariate analytical methods are very powerful for the identification, quanti-
fication, classification, and development of models for prediction. These methods
are described in detail in the previous chapter and also in Ref. [4]. In Fig.6.2, we
illustrate the complementarity between multivariate methods and G-SIMS. Both
methods simplify the mass spectra; multivariate methods reduce the complexity of
the data using statistical methods (i.e., the variance), while G-SIMS simplifies the
spectra by eliminating heavily fragmented and degraded molecules, leaving the
more intact structurally significant ions. This utilizes the underlying chemistry of
the molecules and the physics of the fragmentation process. This latter characteris-
tic enables G-SIMS to extract information that is otherwise not directly accessible
from the spectra, thus providing powerful information on the molecular structure.
Multivariate methods are excellent to identify trends and correlations in spectra and
144 F.M. Green et al.

1 Simplifies spectra.
0.9
0.8

Normalised Intensity
0.7 The most intense peaks are
G-SIMS(7,8,9)
0.6
0.5 directly related to molecular
0.4
0.3 structure.
0.2
0.1
0 Makes identification simpler
0 50 100 150 200 250 300 350 400 450
Mass,u and more direct.

Mass alone may be

insufficient for identification
molecular structure required.
G-SIMS-FPM(9,27)
Enables identification of
possible molecular structures
using latent spectral
information.

A library of simulated
fragmentation pathways helps
SMILES-FPM(35,36) guide analysts through
identification of molecular
structure.

Fig. 6.3 Information available from G-SIMS, G-SIMS-FPM, and G-SIMS SMILESa quick
reference guide for analysts

images with other experimental variables, such as composition or, for example,
production methods, and, coupled with a priori information, for quantification, clas-
sification, and prediction [4]. G-SIMS spectra are not used directly for quantifica-
tion but for identification; however, the G-SIMS peaks are good peaks to select for
quantification. For example, for the quantification of polystyrene materials, the tro-
pyllium ion (C7H7+) is the most intense ion, but quantification is poorer than if the
structurally significant ions identified by G-SIMS (e.g., C15H13+) are used [5].
In Sect.6.2, we review the G-SIMS concept and the practicalities of acquiring
G-SIMS spectra with a simple guide. Examples of the effectiveness of G-SIMS for
a variety of biologically relevant polymers and molecules are given in Sect.6.3. In
Sect.6.4, the use of G-SIMS to elucidate the molecular structure, by examining the
fragmentation pathways, is described. This is known as G-SIMS fragmentation
pathway mapping (FPM). A novel approach based on SMILES (simplified molecu-
lar input line entry specification) [6], to simulate fragmentation pathways and help
identify the molecular structure, is given in Sect.6.5. Section6.6 is a short outline
of recent innovations for G-SIMS imaging. Finally, in Sect.6.7, we discuss the
future outlook.
A summary of the increasing molecular information now available to the analyst
as we progress from a basic G-SIMS analysis through G-SIMS-FPM to G-SIMS
with SMILES is given in Fig.6.3. This indicates the method to select, depending on
the level of information required for any particular analytical situation.
6 Molecular Structure and Identification Through G-SIMS and SMILES 145

6.2 G-SIMS

In this section, we give an overview of the principles of G-SIMS and the role of
fragmentation in SIMS, as well as practical help for analysts. The section begins
with an overview of the G-SIMS concept and follows with an easy guide to obtaining
G-SIMS spectra. This includes an examination of the way fragmentation varies for
ions of different energy and species and helps the analyst choose suitable p rimary
ions to obtain G-SIMS.

6.2.1 Overview of the G-SIMS Concept

We envisage the impact of a single primary ion schematically in Fig.6.4. Typically,

for efficient ion beams, the ion energy is approximately 1025keV; that energy is
dissipated in the surface of the material over a depth range of approximately 25nm
(around 100 atom layers). The energy distribution of excited atoms at the surface
may be approximated by a 2D Gaussian, centered around the original impact site
with an FWHM of around 2nm. In this scheme, we consider a monolayer of
molecules, such as folic acid, covering the surface. It is clear that those folic acid
molecules along the peripheral zone of the impact site are more gently liberated
from the surface and have a higher probability of remaining intact. As we move
closer to the center of the impact site, the energy rises and the molecules begin to

High internal energy

heavily fragmented
10 keV Ar+

Intact molecules
low fragmentation
Temperature

Tp

5 nm Position +5 nm

Fig. 6.4 Illustration of the surface-energy distribution from a single-ion impact on a target mate-
rial with a surface monolayer of folic acid molecules. Typical fragmentation products for folic acid
are shown
146 F.M. Green et al.

Fig. 6.5 Schematic of a hydrocarbon molecule of the series CnH2n

fragment more, so that, at the center, the emitted ions have a high internal energy
and most probably consist of small degraded fragments. The measured SIMS spec-
trum is the volume integral of all these fragment ions and thus contains some intact
molecular ions among a dominant background of smaller degraded ions. We see an
example of this in Fig.6.1. The small fragment ions are not helpful for identifica-
tion, as they are ubiquitous among many different material types. In many cases,
larger, more stable ions, such as polycyclic aromatic ions, may form in this area.
These ions may be misleading and, if used, give poor quantification.
We now consider how fragmentation of a molecule leads to daughter products
and how their intensities are related. In the recoil from the primary ion impact, the
surface zone may be characterized by a surface plasma temperature, Tp. This is a
function of the radius, r, from the point of impact. The plasma temperature is also a
function of the bombarding ion species and the impact energy, E. If we reduce the
plasma temperature by adjusting the primary beam energy, the relative intensities of
the more intact molecular fragments would be expected to increase. In principle,
given the relative intensities of ions at two different values of Tp, it is possible to
extrapolate to the relative spectral intensities at a lower surface plasma temperature.
This is the basic principle of G-SIMS, which is explained below and in more detail
in Refs. [79].
We consider the fragmentation of a hydrocarbon molecule, as shown in Fig.6.5,
into the series CnH2n, CnH2n1, CnH2n2, and so on, and note that the energy, u, to
remove each successive hydrogen atom is approximately the same. Thus, the num-
ber of fragments, Ni, of composition CnH2ni, derived from N0 components of com-
position CnH2n, is given by the simple partition function relation

i u
N i = N 0 exp . (6.1)
kTp

If we now consider these at the two ion beam energies, E1 and E2, with associated
fragment surface plasma temperatures Tp1 and Tp2, we get

N i ( E2 ) N 0 ( E2 ) iu 1 1
= exp . (6.2)
N i ( E1 ) N 0 ( E1 ) k Tp 2 Tp1

The ratio of Ni(E2), for a low ion beam energy, E2, to those at high energy, Ni(E1),
gives a factor, Fx, where x is the index related to each mass peak. This Fx term is
6 Molecular Structure and Identification Through G-SIMS and SMILES 147

related to the effective surface plasma temperatures Tp1 and Tp2 from Eq.6.2. Since
Tp1 and Tp2 are of a similar magnitude and Tp2 is less than Tp1 by T, we may first
consider the factor Fx2, which would be the Fx value for surface plasma temperatures
Tp12T and Tp1. Thus, one could use Fx13 or some high power to deduce the result
at a significantly lower surface plasma temperature than that relevant to either of the
recorded spectra.
We may now generate a low surface plasma temperature SSIMS spectrum by
multiplying an existing spectrum, Nx, with the factor Fxg. This forms the G-SIMS
spectrum with intensities Ix given by
I x = M x N x Fxg , (6.3)

where g, the G-SIMS index, is often set at 13. The additional factor Mx, the mass of
the emitted fragment, is found useful to enhance the natural fall in emission with
mass. In Ref. [8], a simple method is described that constructs a tangent line passing
through the two most significant high-mass fragments. A new ratio, Fx*, is then
calculated by dividing Fx by the function of the tangent line.
It is useful to look in a little more detail at the spectral intensity ratio, Fx, as this
will help with later understanding of the method. The intensity ratio on the left of
Eq.6.2 can be rewritten as Fx,, where

Fx ,i = Fx ,0 exp ( b i ) (6.4)

and i is given by the items in square brackets on the right of Eq.6.2. Thus, an
intensity ratio plot of loge(Fx,i) versus mass will show data for a given CnH2nI series
falling on a straight line with gradient per increment in i. In the case used here,
where the increment in i is a hydrogen atom, the gradient is simply per unified
atomic mass units, u, or Dalton, Da. If u is positive and if the surface plasma tem-
perature, Tp1, at energy E1 is greater than the surface plasma temperature, Tp2, at
energy E2, will be positive, and vice versa. Note that the energy to remove a hydro-
gen atom from C12H24 will be very similar to that from C6H6, so that the gradients,
, will be the same for the hydrogen loss series in the intensity ratio plot for many
values of n.
Figure6.6 shows the intensity ratio plot (Fx,i vs. mass) for the E2 spectra for pri-
mary ion Ar+ with E2=4keV and the E1 spectrum an average for Ar+ with 410keV
energies, with the CxHy ions plotted with a filled circle, distinguishing the gradients,
, for different hydrocarbon cascades. This intensity ratio plot clearly shows the
fragmentation cascades and so will be termed a fragment cascade plot in the rest of
the chapter. Observing the parallelism of the gradients in Fig.6.6 shows that, for
x>3, a single value of u or may be used to describe the fragmentation process.
We will later return to the value of for the selection of optimum ion beam param-
eters. Fortunately for the analysts, we do not need to calculate the value of for
G-SIMS or indeed any other parameterall that is needed is the ratio of spectral
intensities, Fx, as shown above in Eq.6.4.
The G-SIMS procedure blends information about the amount of fragmentation
each peak has undergone (from the fragment cascade plot) with the original spectral
148 F.M. Green et al.

Fig. 6.6 Fragment cascade plot showing the ratio of polystyrene (PS) spectral intensities Fx, for 4
and an average of 410-keV argon primary ions. The filled circles represent fragments of the type
CxHy, whereas the empty circles are peaks for contaminants (After Ref. [7])

intensities. Those fragments that have little degradation are boosted in intensity and
those with significant degradation are suppressed. This is what would be observed
if one could do static SIMS with a very low Tp value (or very low energy). As an
example of the power of G-SIMS for direct analysis, we show in Fig.6.7 the SSIMS
and G-SIMS spectra for polystyrene (PS). The dominant intensity ions in the SSIMS
spectra (Fig.6.7b) are polycyclic aromatic ions with complex linked cyclic arrange-
ments, which exhibit little direct relevance to the PS structure. These structures are
very stable ions produced from the high-energy fragmentation cascade, either as
recombination events or as the end result of a decay process for a molecule origi-
nally with excess energy. These characteristic peaks are common to many other
materials. In contrast, the G-SIMS spectrum (Fig.6.7a) is dominated by ions with
pendant phenyl groups, exactly as would be expected from the molecular structure,
giving direct identification without the need for library data. G-SIMS has been
tested extensively on many materials, including polymers, such as PS [7], PC [7],
PTFE [7], and PMMA [8], and complex molecules, including Irganox 1010 [8],
caffeine [8], cholesterol [8], glucose [8], poly-l-lysine [8], and folic acid [9], biode-
gradable polymers [1012], and adhesives [13]. Typically, for SSIMS, only one or
two peaks above, say, 1% of the maximum peak intensity provide direct interpreta-
tion out of several hundred peaks. For G-SIMS, the situation is quite different; spec-
tra typically contain five or so peaks, which lead to direct interpretation and
identification. This step change opens the door to applications in areas where the
fingerprint approach of collating spectra is not practicable.
6 Molecular Structure and Identification Through G-SIMS and SMILES 149

Fig. 6.7 Positive-ion spectra of polystyrene. (a) G-SIMS spectrum using the ratio between 10-
and 4-keV argon ions;. (b) static SIMS spectrum using 10-keV argon ions (After Ref. [7])
150 F.M. Green et al.

6.2.2 Practicalities of G-SIMS: A Simple Guide

6.2.2.1 Optimizing Instrument Parameters

In practice, to generate G-SIMS spectra, all that is required are two sequential spec-
tra acquired from the same area using different ion beam conditions (providing dif-
ferent surface plasma temperatures). An important prerequisite is that the instrument
should be in good control with repeatabilities of typically better than 2%. This is
readily achievable in modern instruments. In a recent VAMAS interlaboratory study,
around 90% of 32 participating laboratories achieved better than 2% repeatability
[14]. It is recommended that analysts use the standard ISO 23830 for determining
the repeatability and constancy of their instrument [15]. A spreadsheet that does all
of the necessary calculations is freely available from the NPL website; see Ref. [16].
As with all mass spectrometries, for accurate identification it is critical to cali-
brate the mass scale accurately. In a second VAMAS interlaboratory study, it was
found that the fractional mass accuracy for large molecules was typically very poor
[17] (150ppm). A procedure for improving the mass calibration, by typically a fac-
tor of 10, is given in Ref. [18]; an ISO standard for this is now available as
ISO 13084:2011. Similarly, it is important to ensure that the mass resolution of the
instrument is sufficient to resolve chemical differences.

6.2.2.2 Selecting the Ion Beam Parameters

We discussed earlier the index , the gradient of the fragment cascades for CxHy
observed in Fig.6.6. The larger the value of is between conditions, the stronger
the difference will be in surface plasma temperatures. In Fig.6.8, we show the
values of for PS for different primary ion species (Ar+, Mn+, Ga+, Xe+, Cs+, Bi+)
for a range of practical ion beam energies, relative to Ar+ at 7keV. It is clear that
using a Ga+ ion beam at say 12 and 25keV gives only a small change in , and
therefore only a small difference in surface plasma temperatures, and consequently
is not effective for G-SIMS. Selecting primary ions with high and low mass is best.
A number of laboratories have found that Cs (131 u) and Ar (40 u) or Cs and Ga
(69u) provide very effective combinations for G-SIMS [17]. However, by using
different primary ions, we introduce a general mass dependency in Fx (Fig. 11 of
Ref. [7]). In early work, that was removed by a simple cubic fit to the data using a
least-modulus minimization. More recent work shows that an erf function (a func-
tion with a sigmoidal shape giving an output rising from zero to one as the input
increases) gives a more universal fit and is recommended.
For heterogeneous samples (i.e., most samples), it is important that the analyses
are from the same area, and so the ion beams need to be aligned. Clearly, misalign-
ment leads to the ratio of spectra from different areas, causing the method to be less
reliable. If available, a dual-beam source sharing the same ion-optical column is
preferable. This ensures co-alignment at the sample. Recently [19], a novel emitter
for a Taylor cone liquid metal ion source (LMIS) was developed that consists of Bi
6 Molecular Structure and Identification Through G-SIMS and SMILES 151

Fig. 6.8 Average cascade 0.12

gradient, , for Ar+, Ga+, Xe+,
0.1 Cs
Cs+, Bi+, and Mn+ for various
Bi
energies on polystyrene
(After Ref. [19]) 0.08
Xe
0.06
b
0.04

0.02
Ar
0 Ga
Mn
0.02
0 10 20 30
Ion energy, keV

(208.98 u) with a small amount of Mn (54.94 u). This is a very convenient option
for analysts as it may be used in all Taylor cone LMIS, which are now the dominant
primary ion guns in SSIMS. In this way, the two ion beams are inherently aligned
[19]. We discuss this later in Sect.6.6. The alternative use of separate ion columns
will need a careful alignment procedure.

6.2.2.3 Acquisition of Spectra

It is important to ensure that the total fluence of ions is kept below 11016ions/m2
so that ion-induced damage effects are small [20]. If the experiment requires charge
compensation using an electron beam, it is also very important to ensure that the
electron fluence is kept below 61018electrons/m2. More details and guidance are
given elsewhere [21].
Acquire two mass spectra, the first using the ion beam of lowest fragmentation
(for example, Cs+, Bi+, or lower-energy ions), followed immediately by a second,
using the ion beam of higher fragmentation (for example, Ar+, Mn+, or higher-
energy ions). Most modern instruments can be programmed using a batch method,
making acquisition straightforward. To improve and monitor measurement uncer-
tainty, it is recommended that pairs of spectra are acquired from at least three sepa-
rate areas, preferably five.
The next step is exactly as for any SSIMS experiment: The peak intensities need
to be measured from the mass spectrum. Use the low-fragmentation spectrum to
choose the peaks. Most commercial software can perform this automatically by
identifying the peak centroid and width data. When applied to the mass spectrum,
these data yield a data matrix of peak centroid mass and peak intensities. When using
an automatic peak-search method, it is important to review the peaks identified to
check that metastables [22] and other background artifacts are excluded. Metastables
actually provide valuable information about the molecular structure that is very
152 F.M. Green et al.

complementary to G-SIMS. We will not discuss this further here, but interested readers
are directed to Ref. [22], which provides an easy-to-use method to identify metasta-
ble peaks and their parent ions. If peaks are not resolved into separate components,
this can cause problems with identificationas with any analysis.
Use the same table of peak positions and widths to measure the peak intensities
for all the spectra. G-SIMS spectra are then calculated directly using Eq.6.3.
Tomake it easier, a simple Excel spreadsheet Easy G-SIMS is freely available on
the web at Ref. [23]. This spreadsheet simply requires input of the peak masses and
spectral intensities for matching low- and high-fragmentation data in columns. The
spreadsheetalso gives the repeatability of the data and incorporates the tangent
gradient method, which is very useful for large molecules although generally is not
required for polymers. The index g may be altered, effectively allowing the surface
plasma temperature to be adjusted. We shall come back to this later for the analysis
of molecular structure. G-SIMS may be conducted for both positive and negative
secondary ions.

6.3 G-SIMS Examples

In the previous section, we outlined the G-SIMS concept and the practicalities of
producing G-SIMS spectra. We now provide examples to illustrate the utility of the
method. The static SIMS spectrum of polycarbonate (PC) is dominated by polycy-
clic aromatic hydrocarbon ions and other hydrocarbons that do not relate directly to
the PC monomer repeat unit structure but instead are degraded fragments and rear-
rangement product ions (see Fig. 12a of Ref. [7]). In contrast, the G-SIMS spectrum
exhibits a dominant ion at C9H11O+, which is directly related to the parent structure,
as well as other structurally significant ions: C2H3O+, C3H3O+, C7H7O+, and C8H9O+.
The degradation fragments such as polycyclic aromatic hydrocarbon ions are absent
from the G-SIMS spectrum (see Fig. 12b of Ref. [7]). Therefore, the PC reference
material provides a good and clear test for the successful operation of G-SIMS [17].
For G-SIMS to be judged successful for this reference material, each of the follow-
ing requirements must be met: the presence of a strong C9H11O+ peak; the absence
of polycyclic aromatic peaks; good repeatability; and low- and high-fragmentation
conditions that give a well-separated surface plasma temperature (observed by a
clear structure in the ratio Fx values, shown in Fig.6.6 for PS).
Earlier, in the introduction, we illustrated the complexity of the static SIMS spec-
trum with the Irganox molecule, shown in Fig.6.1. For comparison, we show in
Fig.6.9 the G-SIMS spectra, using the tangent gradient, for positive and negative
ions. It is clear that these spectra are much simpler than Fig.6.1 and the peaks are
directly related to the parent structure.
In 2006, Ogaki etal. [11] successfully used G-SIMS to study a series of biode-
gradable polyesters; these are medically important, particularly in drug-delivery
systems and biomedical implants. These materials include polyglycolic acid (PGA),
poly-l-lactic acid (PLA), poly--hydroxybutyrate (PHB), and poly--caprolactone
6 Molecular Structure and Identification Through G-SIMS and SMILES 153

a 1.2
Irganox 1010
R-CH2O H3C CH3

1 R C CH3

R= H2
O

H2
R C R C C
Normalised intensity

C O
O C
H2

0.8 But R C CH3

H3C CH3

R
0.6
CH2 C R
CH3 R
0.4 CH 3
CH 3 -But
CH 2 OH

H
M + CH2
0.2

0
0 200 400 600 800 1000 1200
Mass, u
b 1.2

M-H
1
Normalised intensity

0.8

0.6

C2H3O2
0.4

R-CH4
0.2

0
0 200 400 600 800 1000 1200
Mass, u

Fig. 6.9 G-SIMS spectra of Irganox 1010: (a) positive ions; (b) negative ions, using the ratio
between 10-keV argon and cesium ions (After Ref. [27])

(PCL) [1012]. They have also used G-SIMS to study PLA end-group contribution
to SIMS spectra [12] and random poly (lacticco-glycolic acid) copolymers. As a
typical example of this series of materials, we show in Fig.6.10 the G-SIMS spec-
trum for an 80:20 PLGA co-polymer from Ref. [12]. All of the peaks are directly
related to the molecular structure and can easily be attributed to lactide, glycolide,
and combined fragments. The G-SIMS analysis also provides additional informa-
tion on the formation of the secondary ions, which agrees with deuteration studies.
154 F.M. Green et al.

Fig. 6.10 G-SIMS spectrum of 80:20 PLGA co-polymer. Repeat unit of lactide-only fragments
are marked , glycolide-only fragments are marked , and mixed fragments are marked
(From Ref. [12], reproduced with permission)

In Fig.6.11, we show the G-SIMS spectrum of cholesterol (C27H46O), with mass

386.3549 u. The spectrum is very simple, with a dominant [MOH] ion that is
more similar to traditional electron impact gas-phase mass spectra than the comple-
mentary SSIMS spectrum shown in Fig.6.11b. Similarly, in Fig.6.11c, the G-SIMS
spectrum of caffeine (C8H10N4O2) is much simpler than the SSIMS spectrum of
Fig.6.11d. The G-SIMS spectrum consists of molecular peaks [MCH]+, [M+H]+,
[M+2H]+, [M+CH]+, and [2M+H]+. Figure6.11e, f show the G-SIMS and SSIMS
spectra of the polyamino acid poly-l-lysine. The poly-l-lysine static SIMS spectrum
is dominated by high intensities of fragmentation products in the low-mass range.
The power of G-SIMS is illustrated by the unambiguous peaks representing the
functional group of the amino acid and a dimer fragment at higher mass. The intense
double peak in between is a separate bromide compoundthe material is supplied
as a salt. Samples of bovine serum albumin have also been analyzed [9], showing
the correct fragmentation behavior and allowing identification of individual amino
acid groups.
The molecular structure of folic acid is shown in Fig.6.12 together with six pos-
sible subunits that could assemble to form the molecule, denoted by , , , , , and
. The total molecular weight using the main isotopes is 441.1396 u. Folic acid is
composed of C, H, N, and O arranged in the chemically rather different subunits that
make a chemical composition assignment from the measured fragment mass diffi-
cult, even for time-of-flight systems with typically 10-ppm mass-scale accuracy.
6 Molecular Structure and Identification Through G-SIMS and SMILES 155

a 1 b 1

0.9 Cholesterol M-OH 0.9

Cholesterol
0.8 0.8

0.7 0.7

Normalised Intensity
Normalised Intensity

0.6 0.6

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2 M-OH

0.1 0.1 M-H

0 0
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Mass, u Mass, u

c 1
d
100
M+2H
0.9 Caffeine Caffeine
2M-H
0.8
M+H M+H
0.7
Normalised Intensity

Normalised Intensity
10-1
0.6
M-CH M
M+2H
0.5
M + CH
0.4 M+CH3

0.3 10-2
M-CH 2M-H
0.2

0.1

0 10-3
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
Mass, u Mass, u

e 1 f
Poly-L-lysine
O
0.15 H
0.8 CH Br salt C
N C
H
CH CH2
Normalised Intensity
Normalised Intensity

0.6 O CH2
CH2 O H 0.10
N C N C C N+ CH2
C
CH2 H H H
0.4 CH2 CH2 CH2
CH2 CH2 CH2 NH2
0.05
NH2 CH2 CH2
0.2
CH2 CH2

NH2 NH2
0 0
0 100 200 300 400 0 100 200 300 400
Mass, u Mass, u

Fig. 6.11 Positive G-SIMS spectra (left) and static SIMS (right) for (a) and (b) cholesterol,
(c)and (d) caffeine, and (e) and (f) poly-l-lysine using the ratio between 10-keV argon and cesium
ions (Here and elsewhere, G-SIMS spectra are often shown in green and the complementary static
SIMS in red.) (After Ref. [8])

a b g d

OH

H N
HO C C C C N C N C N
H2 H 2 H H H2
O C O
H N N NH2
O OH

Fig. 6.12 The molecular structure of folic acid; different subsections have been labeled
156 F.M. Green et al.

a1.2 b 10 0

* PDMS peaks
1 { -NH}+
+ +
Normalised intensity

Normalised intensity
*
0.8 10-1

0.6
*
10-2 *
0.4
+NH
*
0.2
*
10-3
0
0 100 200 300 400 500 0 100 200 300
Mass, u Mass, u

Fig. 6.13(a) G-SIMS spectrum for folic acid using the ratio between 10-keV argon and cesium
ions and (b) SSIMS spectrum (After Ref. [27]). In (b) peaks marked with an asterisk are charac-
teristic of PDMS

Figure6.13b shows the SSIMS spectrum for 10keV Cs+ primary ions and Fig.6.13a
the G-SIMS spectrum. The SSIMS spectrum is displayed with a logarithmic inten-
sity scale that clearly shows the large population of degraded fragment products.
The five distinct molecular fragments shown with an asterisk in Fig.6.13b, and
observed at masses 73.054 u, 147.079 u, 207.043 u, 281.067 u, and 324.996 u, are
the well-known fragments from polydimethylsiloxane (PDMS) contamination. The
G-SIMS spectrum is much simpler, with four dominant peaks at masses 120.051 u,
176.065 u, 177.077 u, and 295.094 u. It is interesting that the PDMS peaks are no
longer dominant in the spectrum. The accuracy of the mass-scale calibration is suf-
ficient to assign the fragment formula C7H6NO+ to the peak at mass 120.051.
However, it is not possible to assign a unique fragment formula to the mass peak
295.094. It could be a number of possibilities from the folic acid molecule. For
example, it could be subunits ++++ and an additional oxygen giving the
composition C13H15N2O5+ with a mass of 295.0930, or alternatively it could be sub-
units ++ with a chemical composition C14H11N6O2+ and a mass of 295.0940.
A mass-calibration accuracy of better than 4ppm would be necessary to interpret
this difference. What is needed to identify this peak is structural information, and
we show how this can be done in the following section.

6.4 G
 -SIMS Fragmentation Pathway Mapping
(G-SIMS-FPM)

With a careful calibration of the mass scale [18] and instrument optimization, in
tof-SIMS one can achieve a 10-ppm accuracy of the mass scale [18]. Unfortunately,
this is insufficient to separate the many chemical permutations accommodated by
this level of uncertainty. Traditionally in mass spectrometry, an MS/MS experiment
would be performed, fragmenting the molecular ion of interest. This is usually
achieved via a low-energy collision with target atoms, typically of an inert gas,
6 Molecular Structure and Identification Through G-SIMS and SMILES 157

followed by a mass analysis of the fragmentation products. In ion-trap systems, this

process may be repeated many times depending on the number of initial ions avail-
able and is thus called MSn. This is a very powerful method and is routine for
molecular identification. Unfortunately, at the present time none of the commercial
ToF-SIMS instruments, widely used by industry and academia, has MS/MS capa-
bility. This is because of the need for a high repetition rate (100 us) for fast imaging
and also the small volumes of material that are available from the surface. For exam-
ple, in a monolayer, a 200-nm200-nm square pixel with 1% ionization efficiency,
only 10 molecules are available, assuming a transmission of 50%. New opportuni-
ties now exist, with the advent of organic depth profiling using cluster ion beams
such as C60, where the available material for analysis is no longer constrained to the
outermost molecules at the surface by the static limit (only around 1% of surface
atoms have a primary ion impact). Consequently, the available molecular signal is
much higher (provided the molecule is not just at the surface!). Vickerman and
coworkers are exploiting this in a pioneering new instrument incorporating ToF-ToF
capability [24] in collaboration with Ionoptika Ltd. [25]. This is a very powerful
instrument with a high duty cycle (rapid depth profiling) that gives molecular struc-
ture information through MS/MS and a mass-scale calibration that is decoupled
from the SIMS emission process. Similarly, Winograd and coworkers have fitted a
C60 ion source to a traditional LCMS/MS quadrupoleToF mass spectrometer [26];
such mass spectrometers have a very wide user base in the mass spectrometry com-
munity. These innovative approaches have a very powerful potential, especially for
biological applications.
G-SIMS may also be used to provide molecular structure information. In Fig.6.4,
we illustrated the SIMS emission process with intact fragments emitted from a low
surface plasma temperature (outer zone) and more degraded fragments emitted from a
high-surface plasma temperature. We may explore the fragmentation behavior between
these conditions by changing the parameter g in the G-SIMS equation, Eq.6.3. This
concept is illustrated with the folic acid molecule in Fig.6.14. In the G-SIMS regime
(g=13), the spectrum is dominated by the intact parent molecules; as the plasma tem-
perature is increased (by lowering g), the spectrum is populated by more and more
fragmentation products until at g=0 or 1 the original static SIMS spectra are obtained
with their high populations of fragmentation products. We may therefore map out
these fragmentation pathways; the method to do this, G-SIMS fragmentation pathway
mapping, is explained in the following, exemplified with folic acid.
We may investigate the effect of varying the surface plasma temperature, Tp, by
changing the G-SIMS index, g, of Eq.6.3 from high fragmentation to low fragmen-
tation and mapping out the corresponding variation in G-SIMS intensities.
Figure 6.15 shows this for values of g from 0 to 40. With g=0, we begin in the
SSIMS regime. As g is increased, we progress to the G-SIMS regime (lower surface
plasma temperature). First, we see the substrate and low-molecular-weight frag-
ments decay rapidly. At the same time, the intensities of more intact fragments
begin to grow, but even larger intact fragments that grow more strongly soon over-
take these. Consequently, the intensities of the smaller degraded fragments begin to
decay. This evolves for larger and more intact fragments until eventually the
158 F.M. Green et al.

Fig. 6.14 Illustration of how SSIMS

the fragmentation pathways 0
may be mapped out in
G-SIMS-FPM by varying the
surface plasma temperature

Plasma temperature
using the parameter g

40
G-SIMS

Increasing Tp

102
++

M -H {NH}+
G-SIMS normalised intensity

Si gmax
101
+NH
C2H3O
H
C3H7O
+H2
100

++

H
C2H5

10-1
0 5 10 15 20 25 30 35 40
g

Fig. 6.15 Effect on G-SIMS intensities using the ratio between 10-keV argon and cesium ions
with increasing g from 0 to 40

dominant parent fragment has the highest G-SIMS intensity. For each fragment, the
G-SIMS intensity goes through a maximum at a value gmax that is characteristic of
the plasma temperature [27].
This process is conducted independently for both the positive- and negative-ion
data. The gmax values for each fragment and for both polarities are brought together
in a single plot. In Fig.6.16, we show a plot of the gmax values for folic acid, with
6 Molecular Structure and Identification Through G-SIMS and SMILES 159

Fig. 6.16 Reassembly plot 102

of gmax for folic acid using the
ratio between 10-keV argon
and cesium ions. Two 2
assembly pathways with sets 3
of co-daughters are shown as
solid and dashed lines. One
further pathway for
1

gmax
granddaughters is shown 101 4
dashed. That pathway must
also emit a +CH3O subunit
as either a negative or neutral
fragment and is not shown

100
0 100 200 300 400 500
Mass, u

positive-ion data plotted with a red plus symbol and negative-ion data with a black
hollow diamond symbol. We analyze the fragmentation process by first selecting
the highest-mass fragment with the highest value of gmax. This is the parent fragment
with a mass MP. We next choose a fragment at a lower mass with the next-lowest
value of gmax. This is a trial daughter fragment with mass MD. We now postulate that
the parent fragmented into the daughter (which exhibits a higher surface plasma
temperature, higher gmax) with up to two fragmentation products, co-daughters, with
masses MC1 and MC2, respectively. All possible combinations of up to two fragments
are computed from those fragments present in the G-SIMS spectrum (Fig.6.13a).
The positive fragment ions in the G-SIMS spectra are likely to be protonated, and
so the possibility of one or two additional hydrogen atoms with mass nH is permitted,
such that
M P = M D + M C1 + M C 2 + nH, (6.5)

where 2n2.
The possible co-daughter fragments are then validated by calculating their com-
position from the peak centroid mass and comparing with chemically possible struc-
tures. Typically, MCMP, and the chemical composition may therefore be assigned
with higher certainty. Those postulated co-daughter fragments containing Si from the
substrate or other elements known to be absent in the molecule such as Na or K may
be immediately ruled out. In Fig.6.16, we map out two of the likely principal frag-
mentation pathways (i.e., sets of co-daughters) elucidated using the method above.
With careful mass measurement [18], we may be able to see that fragmentation path-
way 1, in red, leads to three products, which are identified as +2H, H, and H
(with nominal masses 106, 28, and 161 u), so that the parent ion for that pathway is
++ (with mass 295 u), where the symbol order does not necessarily define the
order in the molecule. This represents one end of the folic acid molecule and provides
160 F.M. Green et al.

the identity of the daughter for fragmentation pathway 3, in magenta. The co-daughter
is identified as ++ (with mass 145 u), which then leads to the identification of the
entire molecular structure +++++ (with mass 440 u). Fragmentation path-
ways 2 and 4, in orange and blue, respectively, support these assignments. The frag-
ment at mass 295.0940.003 is now clearly shown to be the fragment [++]+; this
demonstrates the importance of a combined approach with positive and negative ions
and the ability of G-SIMS-FPM to identify a molecule where the mass alone is insuf-
ficient to do this [27]. By following the fragmentation pathways from a large molecu-
lar ion to smaller entities, we have identified those entities. The original parent
molecular structure is then reassembled in a manner similar to traditional MS/MS
experiments. So far, we see how the molecular structure may be reassembled using
G-SIMS fragmentation pathway mapping for folic acid. Of course, in this example,
the molecular structure of the parent molecule is known a priori. Typically, the
molecular structure that must be identified is an unknown, which makes the interpre-
tation of the fragmentation pathways significantly more challenging. In the next sec-
tion, we describe the development of a system for analysts, to guide them through the
reassembly process, providing suggested options.

6.5 SMILES

In this section, we show how a novel approach based on SMILES (simplified molec-
ular input line entry specification) [6] is used to identify the reassembly process,
through evaluation of fragmentation pathways to build up molecular structures.
Abrief introduction to SMILES is given in Sect.6.5.1. To assist analysts in the
interpretation of the fragmentation pathways, we outline in Sect.6.5.2 a method to
simulate the fragmentation pathways. Examples of simulated pathways and their
complementarities to G-SIMS are shown in Sect.6.5.3, for a wide range of different
biological molecules. Section6.5.4 discusses how these can be used to aid the ana-
lyst in producing fragmentation pathways and reassembling molecular structures.
The prediction of mass spectra from structural information, and vice versa, has
been a longstanding challenge since the mid-1960s when computers became suffi-
ciently powerful to study small molecules. The DENDRAL project [28], which
took place over the period 1965 to 1990, was the first. Gasteiger etal. [29] used a
different approach, an automatic algorithm, FRANZ, to predict a mass spectrum
from the molecular structure, including details of fragmentation and rearrange-
ments occurring in the spectrometer. An excellent agreement is shown between
predicted and experimental spectra for small molecules with around 10 constitu-
ents, excluding hydrogen. However, it is clear that these systems are not, as yet,
readily applicable to complex molecules, as evidenced by the usage in the commu-
nity. We show that our simple system based on SMILES, coupled with the unde-
graded G-SIMS spectra, forms a powerful combination for the identification of
complex molecules.
6 Molecular Structure and Identification Through G-SIMS and SMILES 161

6.5.1 S
 implified Molecular Input Line Entry
Specification (SMILES)

Simplified molecular input line entry specification (SMILES) allows the structure of
a molecule to be unambiguously expressed in a logical, computer-readable way using
an ASCII text string. SMILES was originally developed by Weininger etal. [30] and
more recently developed by Daylight Chemical Information Systems [6]. The
SMILES format is a very popular format used in informatics and is integrated into a
wide range of software freely available on the Internet to give, for example, molecu-
lar structure [31, 32], chemical information, such as pKA [33], solubility [33], molec-
ular volume [32], and log P values [32], as well as physical parameters such as
density [33], refractive index [33], molar mass [34], and mass spectral isotope pat-
terns [34]. The syntax and grammar of SMILES are described in detail elsewhere [6,
30]. In brief, each element is denoted by its chemical symbol, with the first letter
given in the upper case. Adjacent atoms in the SMILES string (read from left to right)
are considered to have a single bond unless otherwise denoted by an inserted = or
# character representing double and triple bonds, respectively. Hydrogen is implicit.
Branches are described using parentheses, and it is possible to nest parentheses for
subbranches. Cyclic structures are defined by a number identifying which atoms
open and close the ring structure, so that cyclohexane is represented by C1CCCCC1.
Aromatic atoms are shown in lower case, so that benzene is represented by c1ccccc1.
Ionic bonds may also be described as well as chirality (not used here). Canonical
form ensures that the SMILES structure is unique for a given molecular structure. For
example, CC(C(=O)O)N represents the following structure (without chirality):

6.5.2 Simulated Fragmentation Pathways

A computer program developed at NPL [35] using MATLAB (The MathWorks,

Natick, MA, USA) simulates full fragmentation pathways for any molecule given in
the SMILES format. This starts with the parent molecule of unlimited complexity
and cleaves it at a bond into two parts. This is repeated for each bond in the parent
molecule, and the fragment products are listed in the first level of a tree structure, as
illustrated in Fig.6.17. For example, an illustrative molecule ABCD would be frag-
mented into six substructures, A, BCD, AB, CD, ABC, and D. Clearly, the
162 F.M. Green et al.

Fig. 6.17 Illustrative example of the fragmentation pathways mapped out in G-SIMS for positive
and negative ions (left) and the complementary simulated pathways using the SMILES fragmenta-
tion software (right)

fragmentation terminates for A and D, but, for example, BCD may be fragmented
into four further subunits, B, CD, BC, and D, giving a second level of fragmenta-
tion. Similarly, CD and BC may be fragmented to a third level of fragmentation. In
this example, the third level of fragmentation is the final level, so that at that stage,
the molecule has been fragmented into its constituent elements. At the bottom, in
the right-hand example in Fig.6.17, we would usually finish with fewer possibilities
in the final level and possibly, for example, just C, H, and O. It is also easy to see in
the above simple example that the subunit BC will be found from both the fragmen-
tation of BCD and of ABC. There is no need to fragment the same structure twice;
to save computer time for each fragment, the computer checks in a look-up table to
see if a fragment has been analyzed before. If so, a link is stored and the fragmenta-
tion is terminated. Later we shall see how this can be extended to include fragments
in a larger database for many parent molecules. This method significantly reduces
computational time. The computer program uses a structured array with each
element, containing details of the fragment, including SMILES text string, exact
mass, and links to other parts in the fragmentation pathway with the same fragment.
The exact mass is calculated by a separate program that for a given SMILES text
string sums the main isotope masses of all the constituent elements as well as the
number of implicit hydrogen atoms through the determination of atom valency and
bonding. We use these values later to compare with experimental G-SIMS data.

6.5.3 S
 imulated SMILES Fragmentation Pathways
for Biologically Relevant Examples

The SMILES text for folic acid is readily available from the web as follows:

OC ( =O ) CCC ( C ( =O ) O ) NC ( =O ) c1ccc ( NCc2nc3c ( O ) nc ( N ) nc3nc2 ) cc1.

This structure is provided as input to the fragmentation pathway simulator
software and the fragmentation pathways are calculated. On a Pentium PC with a
6 Molecular Structure and Identification Through G-SIMS and SMILES 163

Fig. 6.18Simulated 0
fragmentation pathways 33
displayed on a plot of 44
1
fragment mass and 1
fragmentation level for folic

Fragmentation step
2
acid (From Ref. [35]). For 2
clarity, not all fragmentation
levels are shown. The 3
fragmentation pathways
identified in Fig.6.14 are also
4
highlighted for comparison

5
16

17
0 100 200 300 400
Mass, u

2.66-GHz processor, this takes approximately 20minutes. Fragmenting the

molecule to the constituent atoms C, N, and O requires 34 levels of fragmentation.
However, it is assumed that stable ring structures, such as benzene, are unlikely to
fragment, such that folic acid only requires 17 levels of fragmentation. The output
of the program contains a structured array, described earlier, containing each frag-
ment structure in SMILES text and the main isotope accurate mass position. For this
system to be useful, it must be able to describe, say, at least 75% of the fragments
observed in the key G-SIMS spectrum. For folic acid, simulated fragments account
for 90% of peaks with an intensity of >0.02. This excellent agreement between
experimental G-SIMS data and our simple system for simulating fragments is
because the G-SIMS spectrum consists of simple fragments that are undegraded and
have not undergone postemission rearrangements [7].
The simulated fragmentation pathways stored in the structural array are essen-
tially organized in a tree structure. However, for analytical purposes, this is not so
convenient, as fragments of a similar mass are not necessarily juxtaposed. Instead,
in Fig.6.18, we show the simulated fragmentation pathways for folic acid as a
mass-based tree structure with the fragmentation level on the ordinate and the frag-
ment mass on the abscissa. So, in the top right we start with the parent molecule
going through successive fragmentations until eventually we end up with the con-
stituent structures (e.g., benzene) and elements C, N, O in the bottom left. This plot
is then similar to the G-SIMS reassembly plot shown earlier in Fig.6.16. Each of
the four fragmentation pathways identified earlier may also be found on the simu-
lated fragmentation pathway, and these are highlighted. At the 16th and 17th frag-
mentation levels, all of the structural information has been lost and is shown here for
completeness. At the 17th level are C, N, O, C6H6, and C6N4H4. In practice, the 10th
level of fragmentation is typically required for structurally significant peaks.
The natural lexicon of amino acids in peptides and proteins provides a useful
system to study the effectiveness of the simulated pathways [36] further as well as
164 F.M. Green et al.

OH
NH2

HO CH CH3

0 C CH
0 H2C

a b
O CH3
CH O
H2N C

CH3 OH

Fragmentation step
1
Fragmentation step

1 HO
H
C
O
H2N
C
CH
CH3
O
C
OH
H2 C OH
H2 H
H2 N CH2

H2C
H CH

2 H3C
C
CH3
H3C CH

CH3
O
2 H2N CH3 H3 C OH
H2
2

3 H2N CH3
3 H2N
CH33

CH2

4 4

5 5
0 20 40 60 80 100 120 0 50 100 150 200
Mass, u Mass, u
CH3

H3C CH

CH NH H2C OH

HO C C CH

H3C

H3C HO CH CH3

0
c
C CH O
HO CH CH3
O HN C

C CH CH NH2

HO CH2
O NH2

1 HO O
HO CH2

CH NH NH2
Fragmentation step

HC
H3 C HN C C CH
CH
CH CH2 O O CH2
C NH2
H2 H3C H3C

2
H3C OH

3 H2N
C
CH3

CH
CH3
H2

5
50 100 150 200 250 300 350 400
Mass, u

Fig. 6.19 Simulated fragmentation pathways for (a) valine, (b) tyrosine, and (c) a simple peptide
valinetyrosinevaline (After Ref. [36]). Those fragments with strong intensities in the static
SIMS library spectra [13] are marked with a filled circle, with the fragment structure shown
directly from the SMILES text for that fragment

being of significant practical importance. In addition, the small base set of only 22
amino acids provides a good starting place for the development of a fragment
library, which, if successful, will be of significant help in the identification of
peptides in static SIMS. Figure6.19a, b show the simulated SMILES fragmenta-
tion pathways for the amino acids valine, OC(=O)C(N)C(C)C, and tyrosine,
OC(=O)C(N)Cc1ccc(O)cc1, respectively. For comparison with static SIMS, those
fragments exhibiting strong secondary ion intensities are marked on the fragmenta-
tion pathways with filled circles. The SMILES fragmentation pathway accounts for
all the major peaks identified in the SIMS library spectra [1] for both valine and
tyrosine. The simulated pathways also include the characteristic ion fragments iden-
tified previously in Ref. [37]. Typically, in the SIMS analysis of proteins, the frag-
mentation is extensive and significant secondary ion intensities are observed for
6 Molecular Structure and Identification Through G-SIMS and SMILES 165

individual amino acids rather than pluralities. The identification of pluralities of

amino acids, even if weak in the SSIMS spectrum, would add significantly to the
ability to identify or characterize proteins. In Fig.6.19c, we show the simulated
fragmentation pathway for a simple, three-amino-acid peptide consisting of valine
tyrosinevaline (VYV). Clearly, there are many fragments containing information
about juxtaposed amino acids. Those information-rich peaks, while weak in the
static SIMS spectra, are more prominent in G-SIMS spectra [35].

6.5.4 I nterpretation of G-SIMS Fragmentation Pathways

Using Simulated Fragments

Section6.4 shows how a molecular structure can be reassembled, by working from

small fragments up fragmentation pathways to reach the molecular structure. For
the illustrative example of folic acid, this was helped by already knowing the molec-
ular structure. In the previous section, we outlined a method for simulating the frag-
mentation pathways for molecules without experimental data. By combining the
two sections, we can produce a simple system to guide analysts through the reas-
sembly of a molecule to reduce the possible choices for an unknown structure.
An analyst may select any fragment in the G-SIMS fragmentation pathway, and
the mass of this fragment will then be automatically compared with a simulated
library of fragments containing many types of parent molecules. A suite of possible
fragmentation routes may then be identified to build the fragment to one of the
larger masses that are contained in the experimental data. This guided process may
then be followed until reaching the largest fragment in the experimental data. This
may not identify a unique solution but may produce a reduced set of options. This
accelerates and improves the reliability of establishing the molecular structure and
molecular identity.
A great advantage of a simulated fragmentation library is that it does not rely on
the contribution of experimental data from the community, which is always the rate-
limiting factor. An additional advantage is the equivalence and reproducibility of
data generated from a controlled computer program compared with a wide range of
instrument designs and configurations in experimental data. One mechanism to
grow the library very rapidly is to provide the SMILES simulation fragmentation
pathway software freely on the web, for the community to use on molecules of
interest to them. The generated simulated fragmentation pathways can then be
added to an open-access central web library, which is freely accessible to all users.
This approach is now being developed and will be tested.
In the development of a fragmentation library, it is clear that archetypal molecu-
lar entities only need to be fragmented and stored once. These become hubs in the
library, marked in Fig.6.19c for valine (V) and tyrosine (Y). In the case of peptides
and proteins, previously fragmented amino acids or peptides will form hubs that can
be directly referenced to the SMILES fragmentation library, simplifying the
166 F.M. Green et al.

Fig. 6.20 An illustration

of the way a SMILES These reflect fragments from V or Y
fragmentation library
will build up modules 0 VYV

Fragmentation step
simplifying the
fragmentation maps
1 V

Pre-existing hubs
2 Y

4
Mass, u

fragmentation map. Figure6.20 illustrates how each new hub becomes an integral
component part (i.e., module) of the SMILES fragmentation library. With more
components in the library, building the library becomes increasingly easy.

6.6 G-SIMS Imaging

A recent innovation by ION-TOF in collaboration with NPL [19] allows G-SIMS

to be conducted using the popular Bi+ liquid metal ion source. The liquid metal ion
source using Aun+ or Bin+ has rapidly become the standard ion gun on all new
instruments, making G-SIMS accessible to a much wider user community.
G-SIMS is now facilitated by a special emitter known as the G-tip, which is com-
prised of Bi and a small amount of Mn. This allows Bin+ or Mn+ primary ions to be
selected and focused to the same point. In the G-tip, the two ion beams are auto-
matically aligned, allowing G-SIMS imaging. These two primary ions produce
well-separated surface plasma temperatures, as may be expected from Fig.6.8.
Indeed, recent calculations show that it would be very difficult to improve on this
choice for generating G-SIMS spectra [38]. An example G-SIMS spectrum using
the G-tip is shown in Fig.6.21a for polycarbonate (PC); it shows the key charac-
teristic C9H11O+ peak. Indeed, the G-SIMS effect is so strong that at a g of 13, this
is the only peak in the spectrum! The capability of the G-tip for G-SIMS imaging
is shown in Fig.6.21c, d using a polycarbonate sample patterned with a surface
coverage of Irganox 1010. The sample is prepared by thermally evaporating
Irganox onto the PC sample masked with a 125-m mesh TEM grid. Figure6.21c,
d show that the dominant G-SIMS peaks may be found for both regions of interest,
and the intensity maps of these key peaks are shown to be strongly characteristic.
A line profile of the image [19] shows that the Mn+ and Bi+ images are correlated
to better than the beam size that defines the resolution of the image here, in this
case about 5m. The ease with which G-SIMS can be conducted, extremely effec-
tively, from a single ion-beam column using the G-tip significantly enhances the
prospects for G-SIMS.
6 Molecular Structure and Identification Through G-SIMS and SMILES 167

a 1
b 1

135.0818 u
M-H R

899.6095 u
CH2 C R
0.8 0.8
Normalised Intensity

Normalised Intensity
R

0.6 0.6

0.4 0.4

205.0746
0.2 0.2

00 100 200 300 0

0 500 1000
Mass, u Mass, u

c d

Fig. 6.21 Positive ion G-SIMS spectra from (a) PC areas and (b) Irganox 1010 areas, of a two-
phase material. Secondary ion images of the corresponding G-SIMS peaks are shown (c) for
135.0818 u, (d) for 899.6095 u (After Ref. [19]). The field of view is 250m250 m

6.7 Cluster Sources

For high-ion yields from organic layers, many analysts use Bi3+ or Bi5+ primary ions.
For the study of thicker layers of organic materials, argon cluster primary ions are
often used as they lead to much less degradation of the sample. These primary ion
beams may be incorporated into the G-SIMS scheme directly, replacing one of the
two ion sources traditionally used for G-SIMS. The choice of Bi3+ or Bi5+ with Ar+
or Mn+ is a practical combination that gives very strong signals. The difference
between Bi3+ or Bi5+ with Ar+ and a more traditional G-SIMS using Bi+ with Ar+ is
quite strong, and this leads to the effect that smaller g values are required to achieve
the G-SIMS result. In an example for the reference molecule of Irganox 1010 [39]
comparing Bi5+ with Ar to Bi+ with Ar, Seah and coworkers found that the effect of
g is about three times stronger. More importantly, the better quality of spectra for
the high-mass peaks using Bi5+ compared with Bi+ means that the G-SIMS for the
higher-mass entities is more reliable. The cluster primary ion sources thus have an
important role to play in the G-SIMS for secondary ions with masses above 500Da.
168 F.M. Green et al.

6.8 Future Outlook

The use of G-SIMS is growing, showing success for a widening range of materials.
The largest barrier to the uptake of G-SIMS in the SIMS community has been the
availability of suitable dual-source ion beams. Recent developments such as the
G-tip make G-SIMS available on the most popular ion beam sold on the majority of
new instruments. This will rapidly increase the accessibility. Together with easy-to-
use analytical toolssuch as Easy G-SIMS [23], a simple spreadsheet analysis of
dataG-SIMS is becoming more accessible to analysts. Future developments may
integrate G-SIMS processing into manufacturers software.
The development of G-SIMS-FPM with SMILES could provide a powerful
approach to identify molecules from the molecular structure that is not restricted by
the availability of experimental libraries. The proposed community-developed
SMILES simulated fragmentation library will aid rapid library development, as has
been demonstrated by many user-developed knowledge resources on the web. A fur-
ther advantage here is that this would all be done using a validated single algorithm.
The challenges analysts face are increasing in the world of rapidly developing,
complex technologies, such as the identification and distribution of molecules at
surfaces for organic electronics, displays, sensors, organically functionalized MEM
devices, and nanobiotechnology. G-SIMS provides the analyst with a powerful
method to simplify spectra of increasingly complex samples and move beyond this
to give a more direct interpretation. The newer developments of G-SIMS-FPM and
SMILES simulated fragmentation pathways show great promise for enabling ana-
lysts to obtain structural information of unknown biologically relevant materials
quickly and simply.
Chapters in this book have shown the power of SIMS for the analysis of biologi-
cal surfaces. G-SIMS adds to this for the identification of complex molecules and
interpretation of data through access to information in the mass spectrum that would
otherwise remain latent.

References

1. Vickerman JC, Briggs D, Henderson A, editors. The Static SIMS library, version 4. Manchester:
SurfaceSpectra. http://www.surfacespectra.com/simslibrary
2. Schwede BC, Heller T, Rading D, Niehius E, Wiedmann L, Benninghoven A. The Mnster
High Mass Resolution Static SIMS library. Mnster: ION-TOF; 2003.
3. Briggs D, Brown A, Vickerman JC. Handbook of static secondary ion mass spectrometry
(SIMS). Chichester: Wiley; 1989.
4. Lee JLS, Gilmore IS. The application of multivariate data analysis techniques in surface analy-
sis. In: Vickerman JC, Gilmore IS, editors. Surface analysis the principal techniques. 2nd ed.
Wiley, Chichester: UK; 2008.
5. Weng LT, Betrand P, Laver W, Zimmer R, Bussetti S. Quantitative surface analysis of styrene
butadiene copolymers using time-of-flight secondary ion mass spectrometry. Surf Inter Anal.
1995;23:879886.
6 Molecular Structure and Identification Through G-SIMS and SMILES 169

6. SMILES, Daylight Chemical Information Systems. http://www.daylight.com/smiles/

7. Gilmore IS, Seah MP. Static SIMS: Towards Unfragmented Mass Spectra - The G-SIMS
Procedure. Appl Surf Sci. 2000;161:465480.
8. Gilmore IS, Seah MP. G-SIMS of Crystallisable Organics. Appl Surf Sci. 2003;203204:
551555.
9. Gilmore IS, Seah MP. Organic Molecule Characterisation G-SIMS. Appl Surf Sci. 2004;
231232:224229.
10. Ogaki R, Green FM, Lic S, Vert M, Alexander MR, Gilmore IS, Davies MC. A comparison of
the static SIMS and G-SIMS spectra of biodegradable homo-polyesters. Surf Inter Anal. 2008;
40:12021208.
11. Ogaki R, Green FM, Li S, Vert M, Alexander MR, Gilmore IS, Davies MC. G-SIMS of
Biodegradable Homo-Polyesters. Appl Surf Sci. 2006;252(19):67977800.
12. Ogaki R, Green FM, Gilmore IS, Shard AG, Luk S, Alexander MR, Davies MC. Study of the
end-group contribution to ToF-SIMS and G-SIMS spectra of poly(lactic acid) using deuterium
labelling. Surf Inter Anal. 2007;39(11):852859.
13. Hawtin PN, Abel M-L, Watts JF, Powell J. G-SIMS of Thermosetting Polymers. Appl Surf Sci.
2006;252:66766678.
14. Gilmore IS, Seah MP, Green FM. Static TOF-SIMS a VAMAS Interlaboratory Study. Part I.
Repeatability and Reproducibility of Spectra. Surf Inter Anal. 2006;37:651672.
15. ISO 23830, Surface chemical analysisSecondary ion mass spectrometryRepeatability and
constancy of the relative intensity scale in static secondary ion mass spectrometry, ISO Geneva.
16. Repeatability and constancy spreadsheet. National Physical Laboratory, UK. http://resource.
npl.co.uk/docs/science_technology/nanotechnology/ssims/repeatabilityandconstancyspread-
sheet.xls
17. Gilmore IS, Seah MP, Green FM. Static TOF-SIMS a VAMAS Interlaboratory Study: part II
Accuracy of the Mass Scale and G-SIMS Compatibility. Surf Inter Anal. 2007;39(10):817825.
18. Green FM, Gilmore IS, Seah MP. TOF-SIMS: Accurate Mass Scale Calibration. J Am Soc
Mass Spectrom. 2006;17:514523.
19. Green F, Kollmer F, Niehuis E, Gilmore I. Imaging G-SIMS: A Novel Bismuth-Manganese
Source Emitter. Rapid Commun Mass Spectrom. 2008;22:26022608.
20. Gilmore IS, Seah MP. Static SIMS: A Study of Damage using Polymers. Surf Inter Anal.
1996;24(11):746762.
21. Gilmore IS, Seah MP. Electron Flood Gun Damage in the Analysis of Polymers and Organics
in Time of Flight SIMS. Appl Surf Sci. 2002;187:89100.
22. Shard AG, Gilmore IS. Analysis of metastable ions in the ToF-SIMS spectra of polymers. Int
J Mass Spectrom. 2008;269(1):8594.
23.
Easy G-SIMS. National Physical Laboratory, UK. http://www.npl.co.uk/server.
php?show=ConWebDoc.610
24. Fletcher JS, Henderson A, Biddulph GX, Vaidyanathan S, Lockyer NP, Vickerman JC.
Uncovering new challenges in bio-analysis with ToF-SIMS. Appl Surf Sci. 2008;252:126470.
25. Hill R, Blenkinsopp PWM, Barber AM. Ionoptika Ltd. www.ionoptika.com
26. Carado A, Kozole J, Passarelli M, Winograd N, Loboda A, Bunch J, Wingate J. Hawkin J,
Murphy J. Biological Tissue Imaging with a Hybrid Cluster SIMS Quadrupole Time-of-Flight
Mass Spectrometer. Appl Surf Sci. 2008;255:157215755.
27. Gilmore IS, Seah MP, Green FM. G-SIMS-FPM: Molecular Structure at Surfaces A Combined
Positive and Negative Secondary Ion Study. Appl Surf Sci. 2006;252(19):66016604.
28. Lindsay RK, Buchanan BG, Feigenbaum EA, Lederberg J. Applications of artificial intelli-
gence for organic chemistry: the Dendral project. New York: McGraw-Hill; 1980.
29. Gasteiger J, Hanebeck W, Schulz K-P. Prediction of mass spectra from structural information.
J Chem Inf Comput Sci. 1992;32:264271.
30. Weininger D. SMILES, a Chemical Language and Information system. 1. Introduction to
Methodology and Encoding Rules. J Chem Inf Comput Sci. 1988;28:3136.
31. SMILES, Daylight Chemical Information Systems, Depict. http://www.daylight.com/daycgi_
tutorials/depict.cgi
170 F.M. Green et al.

32. Molinspiration. http://www.molinspiration.com/cgi-bin/properties

33. Karickhoff SW, Carreira LA, Hilal SH. SPARC on-line calculator. http://ibmlc2.chem.uga.
edu/sparc/
34. Molar mass and mass spectral isotope pattern. http://www.colby.edu/chemistry/NMR/scripts/
smileMM.html
35. Gilmore IS, Seah MP, Green FM. Identification of Complex Molecules at Surfaces: G-SIMS
and SMILES Fragmentation Pathways. Int J Mass Spectrom. 2008;272:3847.
36. Gilmore IS, Seah MP, Green FM. G-SIMS and SMILES: Simulated Fragmentation Pathways
for Identification of Complex Molecules, Amino Acids and Peptides. Appl Surf Sci.
2008;252:852855.
37. Wagner MS, Castner DG. Characterization of Adsorbed Protein Films by Time-of-Flight
Secondary Ion Mass Spectrometry with Principal Component Analysis. Langmuir.
2001;17:46494660.
38. Seah MP, Gilmore IS, Green FM. G-SIMS: Relative Effectiveness of Different Monatomic
Primary Ion Source Combinations. Rapid Commun Mass Spectrom. 2009;23:599602.
39. Seah MP, Green FM, Gilmore IS. Cluster Primary Ion Sputtering: Secondary Ion Intensities in
Static SIMS of Organic Materials, J Phys Chem C. 2010;114:5351 5359.
Chapter 7
Imaging with the Helium Ion Microscope

John Notte and Bernhard Goetze

Abstract A newly developed technology, the helium ion microscope (HIM),

provides high-resolution imaging with several benefits compared to the standard
scanning electron microscope (SEM). First, the images provide high resolution
because the helium beam can be brought to a focused probe size that can be as small
as 0.25 nm. Second, the images provide contrast mechanisms that are often mark-
edly different from the SEM. These contrast mechanisms can reveal topographic,
composition, and other types of information about the sample. Third, compared to
the SEM, the HIM images tend to be more surface-specific revealing information
about the surface without the confusing subsurface information. Fourth, the HIM
can obtain high-resolution images even of insulating samples that would otherwise
charge excessively in the SEM. The HIM is still in its infancy compared to the SEM,
having only been commercially available for 7 years; however, it has already pro-
vided several unique advantages for the imaging of biological materials.

7.1 Introduction: On the Importance of Surfaces

In many imaging and analysis applications, surface specificity is crucial to produce

an easily interpreted image or data set. In fact, the nature of the human retina (essen-
tially a 2D sensor) limits our eyes to gathering 2D information. Also, for uncounted
years, our brains have been optimized for interpreting this 2D information as sur-
face-specific information. The mixing of deeper information with the surface infor-
mation tends to make the images harder to understand or downright deceptive (think
smoke and fog). In a mathematical sense, the determination of 3D information from
a 2D data set is a noninvertible problem. In other words, a single image does not

J. Notte (*) B. Goetze

Director of Research and Development, Carl Zeiss Microscopy,
Peabody, MA, USA
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 171

DOI 10.1007/978-3-319-01360-2_7, Springer International Publishing Switzerland 2014
172 J. Notte and B. Goetze

provide enough information to unambiguously reveal both surface information and

the deeper information. Hence, an imaging technique that provides surface-specific
information offers a less ambiguous interpretation.
In many applications, surface specificity is of primary importance because the inter-
face itself is the subject of the investigation. To a large extent, this is because the
interfaces define the boundary between objects, and it is through these interfaces
that objects can interact. Challenging applications of this nature occur in biology
(e.g., cell membranes), semiconductor physics (e.g., doped silicon junctions), and
material science (e.g., catalysis and corrosion). In most of these applications, the
surfaces are critical because the most interesting chemical processes are limited to
interfaces between different domains.
Many of the well-established high-resolution imaging techniques provide an
ambiguous image that mixes the surface information with the deeper information.
The well-established techniques of transmission electron microscope (TEM) and
scanning transmission electron microscope (STEM) provide excellent resolution
only after preparation of a ~100 nm thick lamella through which the beam passes.
The lamella preparation can be time-consuming and there is a risk of damaging the
specimen in the process. Because the beam passes through the lamella, the contrast
in the image represents an average along the beam path through the sample rather
than the true surface information. In addition, while the atomic force microscope
(AFM) provides high-resolution surface analysis, it is limited to the subset of sam-
ples that can be directly contacted and have a very limited topography a very small
subset of biological specimens. The most routinely used high-resolution imaging
instrument for biological samples is the scanning electron microscope (SEM). The
images it provides are high-resolution and do provide a good image contrast that
reveals different properties of the specimen. However, the SEM is known to produce
much of its signal from the reemergence of backscatter electrons, which reveal
SE2 information about deeper layers a topic of discussion in later sections here.
This effect can be mitigated by operating the electron beam at lower energies, but
this tends to limit the lateral resolution by chromatic aberration effects. The opera-
tion of the SEM also induces charging on and under the surface of the sample
effects that can compromise the image quality and can even damage the specimen.
Metal coatings are sometimes used to improve the charging and surface specificity
of the SEM, but the coatings often obscure the fine-scale features of interest and
even also damage the sample. Operating the SEM in the presence of gases (such as
water or nitrogen) can help to mitigate the surface-charging effects, but subsurface
charging is not resolved, and the gases tend to limit the resolution.
In contrast, the newly developed technology of the helium ion microscope (HIM)
produces high-resolution images with inherent surface specificity. The surface-specific
nature of the images is a direct consequence of the physics of the helium ion as it
interacts with the specimen. The charging effects are also fundamentally different
from the SEM and are much more readily mitigated. Figure 7.1 is an HIM image of
the iron oxidizing Acidovorax Proteobacteria (strain BoFeN1). This sample was
originally isolated from the anoxic freshwater sediments of Lake Constance in
Germany and was provided by Martin Obst and Fabian Zeitvogel, University of
7 Imaging with the Helium Ion Microscope 173

Fig. 7.1 A high-magnification helium ion microscope image of Acidovorax (Sample provided by
Martin Obst and Fabian Zeitvogel of the University of Tuebingen)

Tuebingen, Germany. When grown in the presence of Fe(II), some cells tend to
form a crust of Fe(III) mineral needles. Imaging with an HIM reveals the structure
of those mineral crusts on the cell surface for the first time. Imaging in the SEM
requires platinum coatings, which incidentally adulterate the real structure and
introduce artifacts [1].
In this chapter, the basic technology of the HIM is first explained. Subsequently,
the unique interaction of the beam with the sample is described in detail for a range
of beam energies and samples. Specifically addressed are the image formation pro-
cess and the properties of the sample, which are revealed in the resulting image. The
unique charging advantages and the minimal sample preparation requirements are
then described in detail. Lastly, the future outlook of this technology is provided.

7.2 Technology of the Helium Ion Microscope

7.2.1 Overview

Much of the technology discussed here is contained within the Zeiss family of
helium ion microscopes: the ORION PlusTM and the ORION NanoFabTM. These
models have only recently become commercially available [2] after many years of
174 J. Notte and B. Goetze

Fig. 7.2 Diagram of the helium ion microscope showing the ion source, the ion column, the
sample, the detectable particles, and the detector

development [3]. A simplified diagram of the HIM is shown in Fig. 7.2. The helium
ion source produces an ion beam with a typical energy of 30 keV and an ion beam
current of about 100 pA. The beam enters the ion column, which includes steering
deflectors, apertures, stigmators, scanning deflectors, and focusing elements. As the
beam exits the column, it is focused to a very small probe size on the surface of the
specimen. Typically, this probe size is 0.5 nm although measurements as small as
0.21 nm have been attained. The specimen can be virtually any shape or size, and a
mechanical stage allows the sample to be reoriented to provide alternative perspec-
tives. The entire beam path and specimen are maintained under a vacuum of better
than 2 107 Torr.
As the focused beam strikes the sample at a particular location, it produces a
number of particles that can subsequently be detected. The properties of the gener-
ated particles (their abundance, energy, angle, etc.) reveal some property about that
particular location. The beam is then advanced to a new location (perhaps just 1 nm
away), and the emitted particles are again detected. The variation in the quantity or
properties of the generated particles provides the contrast from location to location
on the sample. The focused helium beam is advanced in a raster pattern across a
rectangular region of the sample (as shown in Fig. 7.2). The image is then assem-
bled on a pixel-by-pixel basis as the beam is advanced. The gray level of each pixel
is based upon a chosen property of these generated particles. For example, the pixel
may be assigned black if there are no secondary electrons produced, or white if 10
or more secondary electrons are produced at that location. The typical time to
acquire such an image can vary from 5 s to 5 min, depending on the signal-to-noise
ratio required and the number of pixels in the final image (Fig. 7.3).
In some respects, the HIM operates much like the traditional SEM or gallium
focused ion beam (FIB). In these regards, there are excellent textbooks that detail
7 Imaging with the Helium Ion Microscope 175

Fig. 7.3 The focused helium beam is moved from location to location in a raster pattern across the
sample. Each of the locations on the sample corresponds to a pixel in the final image. Each pixel is
assigned a gray level based upon the generated particles at the corresponding position

the basic physics that is common to their operation [4]. However, in many important
ways the technology of the HIM is distinctly different. Only the most significant of
these differences are therefore detailed in the following sections.

7.2.2 The Helium Ion Source

The helium ion source [5] is the key enabling technology for the HIM. A fuller
description of the technology can be found in the established literature [6], but a
cursory description is provided here. The ion source consists of a needle that is
drawn to an atomically sharp end form (Fig. 7.4). The ion source operates at a tem-
perature of about 70 K, in an ultrahigh-vacuum (UHV) vessel, with a large positive
voltage applied to it. The apex of this tip has an underlying spherical shape (radius
~1,000 ). Superimposed on this spherical shape is a three-sided pyramid terminat-
ing in an atomically sharp vertex. At this vertex, the single most protruding atom
experiences an electric field, which can be as large as 4 V/. In this otherwise UHV
region, ultrapure helium gas is admitted (impurity concentrations are typically 1
part in 106). Although neutral, the helium atoms become polarized and are drawn in
toward the tip in the presence of the field gradient. This effect creates a region of
elevated pressure surrounding the apex. The helium atoms are cooled by the process
of repeated collisions with the cryogenic tip and eventually come to thermal equi-
librium with the tip. As the low-energy helium gas atoms pass in the vicinity of the
most protruding atom, the large electric field can cause a single electron to quantum
mechanically tunnel [7] out of the helium atom and into the emitter tip. The remain-
ing positive helium ion is now repelled by the positively biased tip and is
176 J. Notte and B. Goetze

Fig. 7.4 The helium ion source produces helium ions originating from an atomically sharp asper-
ity at the end of a positively biased needle maintained at cryogenic temperatures

immediately accelerated away. The field ionization process, which is key to this new
microscope, was discovered over 50 years ago in the context of the field ion micro-
scope (FIM) [8, 9].
Several unique properties of this ion source make it so desirable for high-resolution
microscopy. First, the generated ions are produced from a region of atomic dimen-
sions, and the virtual source size (established from back-tracking the ions final
trajectories) is less than 1 in size. Second, the beam diverges very gradually, with
a typical emission cone semiangle less than 1. Together, these two properties are
reflected in the very high brightness of the ion source routinely measured to be
4 109 A cm2 sr1 for a 25 keV beam [10]. Another important attribute is the mono-
chromatic character of the ions. The energy spread, E, of the beam is found to be
about 1 eV or less, representing less than 1 part in 104 of the beam energy [11]. The
low-energy spread is important to minimize the energy-dispersive effects as the
beam is shaped and steered with electrostatic lenses and deflectors.
The usage of helium ions as opposed to lighter- or heavier-charged particles
is ideal for imaging applications and offers advantages over the competing tech-
nologies of electron and gallium ion beams. The mass of an electron is so small that
its wavelike properties begin to manifest themselves as the electron beam is focused.
In fact, a highly optimized SEM will have its probe size significantly limited by dif-
fraction [12]. For the helium ion beam, the de Broglie wavelength can be as small
as 100 fm not significantly affecting the focused probe size. Compared to gallium,
helium is light enough that it does not cause excessive damage to the sample. In
contrast, the massive gallium atom (atomic weight ~ 69 amu) is very effective [13]
in sputtering away any specimen in which it strikes. Helium is also optically trans-
parent, is chemically inert, and can diffuse out of biological specimens in relatively
short times. For these reasons, helium is a convenient ion species for a charged
7 Imaging with the Helium Ion Microscope 177

Fig. 7.5 A larger field of view of the same specimen shown in Fig. 7.1. This image shows that the
sharpness is uniform throughout the image for both near (lower boxed region) and far features
(upper boxed region). These are separated by an estimated 2 m in depth

particle microscope. It will be mentioned in later sections that the same technology
can be made to work with the heavier noble gas, neon. Such a neon beam offers a
different type of sample interaction that can induce deliberate erosion of the speci-
men in a controlled manner with nanometer-level precision.

7.2.3 Probe Formation

After the ions are emitted from the ion source, they are accelerated into the optical
column, which manipulates the beam to achieve the smallest attainable probe size
on the sample. The beam is extracted from the ion source and emerges with an
energy that varies from 15 to 45 keV. An electrostatic condenser lens is used to limit
the rate of divergence of the helium beam. The column includes static deflectors for
aiming the beam down the column and dynamic deflectors for scanning the beam in
a raster pattern across the sample. An aperture is used to select only the central por-
tion of the beam before it enters the final lens. Finally, the beam is focused with an
electrostatic lens to achieve the smallest probe size at the surface of the sample.
As the beam is approaching the sample, it is roughly conical in its shape, with a
convergence semiangle of less than 1 mrad. The small convergence angle also pro-
vides for a relatively long depth of field, making it easy to visualize samples with
high aspect ratios (Fig. 7.5). The size of the imaged area (field of view) can span
178 J. Notte and B. Goetze

from 100 nm to 1 mm, with as many as 2048 2048 pixels per image. The sample
can be mounted to a standard stub or glass slide and be positioned anywhere from
4 to 40 mm below the final lens. Under optimal conditions, the helium ion beam can
be focused to a probe size as small as 0.25 nm. Such images can resolve details
otherwise not seen with an SEM or FIB. The sample can also be biased positively
or negatively to enhance or diminish specific contrast mechanisms.

7.3 BeamSpecimen Interaction

As in an SEM or gallium FIB, the image generation in the HIM depends critically
on how the particles comprising the focused beam interact with the specimen.
Because the helium beam interaction is distinctly different compared to an SEM or
a gallium FIB, the image contrast is distinctly different. The physics of the helium
beam interaction is not fully understood, but several researchers [1416] have begun
unfolding the phenomena that underlie the images. These efforts will ultimately
explain the mechanisms by which the HIM produces its high-contrast, high-resolu-
tion images. But even without our having a complete theoretical understanding of
the contrast mechanisms, the HIM is establishing itself through the unique images
it produces.
A basic understanding of beamspecimen interaction is best approached by con-
sidering the behaviors of individual helium ions incident upon the specimen. The
fate of a single helium ion impinging upon a specimen can be understood from
fundamental physics primarily electrostatics and atomic-level scattering physics.
In almost all circumstances, the incident particles do arrive one at a time, and their
collision cascades are completed before the next particle arrives. The behavior of
individual ions can then be combined by statistical methods to provide the average
behavior of the helium ion beam. Such methodologies are commonly undertaken
with Monte Carlo computer simulations. The remainder of this section relies heav-
ily upon the IoniSE [16], TRIM [17], and Casino [18] computer programs to simu-
late the charged particles within the specimen.

7.3.1 Beam Penetration

A single helium ion interacts with the sample through a series of electrostatic inter-
actions between the ion and the target nuclei and electrons that comprise the speci-
men. Most of these interactions produce small angular deflections whereby the
trajectory of the helium ion is only slightly altered. These interactions tend to reduce
the energy of the incident ion with a statistically averaged stopping power mea-
sured in energy loss per traveled path length (eV/). The stopping power deter-
mines how the particle is slowed down and eventually comes to a stop, and
correspondingly how its energy is transferred to the specimen. The exact value of
7 Imaging with the Helium Ion Microscope 179

Table 7.1 The stopping power for 25 keV helium ions into select materials
Stopping power for 25-keV incident helium
Specimen (eV/)
Water 6.1
Adipose tissuea 7.3
Cortical bone 9.7
Zinc 13.0
Osmium 14.8
Gold 16.0
a
Berger MJ. Stopping powers and ranges for protons and alpha particles, International Commission
on Radiation Units Report ICRU-49. Bethesda, MD, USA: ICRU; 1993

Fig. 7.6 The stopping power tends to decrease with decreasing energy

the stopping power depends on the composition of the specimen and the energy of
the helium ion. The stopping power for 25 keV helium ions into select materials is
shown in Table 7.1.
The stopping power tends to decrease with decreasing energy, as shown in
Fig. 7.6. This energy is transferred from the helium ion to the electrons and nuclei
of the specimen by several mechanisms. The relative importance of the different
mechanisms may vary with beam energy and with the composition of the specimen,
but the following estimates are valid for 25 keV helium into silicon: About 80 % of
the energy transfer is to electrons including ionization of the atoms in the speci-
men. These excited electrons are responsible for the production of secondary
electrons which will be discussed in the next section. The remaining 20 % of the
energy is transferred to the nuclei of the sample, resulting in lattice vibrations
(phonons), recoiled target atoms, and occasional backscattering events. As the
180 J. Notte and B. Goetze

incident ion loses energy, the energy transfer to the nuclei tends to dominate over
the energy transfer to the electrons. Ultimately, the stopping power is responsible
for determining the average penetration depth of the helium beam.
As the helium ions penetrate into the surface of the specimen, there is a high
probability that the helium ion will capture an electron within a few nanometers of
the surface. Consider that helium is the most electron-greedy of all the residents
of the periodic table. The helium ion will spend most of the rest of its trajectory
(perhaps hundreds of nanometers) as a neutral helium atom. By this process of elec-
tron capture, the incident helium produces a very thin layer of positive surface
charge over the top few nanometers of the surface. This is distinctly different from
electrons, which are destined to keep their charge with them wherever they go
producing a deeper and widely distributed negative charging artifact. This simple
difference is a distinct advantage for the HIM over the SEM and will be discussed
in greater detail in a later section.
The helium ions predominant interaction with electrons tends to produce a rela-
tively small angular deflection of the incident ion, a consequence of the disparity in
the masses: MHe/me 7,300. But there is a nonzero probability that the helium ions
trajectory will put it in line with the nucleus of a target atom. In this case, there is a
strong electrostatic repulsion between the helium nucleus and the nucleus of the
target atom. To some extent, these nuclei will be partly shielded by the remaining
electrons. This resulting deflection is commonly known as Rutherford scattering
[19]. In some cases, the helium atom is scattered backward out of the sample. These
helium atoms are termed backscattered, and their detection for the purposes of
imaging is a subject addressed in the next sections. The probability of backscatter-
ing is typically 0.1 to 1 %, but this number increases for higher-atomic-number
targets or for lower-incident-energy helium ions.
Due to the random nature of the collisions, the trajectories of the individual
helium ions in the specimen vary considerably. Using the simulation software to
simulate many thousands of ion trajectories reveals the general depth and shape of
the interaction. From these simulations, it is also possible to determine the statisti-
cally averaged results, such as the average penetration depth and the average sub-
surface dispersion of the beam. The leftmost portion of Fig. 7.7 shows the general
shape of the interaction volume for a 30 keV helium beam incident on a silicon
specimen. The statistical nature of the scattering is evident in the varied trajectories
(100 are shown here). Below the surface, the ions diverge in the shape of a well-
defined cone (the typical cone angle is less than 5) before broadening into a more
spherical volume (not shown). The overall teardrop shape is common for many
target materials and medium- to high-helium beam energies. In this particular case,
the average incident helium beam penetrates to a depth of 350 nm. The shape of the
interaction volume and the width near the surface are of critical importance for
high-resolution-image formation the topic of the next section. For comparison, the
penetration of a 30 keV gallium beam is shown in the center of Fig. 7.7. For the
more massive gallium beam, the nuclear scattering dominates, and the beam pene-
trates less deeply and broadens rapidly as it produces many displacements (shown
as green dots) to the sample atoms. Surface sputtering is also quite significant for
7 Imaging with the Helium Ion Microscope 181

Fig. 7.7 Monte Carlo modeling results for representative beams into silicon. The red shows the
trajectories of the incident particles (100 are shown for each case). The green dots indicate atoms
of the specimen that have been displaced

such heavy ion beam species, causing the sample to be eroded as it is imaged. On
the right of Fig. 7.7, a low-energy electron beam is shown for comparison. The
electron beam interacts predominantly with the abundant electrons in the sample,
and owing to their equal mass, they can scatter through large angles. Consequently,
the beam disperses under the surface, and the interaction volume can be quite wide
at the surface.

7.3.2 Generated Particles and Suitable Detectors

As the helium ion beam penetrates into the specimen, there are several particles that
may be ejected from the specimen. The properties of these particles (including their
abundance, charge, emission angle, or energy) can reveal information about the
specimen that can vary from point to point (e.g., local topography or local composi-
tion). The detection of these particles, and their assumed variation from one location
to another, form the basis for the contrast in the image. In the following section, the
commonly encountered detected particles are discussed in detail.

7.3.2.1 Secondary Electrons

A secondary electron (SE) is any electron ejected from the surface with a kinetic
energy below 50 eV. As with the SEM and gallium FIB, the SEs are the most com-
monly used detectable particle for image formation in the HIM. This is in part
because of the relative abundance of SEs produced, and in part because of the
182 J. Notte and B. Goetze

relative ease with which the SEs can be detected. Images that are based upon SE
detection also reveal high-resolution and topographic information. The HIM has an
advantage in producing many more SEs per incident particle (2 to 5) compared to
the SEM, where the SE yield is typically less than 1.
As the helium ion enters the specimen, electrons are excited all along the ions
trajectory, but they can travel only a short distance (typically <3 nm) before their
energy is dissipated. So it is only those electrons excited sufficiently close to the
surface that may escape and be detected. Refer back to Fig. 7.7: The HIM has a
unique advantage over the SEM or Gallium FIB in that the excited volume is quite
narrow within the top few nanometers of the surface. Hence, the helium-induced
SEs that can be detected will convey information about the intended location of the
incident beam [20]. Compare this to the other beams, where the SEs can convey
information that is less local because the near-surface excited volume is consider-
ably wider. The result is a sharper image with the HIM.
The SEs that are generated from the incident beam as it first enters the sample are
termed type-one secondary electrons, or SE1, whereas if the SE is generated from
the incident beam as it backscatters deeply and again passes near the surface, it is
termed SE2. Because of the relatively low backscatter yield of helium (~1 % for
typical biomaterials), the SE signal consists almost entirely of SE1s and hence convey
only surface information. For the electron beam, however, there is often considerable
backscattering from deeper within the specimen, producing any number of SE2s as
the incident electrons reemerges from the surface. The likelihood of the high angle
scattering depends critically upon the subsurface composition, and hence the SEMs
detector sees both surface and subsurface information conflated together [21].
The helium-induced SEs that escape the specimen typically have an energy [22]
less than 2 eV. Due to their relatively low energy, their detection is relatively
straightforward; they can easily be drawn toward any nearby electrode that is posi-
tively biased. The same technology used in the SEM, the EverhartThornley detec-
tor, has been adapted for usage in the helium ion microscope (Fig. 7.8). It consists
of a highly transparent metal grid that is biased positively to about +500 V to attract
the electrons. Once within the grid, the electrons are accelerated toward a scintilla-
tor that is biased to +10 kV. Upon striking the scintillator, a single electron will
produce about 100 photons, which are optically guided down a light pipe and deliv-
ered to a photomultiplier tube (PMT). With a suitable gain, the PMT produces an
easily measured electrical signal for each detected electron. The detection efficiency
of such a detector depends critically upon the geometry of the sample relative to the
detector grid and the final lens. The detection efficiency is commonly 80 %, but it
tends to be reduced when the sample is very close to the final lens. A small positive
bias can be applied to the sample to improve the performance of the detector.
The HIM images that are based upon the detection of SEs convey a detailed sur-
face topography that is intuitively interpreted. The topographic information is evi-
dent in Fig. 7.9, which shows a complex three-dimensional arrangement of collagen
fibers from a rabbit knee. Even to the untrained eye, each fibers size, shape, banded
texture, and three-dimensional arrangement are easily recognized. The visual cues
that aid our interpretation include the characteristic bright-edge effect along the
7 Imaging with the Helium Ion Microscope 183

Fig. 7.8 A standard EverhartThornley detector is very effective in collecting secondary electrons
and producing an electrical signal

Fig. 7.9 The three-dimensional structure of collagen fibers is readily interpreted in this SE image
from the HIM. Below the image, an individual line profile between the two arrows shows the
bright edge effect (Sample provided by Wendy van den Berg-Foels, ClemsonMedical University
of South Carolina Bioengineering Program)
184 J. Notte and B. Goetze

Fig. 7.10 As the helium beam

enters the sample, electrons are
excited all along its trajectory.
When the beam enters at a glancing
angle, more of these can escape and
become detectable SEs, leading to a
bright edge effect. The dotted line
represents the characteristic escape
depth of secondary electrons

edge of each fiber and can be explained as follows (Fig. 7.10): Where the beam
strikes the sample at normal incidence, a small number (<5) of secondary electrons
are produced, resulting in a darker gray pixel. In contrast, when the incident beam
strikes the sample at a glancing angle, the beam remains close to the surface for a
longer path length, and hence more SEs can escape and result in a brighter pixel.
Experimentally, the bright-edge effect roughly corresponds to a brightness enhance-
ment factor proportional to sec(), where = 0 corresponds to the incident beam
being parallel to the local surface normal.
Figure 7.9 also demonstrates the excellent depth of field afforded by the HIM.
The more deeply situated collagen fibers are inherently darker because of the lower
probability of SE detection a visual cue that provides additional spatial informa-
tion. Note also that the thinnest fibers are somewhat transparent: There are detect-
able SEs from the front surface as well as the back surface, as well as in further
portions of the sample. This can lead to a transparency-like effect where the HIM
conveys front surface information, superimposed on a more diffuse back surface
information, superimposed upon a further portion of the sample.

7.3.2.2 Backscattered Helium

The small fraction of incident helium ions that undergo large angle scattering from
the nuclei of the specimen can occasionally escape from the specimen and subse-
quently be detected. These backscattered helium ions may still be in the form of
positive ions or they may be neutralized. The probability of backscattering is directly
related to the scattering cross section, which in turn depends on the atomic number
of the specimen atoms. According to the simple Rutherford scattering principles,
the scattering probability should increase with the target atoms atomic number
squared. Thus, an embedded gold nanoparticle would be more likely to produce a
7 Imaging with the Helium Ion Microscope 185

backscatter event than the lighter elements found in biological specimens.

Experimentally, this general trend is readily observed, but there are additional fluc-
tuations that are not fully understood [23]. The unexplained fluctuations seem to
correlate strongly with the group number (column number) of the periodic table,
with copper, silver, and gold having scattering probabilities that are higher than
most other elements in the same period (row).
The outcome of helium scattering can be most simply understood by considering
the conservation of momentum and energy. Solving these equations yields the fol-
lowing equation [24]:

2
M 2 - M 2 sin 2 q + M cos q
EB = E0 2 1 1
. (7.1)
M1 + M 2

Here EB is the energy of the helium atom just after the scattering, and E0 is the
energy just before the scattering. The masses of the incident helium ion and the
target atom are M1 and M2, respectively. is the angle by which the helium atoms
trajectory was changed (with 180 corresponding to a true backscatter). The simple
physics is complicated by the fact that the helium atom loses energy through its
many inelastic collisions throughout its trajectory (both before the scattering event
and after).
In the ORION Plus instrument, the backscattered helium can be detected by two
different methods. The first is a detector that produces a signal proportional to the
abundance of detected backscattered helium. This is achieved with an annular
microchannel plate (MCP) detector, which can be either retracted or inserted
between the sample and the final lens, as shown in Fig. 7.11. When inserted, the
MCP subtends a solid angle of about 1.8 sr. Because the backscatter rate is depen-
dent on the atomic number, this detector provides a useful contrast to distinguish
between heavy and light elements, with minimal topographic information. Because
of the low rate of helium backscattering, this detector requires comparatively high
exposures to generate images. The second option for detection of backscattered
helium is a detector capable of simultaneously measuring the angle and energy of
the individual backscattered helium atoms. This is achieved with a solid-state,
energy-resolving detector with a limited acceptance angle. In the ORION Plus
instrument, this detector has been used to identify unknown elemental composition
or to determine the thickness of thin films [25].

7.3.2.3 Other Detectable Particles

In addition to secondary electrons and backscattered helium ions, photons have been
observed to be emitted from some types of samples (Fig. 7.12). These have been mea-
sured with a simple PMT with a borosilicate glass window and a bi-alkali photocath-
ode. Several materials have been tested, but only a small fraction of them seem to
produce photons under the helium ion beam. The mechanism is not fully understood,
186 J. Notte and B. Goetze

Fig. 7.11 The MCP detector can be inserted for use and provides an electrical signal for each
helium atom (ion or neutral) that strikes it. Alternatively, the spectrometer can be used to analyze
the angular and energy distributions

Fig. 7.12 The specimen is a collection of table salt. In this image, the grayscale was assigned
based upon the abundance of detected photons

but two classes of mechanisms have been proposed. First, the photons can be the result
of relaxation of the target atoms that have been raised to a higher energy state by action
of the helium beam. This is akin to the cathodoluminescence (CL) effect, similar to
what has been observed with SEMs. The spectrum of this emitted light could then be
used to positively identify the target atoms by their emission spectra [26]. The other
proposed explanation (which has no SEM counterpart) is that the photons are the result
of the incident helium ions returning to a lower energy state. This process will produce
deep UV photons (~20 eV) as well as a broad range of visible and infrared photons.
7 Imaging with the Helium Ion Microscope 187

Fig. 7.13 Myelin sheets from a mouse cell as imaged with transmitted helium (Sample provided
by Prof. Schroeder from MPI Heidelberg)

For either mechanism, the detection of these photons may reveal important informa-
tion about the optical properties of the specimen in question. Recently, some HIM
owners have equipped their systems with optical spectrometers for collecting and ana-
lyzing the photons that are produced [27]. Of special interest are the well-established
fluorescent markers used in biomedical research.
Secondary ions and neutrals are known to be ejected from the sample when
exposed to the focused helium beam. These are sputtered atoms from the sample
and, as such, could provide useful imaging or analysis capabilities. A detector that
can measure the gross abundance of all such secondary ions (SI) can provide an
alternative imaging mode with a contrast that may be complementary to the SE
contrast. Beyond this, one of several mass spectrometry techniques [28] can ascer-
tain the atomic or molecular mass of the sputtered materials at the indicated loca-
tion. Although a secondary ion mass spectrometer (SIMS) is not presently available
for the HIM, there is ongoing work to develop such a commercially available detec-
tor for imaging and analysis [29]. Such an analysis technique could provide a mass
resolution sensitive enough to distinguish different isotopes of the same element
enabling the use of isotopic markers.
Another type of detectable particle is the transmitted helium. For samples that
are prethinned (typically 100 nm or less), the helium ions have a probability of pass-
ing through it with some angular deflection. The detection of the transmitted helium
at a certain angle can provide a useful contrast mechanism that could complement
the standard STEM imaging modes. An example of such an image is shown in
Fig. 7.13. The image shows the myelin sheets from a mouse cell after it had been
prepared by microtome and imaged in the ORION HIM with a transmission
188 J. Notte and B. Goetze

detector. In this case, the detector was configured for brightfield mode, collecting
transmitted helium ions that suffered minimal angular deflections.
Lastly, it should be pointed out that the HIM is a relatively new instrument, and
so there may be other particles that may be produced and may be detectable but have
not yet been investigated. As we do with other new imaging technologies (i.e., the
SEM in the 1950s), we anticipate the development of new detectors and new imag-
ing modes as the instrument gains wider usage.

7.4 Sample Charging and Sample Preparation

Imaging biological specimens in the SEM requires several preparatory techniques

to stabilize the materials to tolerate vacuum, to provide adequate contrast, and to
minimize charging artifacts. While these preparations have been well established
[30], they are known to introduce artifacts [31]. These artifacts are tolerable for
some applications, but under highest magnifications, the sub 10 nm details reveal
many distortions compared to the native state. In particular, the metallization of
insulating samples with platinum or carbon will destroy or at least obscure the finer
details. Operation of the SEM under high gas pressures (e.g., the environmental
SEM) is one alternative to metal coating, but again there are resolution disadvan-
tages that hide the sub 10 nm details.
Imaging with the helium ion microscope offers a unique advantage relative to the
SEM in minimizing charging artifacts when imaging insulating samples or samples
mounted to glass substrates. The first advantage is simply a result of the lower cur-
rents used in the HIM (1 pA typically) compared to the SEM (10 pA or more).
A more significant difference between the HIM and SEM is in the distribution of the
positive or negative charge. As mentioned earlier, the helium ions arrive in a single
ionized state, and as they enter the specimen, they are apt to become neutralized and
statistically spend most of the rest of their trajectory in a neutral state. Thus, beneath
the surface, there is no net charge transport with the helium beam. In contrast, the
electrons are deposited under the surface, where a negative space charge will accu-
mulate. At the surface there is a relatively minor difference between the HIM and the
SEM. The helium beam will produce positive surface charging because of two rea-
sons: (1) the neutralization of the incident helium, and (2) the ejection of secondary
electrons from the top few nanometers of the surface. The incident electron beam will
also produce a positive surface charge, but only because of the ejection of secondary
electrons from the surface. The effect is that the HIM produces only surface charging,
and it is always positive in sign, whereas the electron beam induces positive surface
charging in combination with negative subsurface charging. The net charging for the
SEM can be either positive or negative depending on the relative contribution of these
two effects. But even if they are exactly balanced and there is no net charging, there
is still an electric field established between the subsurface (negative) and surface
(positive) charging. The HIM comes equipped with a low-energy electron flood gun
that easily mitigates the HIMs positive surface charging. However, an electron flood
7 Imaging with the Helium Ion Microscope 189

Fig. 7.14 The charging characteristics of the HIM (left) and the SEM (right). The HIM produces
only positive surface charging within a few nanometers of the surface, which is easily mitigated
with an electron flood gun. The SEMs subsurface charging has no known remedy

gun, or any surface treatment, cannot resolve both the surface and subsurface accu-
mulated charge for the SEM. Figure 7.14 is a diagram comparing the charging situa-
tion for insulating samples from the HIM (left) and the SEM (right). For these
reasons, the HIM has provided some exceptional imaging results on insulating sam-
ples that otherwise provide a challenge to image in the SEM.
An example of HIM imaging of an insulating sample is shown in Fig. 7.15. This
image reveals the large-scale ultrastructure of mouse tooth enamel as imaged in the
HIM without any special metal coating. Tooth enamel consists of extremely long
and thin crystals of carbonated hydroxyapatite. The single crystals (70 nm wide and
hundreds of microns long) are arranged in bundles (so-called enamel prisms) that are
oriented in a species-specific and tooth-specific way to form a hierarchical structure
that is very hard, but not brittle, and lends the tooth its mechanical properties (hard-
ness and fracture resistance). The bundles represent former pathways of cells, con-
trolling the growth of both the crystals and the bundles. The unique three-dimensional
ultrastructure of the enamel represents a frozen map of the cell migration in the early
development of the tooth. An SEM microscopist might have turned to a metalliza-
tion technique in order to image this sample, but this would have compromised the
fine detail and would likely have produced incomplete coverage due to the wildly
varying topography.
The HIMs ability to image samples with minimal charging artifacts and with high
resolution and contrast offers the microscopist a chance to image his or her samples
190 J. Notte and B. Goetze

Fig. 7.15 Ultrastructure of mouse tooth enamel imaged without any metal coating in the helium ion
microscope (Samples provided by Felicitas Bidlack of the Forsyth Institute, Cambridge, MA, USA)

with much less sample preparation. For highly three-dimensional structures, this
avoids possible damage. Figure 7.16 shows a collecting duct of an intercalated renal
cell. The collecting duct of the kidney absorbs water and nutrients out of the primary
urine an ultra filtrate of the blood. The intercalated cells play an important role in the
acidbase homeostasis of the kidney. The intense membrane system of these cells
bears H+ ATPases regulating the pH of the primary urine. In the foreground (indicated
by arrows), a primary cilium of the principle cells is visible. The exact function of this
structure in the kidney remains unknown; one theory is that it has a role as a flow sen-
sor in the collecting duct. HIM imaging of those structures reveals details that have not
been observed before using conventional SEM.
The HIM offers the unique possibility to image biological samples with the with-
out the necessity of conductive coatings. This capability, together with the extremely
high resolution, can result in completely new insights on biological specimen. As
with all microscopy techniques, improving resolution always unveils new and excit-
ing details. It also means, though, that one enters the unknown with respect to what
exactly can be observed and interpreted. Preparing a wet biological sample, fixing it,
and drying it to make it vacuum-friendly inevitably changes the nature of the sample.
If the sample is then imaged uncoated, it represents nothing but reality or does it?
What happens if one can image without coatings in the sub 10 nm range is that
sample preparation artifacts become obvious instantly. The difficulty one faces in
omitting the well-known metal or carbon coatings is interpreting the results. Using
the HIM on biological samples is rewarding in many aspects, and this chapter tries
7 Imaging with the Helium Ion Microscope 191

Fig. 7.16 The intercalated and principal cells of the kidney (Sample provided by Dennis Brown
and Teodor Paunescu of Massachusetts General Hospital, Boston, MA, USA)

to convince the reader that it is worthwhile. It also means that the sometimes well-
established sample preparation protocols for the SEM are a good starting point for
HIM imaging but need to be modified for each sample type. The imaging of the
collagen fiber network (Fig. 7.9) needed extensive optimization steps [32] in the
preparation of the cartilage tissue. Figure 7.16 shows a rat kidney preparation that
was possible only after a thorough development of a sample preparation protocol
[33] for tissue samples. This protocol has been successfully adapted in the mean-
time for testis, retina, and inner ear samples (unpublished). In general, one has to be
extremely careful not to damage the ultrastructure of the sample; what this involves
is almost always very much dependent on what type of sample is under investiga-
tion. Once this hurdle is overcome, the results are very rewarding. Researchers are
only starting to investigate their structures of interest using this new technology, and
the results so far are very encouraging.

7.5 Future Outlook

The helium ion microscope remains a relatively new technology much as the SEM
was in the 1950s. Its strengths and weaknesses are still being recognized for bio-
logical and other application areas. New detectors and new methodologies are still
evolving. Already mentioned was the prospect of generating images from the
192 J. Notte and B. Goetze

detection of specific secondary ions, characteristic photons, and transmitted helium.

Other nonimaging applications include manipulation of the specimen to support
imaging. For example, recent work has demonstrated that this same technology can
be extended to ion beam species other than helium. A focused neon beam, for exam-
ple, has been generated with the same gas field ion source by changes to the emitter
and the gas supply. The heavier mass of the neon atom will sputter away materials
at a rate about 50 times faster than helium. Such a nanometer-sized focused neon
beam can serve as a nanoscale scalpel, able to remove material and expose hidden
features that can be subsequently imaged. It is conceivable that three-dimensional
information can be reconstructed through an alternating series of slicing and imag-
ing procedures. In a recent publication by Joens et al. [1], the sheath of a predator
nematoad was removed to expose the otherwise hidden tooth.

7.6 Summary

The newly developed helium ion microscope offers new imaging capabilities that
are distinctly different from the traditional gallium focused ion beam or scanning
electron microscope. The helium beam can be focused to a smaller probe size (as
small as 0.25 nm) and can provide a greater depth of focus than the competing tech-
niques. Most importantly, the helium beam interacts with the specimen in a dis-
tinctly different manner than electrons or heavier ions. The generated particles
(including secondary electrons, backscattered helium, and others) provide rich con-
trast mechanisms that give information about the topography, composition, and
other properties of the sample. For the imaging applications in biology, the HIM
offers some unique advantages. In addition to the high resolution and long depth of
focus, the helium beam can provide excellent contrast even on low-atomic-number
materials such as carbon. The images that are produced from secondary electrons
provide information that is specific to the top several nanometers of the sample.
Most importantly, the HIM can provide imaging with a minimal sample prepara-
tions, allowing the researcher to have greater confidence that preparation artifacts
are not occluding the features of interest. Also, insulating samples can be easily
imaged at high magnification without the usual degradations and artifacts seen in
the SEM.

References

1. Joens MS, Huynh C, Kasuboski JM, Ferranti D, Sigal YJ, Zeitvogel F, Obst M, Burkhardt CJ,
Curran KP, Chalasani SH, Stern LA, Goetze B, Fitzpatrick JAJ. Helium ion microscopy (HIM)
for the imaging of biological samples at sub-nanometer resolution. Sci Rep. 2013;3:3514.
2. Ward B, Notte J, Economou NP. Helium ion microscope: a new tool for nanoscale microscopy
and metrology. J Vac Sci Technol B. 2006;24:28714.
3. Economou NP, Notte JA, Thompson WB. Scanning. 2012;34(2):839.
7 Imaging with the Helium Ion Microscope 193

4. Goldstein J, Newbury DE, Joy DC, Lyman C, Echlin PE, Lifshin E, Sawyer L, Michael J.
Scanning electron microscopy and X-ray microanalysis. 3rd ed. New York: Kluwer/Plenum
Publishers; 2003. p. 49.
5. Tondare VN. Quest for a high brightness monochromatic noble gas ion source. J Vac Sci
Technol A. 2005;23(6):1498507.
6. Hill R, Notte JA, Scipioni L. In: Hawkes PW, editors. Advances in imaging and electron phys-
ics, vol. 170. Elsevier; 2013. p. 65128.
7. Oppenheimer JR. Three notes on the quantum theory of aperiodic effects. Phys Rev.
1928;31:6681.
8. Tsong TT. Atom probe field ion microscopy. New York: Cambridge University Press; 1990.
9. Mller EW, Tsong TT. Field ion microscopy, principles and applications. New York: Elsevier;
1969.
10. Hill R, Notte J, Ward B. The ALIS helium ion source and its application to high resolution
microscopy. Phys Procedia. 2008;1:13541.
11. Ernst N, Bozdech G, Schmidt H, Schmidt WA, Larkins GL. On the full-width-half-maximum
of field ion energy distributions. Appl Surf Sci. 1993;67:11117.
12. Reimer L. Scanning electron microscopy. 2nd ed. New York: Springer; 1998. p. 27.
13. Orloff J, Swanson LW, Utlaut M. Fundamental limits to imaging resolution for focused ion
beams. J Vac Sci Technol B. 1996;14(6):375963.
14. Inai K, Ohya K, Ishitani T. Simulation study on image contrast and spatial resolution in helium
ion microscope. J Electron Microsc. 2007;56:1639.
15. Bell DC. Contrast mechanisms and image formation in helium ion microscopy. Microsc
Microanal. 2011;17:14753.
16. Ramachandra R, Griffin B, Joy DC. A model for secondary electron imaging in the helium ion
microscope. Ultramicroscopy. 2009;109:74857.
17. Ziegler JF, Biersack JP, Littmark U. The stopping and range of ions in solids, vol. 1, Stopping
and ranges of ions in matter. New York: Pergamon Press; 1984. For the updated version, see
SRIM Version 2013 (www.SRIM.com).
18. Drouin D, Couture AR, Joly D, Tastet X, Aimez V, Gauvin R. CASINO V2.42 a fast and
easy-to-use modeling tool for scanning electron microscopy and microanalysis users.
Scanning. 2007;29(3):92101.
19. Rutherford E, Geiger H, Marsden E. The scattering of and particles by matter and the
structure of the atom. Phil Mag. 1911;Series 6, 21:66988.
20. Sijbrandij S, Notte J, Sanford C, Hill R. Analysis of subsurface beam spread and its impact on
the image resolution of the helium ion microscope. J Vac Sci Technol B. 2010;28(6):C6F69.
21. Goldstein J, Newbury DE, Joy DC, Lyman C, Echlin PE, Lifshin E, Sawyer L, Michael J.
Scanning electron microscopy and X-ray microanalysis. 3rd ed. New York: Kluwer/Plenum
Publishers; 2003.
22. Petrov YV, Vyvenko OF, Bondarenko AS. Scanning helium ion microscope: distribution of
secondary electrons and ion channeling. J Surf Invest X-Ray Synchrotron Neutron Tech.
2010;4(5):7925.
23. Kostinski S, Yao N. Rutherford backscattering oscillation in scanning helium ion microscopy.
J Appl Phys. 2011;109:064311.
24. Rabalais JW. Principles and applications of ion scattering spectroscopy. Hoboken: Wiley
Interscience; 2003.
25. Sijbrandij S, Notte J, Scipioni L, Huynh C, Sanford C. Analysis and metrology with a focused
ion beam. J Vac Sci Technol B. 2010;28(1):737.
26. MacRae CM, Wilson NC. Luminescence database I minerals and materials. Microsc
Microanal. 2008;14(2):184204.
27. Boden SA, Franklin TMW, Scipioni L, Bagnall DM, Rutt HN. Ionoluminescence in the helium
ion microscope. Microsc Microanal. 2012;18(06):125362.
28. Benninghoven A, Rdenauer FG, Werner HW. Secondary ion mass spectrometry: basic con-
cepts, instrumental aspects, applications, and trends. New York: Wiley; 1987.
194 J. Notte and B. Goetze

29. Wirtz T, Vanhove N, Pillatsch L, Dowsett D, Sijbrandij S, Notte J. Towards secondary ion mass
spectrometry on the helium ion microscope: an experimental and simulation based study with
He+ and Ne+ bombardment. Appl Phys Lett. 2012;101:041601.
30. Pawley J, Schatten H. Biological low-voltage scanning electron microscopy. New York:
Springer; 2008.
31. Crang FE, Klomparens KL. Artifacts in biological electron microscopy. New York: Plenum
Press; 1988.
32. Vanden Berg-Foels WS, Scipioni L, Huynh C, Wen X. Helium ion microscopy for high-reso-
lution visualization of the articular cartilage collagen network. J Microsc.
2012;246(2):16876.
33. Rice WL, Van Hoek AN, Punescu TG, Huynh C, Goetze B, Singh B, Scipioni L, Stern LA,
Brown D. High resolution helium ion scanning microscopy of the rat kidney. PLoS One.
2013;8(3):e57501.
Chapter 8
Sum Frequency Generation Vibrational
Spectroscopy: A Sensitive Technique
for theStudy of Biological Molecules
at Interfaces

Andrew P. Boughton and Zhan Chen

Abstract The behavior of biological molecules in interfacial environments is critical

to understanding a broad range of phenomena, from biocompatibility to the functions
of membrane-associated peptides and proteins. Sum frequency generation (SFG)
vibrational spectroscopy is a nonlinear optical vibrational spectroscopic technique
with an excellent sensitivity to interfacial molecules and molecular ordering and is
well suited to probing biomolecules in a native interfacial environment. Using this
technique, one can obtain unique information on the biomolecular orientation and
conformation at interfaces. SFG also provides additional measurements that are
complementary to other vibrational studies; more complicated protein structures
and orientation distributions may be studied in greater detail when SFG is combined
with other vibrational techniques.

8.1 Introduction

Although many biological phenomena have been studied in solution, interfaces are
more difficult to probe. The structures of biomolecules at interfaces control many
important phenomenaexamples include blood coagulation on implant surfaces,
membrane protein functions, marine biofouling, biosensing, and antimicrobial
potency and selectivity [16]. Protein adsorption is the earliest response of the body
to biomaterial implantation; depending on the properties of the material in question,
this can lead to quite dramatic changes in protein structure relative to the X-ray crystal
structures that have proven so valuable in biology. These structural changes mediate
further body reactions and may cause unfavorable responses, such as blood clotting

.P. o h o . he *)
e ar me o hemi r i er i o Mi hi a or h i er i e e
r or MI S
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 195

I . S ri er I er a io al P li hi S i erla
196 A.P. Boughton and Z. Chen

and unnecessary immune response [ ]. Biomolecules such as adhesive proteins from

marine organisms can also interact with coating surfaces, leading to biofouling [ ].
Proteins exist at interfaces quite naturally as well; for example, the lipid mem-
branes that separate cells from their surroundings contain a wide range of proteins.
These molecules are vital for the transport of nutrients, the survival of the cell, and
the ability to sense and adapt to changes in the cells surrounding environment. Yet
high-quality crystals of many membrane proteins have proven elusive for crystal-
lographers, and the constraints of crystallization may alter native structures in the
cases where X-ray diffraction data do exist. Further, such static pictures may fail to
reveal the changes in protein structure that occur during biological processes or
from changes in the protein environment (such as temperature, pH, or varying
charge and hydrophobicity of the lipid environment). Therefore, it is necessary to
investigate the structures of membrane-associated proteins in situ. A direct probe of
peptides in membranes also has applications to the rational design of antimicrobial
e i e MP ) 1, ], for which orientation measurements can reveal the nature
o i era io i h he mem ra e ho h o e ri i al o MP io .
Surface-sensitive in situ probes of biomolecules at interfaces are needed to
address these and other questions, yet such techniques are often limiteddue to
either a requirement for high-vacuum, interfering signals from the surrounding
environment or a need for chemically added labels that may alter the native behavior
of the biomolecule in question.
To this end, sum frequency generation vibrational spectroscopy (commonly
abbreviated as SFG) has recently been applied to a range of biological systems. The
focus of this chapter is on the biological applications of SFG for a broad audience
in science and engineering. Here, instead of an inclusive review with an extensive
discussion of SFG methodology and applications, we will summarize the method,
compare advantages and disadvantages, and demonstrate examples of how this
emerging technique can expand our knowledge of the interfacial behavior of impor-
tant biomolecules. For further background, and details of many nonbiological appli-
cations of SFG, numerous review articles are available to the interested reader
[ ]. Examples of nonbiological systems studied using SFG include surface
structures of liquids [15 ], interfacial structures of surfactants [19, ], structures
of molecules on electrodes [ ], molecular adsorption and reaction on catalyst
surfaces [ ], chirality [ ], and polymer surfaces and interfaces [9, , ].

8.2 S
 um Frequency Generation (SFG): Theory,
Experiment, and Data Analysis

8.2.1 SFG: A Nonlinear Optical Process

SFG is a second-order nonlinear optical process. One practical consequence of this

fact is that output signals will be observed only in media that lack inversion sym-
metry (under the electric dipole approximation [ ]). Bulk materials (such as air,
 S m re e Ge era io i ra io al S e ro o

Fig. 8.1 SFG process

liquids, amorphous solids, and many crystalline solids) possess inversion symmetry.
At many surfaces and interfaces, inversion symmetry is broken; thus, SFG is highly
sensitive to surface-adsorbed molecules and interfacial molecular ordering.
I he i al S G e erime a e re e i i le eam i o erla e i
space and time with a frequency-tunable infrared laser, producing an output beam
whose frequency is the sum of the two input beams (Fig. . ). This process is greatly
enhanced when one of the beams (usually the infrared) is in resonance with the
(vibrational) transitions of the molecule, resulting in a surface-sensitive vibrational
spectroscopic technique; by using polarized beams, we can also study the molecular
orientation. Data are presented as the output (sum frequency) signal intensity plot-
ted against the frequency of the tunable infrared beam to take advantage of this fact.
here ore S G e ra a ear imilar o o rier ra orm i rare I )a
ama e ra.
However, there are important differences. Vibrational modes are observed in
I or ama e ra he ei her he i ole or olari a ili ra i io mome
change, respectively. The response of the molecule in SFG depends on a property
known as the hyperpolarizability (), which is the product o he I i ole a
a * m n
ama olari a ili ra i io mome h ha blmn , q lm . I ei her he
Qq Qq
change in polarizability () or the change in dipole moment () of a vibrational
mode is equal to zero, then no SFG signal will be observed: The vibrational peaks
observed in SFG are only those that are present in both infrared and ama e ro -
copy. This can be advantageous in the study of adsorbed proteins, as the elimination
of interfering peaks simplifies the resulting spectrum. Further, the selection rule of
SFG is such that the interfacial layer is probed selectively, and with good sensitivity.
Both experiments and theoretical simulations indicate that SFG is submonolayer
surface-sensitive. By contrast, the surface sensitivity in attenuated total reflectance
o rier ra orm I e ro o I a ommo l e li ear e ro -
copy) is determined by the penetration depth of the evanescent wave (on the order of
h re o a ome er or e e mi ro ). I ra i e I ro e ma la -
ers of molecules, beyond those at the surface or interface. As a result, large water-
bending signals must be subtracted when studying backbone signals from proteins at
the surfacewater interface. This background subtraction is not necessary for SFG.
 A.P. Boughton and Z. Chen

l ho h r a e e ha e ama a eri S S) ha ee i el a lie

to study the surface and interfacial structures of various molecules, it also lacks the
intrinsic surface specificity of SFG and requires the use of metal substrates for the
signal enhancement to occurthis is not always ideal for biological samples.
Another distinctive feature of SFG is that it is a coherent process, meaning that
different vibrational modes can interfere with each other. Although this may compli-
cate fitting in some cases, it can also provide an opportunity to measure the absolute
orientation and to test the fitting results empirically. SFG spectra are fit according to
the following equation:
2
ASFG , q
= c nr +
2
I SFG c eff ,
q w 2 w q + iG q

where A represents the strength of the qth vibrational mode, is the damping coef-
ficient (peak width), is the infrared frequency, q is the vibrational frequency of
the qth vibrational mode, and eff is the effective second-order nonlinear susceptibil-
ity (see later discussion). One practical consequence of this equation is that vibra-
tional modes of similar energy (peaks close in frequency) can interfere with each
other constructively or destructively, depending on their relative phases. This inter-
ference allows SFG to measure the absolute orientation, such as whether a group
points up or down; such information is difficult to obtain from other commonly used
vibrational techniques.
o a e m h or ha ee o e i S G o he re hi mo e i
ario hemi al ro h a me h l me h le e a he l ro ). More
re e l he hi hl i orma i e ami e I a ha ee ie a ell e ami i
signals from backbone carbonyl groups in the protein [ . he ami e I ea e er
frequencies and peak widths are sensitive to the secondary structure and conforma-
io al ha e o he ro ei a i I h S G ie a e ho h o a a
valuable extension of earlier vibrational spectroscopies, with a range of new advan-
tages (discussed herein) that are well suited to complex biological systems. By
adopting similar methods of sample preparation and combining multiple techniques,
it is possible to study complex biomolecules with superb interfacial sensitivity.

8.2.2 SFG Data Analysis

8.2.2.1 SFG Orientation Analysis for CH Groups

By taking advantage of the polarized nature of the output sum, input visible, and
input infrared radiation in SFG experiments, we may extract important information
o mole lar orie a io . I a i io o ha i he e elle ra e e i i o -
lined earlier, SFG uses three beams, providing more combinations of polarizations
a o e iall more mea reme ha I o e eam) or ama o eam )
this allows for more complicated orientations and orientation distributions to be
characterized.
 S m re e Ge era io i ra io al S e ro o 199

The fitted signal intensity in SFG is related to the surface susceptibility tensor ijk
in the lab-fixed coordinate system, as shown above. Different combinations of
polarized beams allow the measurement of different components of this tensor [ ]:

, ssp = Lyy (w ) Lyy (w1 ) Lzz (w 2 )sin b 2 c yyz ,

(2)
c eff

, sps = Lyy (w ) Lzz (w1 ) Lyy (w 2 )sin b1 c yzy ,

(2)
c eff

, pss = Lzz (w ) Lyy (w1 ) Lyy (w 2 )sin bc zyy ,

(2)
c eff

, ppp = Lxx (w ) Lxx (w1 ) Lzz (w 2 )cos b cos b1 sin b 2 c xxz

(2)
c eff
Lxx (w ) Lzz (w1 ) Lxx (w 2 )cos b sin b1 cos b 2 c xzx
+ Lzz (w ) Lxx (w1 ) Lxx (w 2 )sin b cos b1 cos b 2 c zxx
+ Lzz (w ) Lzz (w1 ) Lzz (w 2 )sin b siin b1 sin b 2 c zzz ,

where Lij denotes Fresnel factors, and , 1, and are angles between the surface
ormal a he S G i al i i i le a i I eam re e i el . he
surface susceptibility tensor in the lab-fixed coordinate system is proportional to
theresponse of the molecule as described by the molecular hyperpolarizability
tensor, lmn, related by a coordinate transformation:
     
(2)
c ijk ,q = N i l( )( )( )
j m k n blmn,q .
l ,m,n

Therefore, SFG measurements can be related to the orientation angle of the molecule
relative to the lab frame, where is typically defined as the angle between the princi-
pal axis of the molecule or function group and the z-axis perpendicular to the plane
of the interface. For example, for the symmetric stretching mode of a methyl group,

1
c xxz , s = c yyz , s = N s b ccc cosq (1 + r ) cos3 q (1 r ) ,
2
1
c xzx , s = c yzy , s = c zxx , s = c zyy , s = N s b ccc cosq cos3 q (1 r ) ,
2
c zzz , s = N s b ccc r cos q + cos3 q (1 r ) .

Here, aac=rccc. The number density, Ns, is a constant property of the surface and
does not change with the polarization combination of the input laser beams.
Therefore, important information on molecular orientation may be extracted from
experimental measurements as a ratio of the signal strength measured in two polar-
izations. This approach does not require knowledge of the surface coverage, which
cancels when a ratio is taken; this somewhat simplifies the experimental procedure.
However, it requires the assumption that every methyl group has the same orienta-
tion angle . I me h l ro a o i ere orie a io he i he a o e e a io
cos and cos will be replaced by their averages: <cos > and <cos >. To
 A.P. Boughton and Z. Chen

characterize these average values for a distribution, it is necessary to know the

surface coverage (Ns) [ ]. A number of articles discuss the mathematical relation-
ships in greater depth [ ].
Ma S G ie ha e ee er orme o he orie a io o ro h
as methyl, methylene, and phenyl groups), and the details will not be repeated here;
instead, our focus will be on the orientation of protein secondary structures as deter-
mi e i ami e I i al . o e er i i or h o i ha me h l ) signals
are also of great biological utility, and their orientation can be used to study struc-
tural changes that affect protein side chains and model cell membranes.

8.2.2.2 SFG Amide I Protein Signals

Si al i he ami e I re io mai l ome rom he ro o he a o e o

the protein and are sensitive to secondary structure. Two common secondary struc-
tures that form the building blocks of larger proteins are -helices [ ] and -sheets
[ ]; both have been studied using SFG. Signals from these structures give peaks at
different frequencies, which in SFG can be separated and correlated to the overall
orientation of -helices or -sheets in interfacial biomolecules.

8.2.2.3 Determining the Orientation of an -Helix Using SFG

I he orie a io o ea h e o ar r re o a ro ei a a i er a e a e
deduced, the overall orientation and conformation of the protein can be inferred (for
further discussion, see Sect. . . . . ). As was discussed above for the simple case
of a methyl group, the orientation of an -helix can be deduced from polarized SFG
experiments. The different susceptibility tensor components abc measured in each
polarization are related to various hyperpolarizability elements ijk as a function of
the molecular orientation. For -helices, the observed signal intensity varies with the
tilt angle relative to the surface normal according to the following equations [ ].
For the A mode:

1
c A , xxz = c A , yyz = N s (1 + r ) < cosq > (1 r ) < cos3 q > b ccc ,
2
1
c A , xzx = c A , yzy = c A , zxx = c A , zyy =
2
( )
N s (1 r ) < cosq > < cos3 q > b ccc , . )

c A , zzz = N s r < cosq > + (1 r ) < cos3 q > b ccc ,

b aac
where r = .
b ccc
For the E1 mode:
 S m re e Ge era io i ra io al S e ro o

( )
c E , xxz = c E , yyz = N s < cosq > < cos3 q > b aca ,

( 3
)
c E, xzx = c E, yzy = c E, zxx = c E, zyy = N s < cos q > b aca ,

. )

( )
c E , zzz = 2 N s < cosq > < cos3 q > b aca .

Because of the limited resolution of many SFG spectrometers (~5cm or more),
it is necessary to add the response of the two vibrational modes of the helix together
before taking a ratio between polarized measurements. This assumes that the two
peaks are too close in frequency to be separated in the fitting, so that the total inten-
sity of the measured peak will be a combination of both of the closely spaced modes.
erime al mea reme rom i rare a ama ha e ee e o e ermi e
the values of required to determine the orientation [ ]. Using these parame-
ters, SFG has recently been applied to study the multiple orientations of the helical
peptide melittin in a lipid bilayer, as well as the orientation distribution of helical
fibrinogen coiled-coils at interfaces [ , ]. The total hyperpolarizability of a
multi-helix protein has been calculated and has also been applied to the study of
more complicated proteins containing multiple distinct helices, such as the sub-
unit of a heterotrimeric G-protein in a model membrane [ ].

8.2.2.4 Studying the Orientation of a -Sheet with SFG

While helices represent one of the most common secondary structural motifs,
-sheets represent another, and some preliminary work has been done on their ori-
entation analysis. This is somewhat more complicated by the fact that upon averag-
ing to calculate the surface susceptibility, one can no longer assume that the molecule
is axially symmetric. The result is that the orientation of a -sheet is characterized
by two angles: a tilt angle, , and a twist angle, [ ].
For the B1 mode:

(
c xxz = c yyz = c xzx = c yzy = c zxx = c zyy = < cosq sinf cosf > < cos3 q sinf cosf > b abc , )

(
c zzz = 2 < cosq sinf cosf > < cos3 q sinf cosf > b abc , )
1
c zxy = c zyx = c yzx = c xzy =
2
(
< sin 2 q cos2 f > < sin 2 q sin 2 f > b abc . )

1
(
For the B mode: c zxy = c zyx = c yzx = c xzy = < cos2 q > < sin 2 q cos2 f > b acb .
2
)
1
(
For the B mode:c zxy = c zyx = c yzx = c xzy = < cos2 q > < sin 2 q sin 2 f > b bca .
2
)
The achiral susceptibility tensor elements for the B and B modes are identical
to those for the B1 mode, except that abc is replaced by acb and abc, respectively.
 A.P. Boughton and Z. Chen

Fig. 8.2 Layout of a typical SFG instrument

o e ha li e re io e am le he orie a io o a -sheet is described by

a dditional parameters besides <cos > and <cos >; as we will discuss later, the
availability of additional measurements can be very valuable when studying larger
proteins.
Due to the symmetry of -sheets, a new set of chiral tensor elements is available
for measurement in addition to the elements measured normally. (For the chiral ele-
ments, all three indices i, j, k in the element ijk are i ere .) I a he -sheet
structure can potentially yield strong chiral signals [ , ]. These additional mea-
surements may make it possible to analyze the orientation of -sheets at interfaces;
the relevant molecular hyperpolarizability components required for quantitative
orientation analysis of -sheets are currently being investigated.

8.2.3 SFG Experimental Procedures

The typical SFG experiment employs a fixed-frequency visible beam and an infra-
re la er ei her a le re e or roa a ). e re e a i e S G i r me
ha e ee e ri e i e . 51 ], and in recent years commercial instruments
have become increasingly available from several manufacturers. The input laser
beams are spatially and temporally overlapped to produce the third output fre-
e hi h i olle e a ho om l i lier e or o her e e io
heme i e he eam i e are i e mall m), sample requirements are
reasonable. An example SFG system is shown in Fig. . . This system has four
om o e )a i o e o G la er ) a harmo i i ih o P
r al ) a o i al arame ri e era io PG) o i al arame ri am li a -
tion (OPA) and difference frequency generation (DFG) system based on LBO and
 S m re e Ge era io i ra io al S e ro o

Fig. 8.3 The SFG near-

total-reflection experimental
eome r e ro e i h
ermi io rom e . ].
meri a hemi al
Society)

AgGaS r al a ) a e e io em. he i i le eam m) i e er-

ae re e o li he ame al o le o le i h
rom he G la er. he a le I eam i e era e rom he PG P a
G em a a e e rom o m ; by replacing the AgGaS
r al i h a GaSe r al hi ra e ma e e e e o o m . I hi
example, the output signal is collected by a photomultiplier and processed with a
a e i e ra or. o e er e e or are e omi i rea i l a aila le.
mo i ori he o er o he i i le a I eam i ho o io e e a
normalize SFG spectra by the power of the input laser beams.
I or er o olle ro ei i al i i e e ial o o i er he e erime al
geometry. As shown in Sect. . . . , the output signal is affected by the Fresnel fac-
tors (Lij), which depend on the incident angles of the beams and the indices of
re ra io o he me ia hro h hi h he eam a . Ma e i i S G e eri -
ments employ a straightforward reflection geometry, which has the advantage of
being relatively easy to align. Yet if we change it to a near total internal reflection
right-angle prism geometry (Fig. . ) instead, the angle-dependent Fresnel factors
are increased substantially, making it easier to observe weak signals experimentally
[ , 55]. This provides advantages for fitting and quantitative analysis, as well as
improved detection limits.
One important advantage of SFG in the study of biomolecules is that unlike
vacuum techniques, SFG is capable of measuring protein orientation and conforma-
tion in a biologically relevant, aqueous environment. The surface sensitivity of SFG
is such that as little as a few micrograms of the protein of interest are enough for
reasonable signals to be observed, making it feasible to study protein behavior even
at low concentrations. This is also beneficial when employing expensive isotope-
labeled materials, large quantities of which are often difficult to obtain.
Lastly, dramatic signal and selectivity enhancements may be possible by employ-
i he emer i e h i e o o l re o a S G S G) , 5659], which
can enhance the signal intensities by exciting both vibrational and electronic transi-
io lea i o i al e ha eme o e eral or er o ma i e. ormall he
S G i al i e i i e ermi e he ro o he I a or io a
 A.P. Boughton and Z. Chen

he ormal ama i al i e i . he S G i al i e i i he ro o
he I a or io a he resonance ama i al i e i , ], and therefore
i al o ai e i he S G i al a e m h ro er ha ho e rom
normal SFG. This promises superior detection limits, among other uses. However,
there have as yet been few applications of this technique to biomolecules, and so
this technique will not be discussed further in this chapter.

8.3 Applications of SFG to Biological Molecules at Interfaces

8.3.1 S
 FG Studies of Proteins: The CH and NH Spectral
Region

Partially due to limitations in the tunable frequency range of older laser instruments,
ma earl S G ie o ro ei o e o re hi i al me h l
groups, phenyl rings, and other functional groups of interest in protein amino acid
side chains. Although the sheer number of similar amino acids would hamper any
serious attempt to study the orientation and conformation of the entire protein at the
i er a e a e o i al he re ireme ha a me i m la i er io m-
metry allows signals to be interpreted in terms of side-chain ordering at the inter-
a e. arl ie ho e ha ro ei i al o l e e e e rea il o a
range of biomedically important polymers [ ], and these signals were interpreted
as demonstrating the segregation of hydrophobic and hydrophilic residues at the
polymerwater boundary. As a later test, Somorjai and coworkers used a simple
peptide designed to form a facially amphiphilic helix at interfaces and found that
ro i al o l i ee e o ai e 61, ]. These results confirmed that
the hydrophobic residues are strongly ordered at a hydrophobichydrophilic inter-
face (e.g., polymersolution). Such ordering of hydrophobic groups has also been
seen previously for a variety of amino acids at an oilwater interface [ ]. By also
i he re h a e a ollea e emo ra e ha e i
amino acids remain buried after the conversion of fibrinogen to fibrin [ ]. Water
molecules can also order in the vicinity of a protein, providing additional clues
about protein adsorption and the burial of hydrophobic regions [65, 66].
Side-chain ordering at an interface can be influenced by changing the contacting
media of the protein [61, , ]. Different spectra were collected from bovine
serum albumin (BSA) at the polymerwater interface as opposed to the polymerair
i er a e a a re l o a i ere or eri o h ro ho i ro . e la i he
water with a hydrophobic solvent (such as benzene) also led to spectral changes.
Stronger ordering was observed at hydrophobichydrophilic interfaces, whereas
only weak signals were observed from hydrophilichydrophilic interfaces (such as
SiO a er). I i im or a o o e ha S G i er e i i e o mole lar orie a-
tion; as a result, weaker signals can sometimes be observed even in the case where
more adsorbed protein is present (if those molecules are less well ordered) [69].
 S m re e Ge era io i ra io al S e ro o

ha e i he i er a ial e iro me a ma i e a o her ha e i ro ei

behavior as well. A difference in pH can affect side-chain ordering at an interface,
as was observed for proteins such as lysozyme [ ]. This may dramatically change
he omi a ro ei r a e i era io i ome a e . ormal al e
groups in human fibrinogen were found to order; at acidic pH values, this interac-
tion shifted to more favorable hydrogen bonding [ ]. Thus, SFG can provide a
molecular-level explanation for the pH-dependent difference in protein affinity for
the surface. Such changes may also shift the balance in the process of competitive
adsorption, as the nature of the favorable proteinsurface interactions shifts.
A change in pH may also affect the protonation state of amino acids on the pro-
tein, which in turn will alter the net charge of the adsorbed protein layer and result-
ing interactions with other molecules. Urea, a common protein denaturant, was
found to orient differently toward a protein layer at high and low pH values; since
the actual effect of urea denaturation is not sensitive to charge, a direct interaction
between the protein and the denaturant was ruled out as the mechanism [ ]. The
interfacial and in situ sensitivity of SFG makes it possible to directly study molecu-
lar interactions at the interface as a function of changes in the surrounding condi-
io ha . S G ro ei i al ha e al o ee e o mole lar
recognition events such as protein binding of a ligand, of great importance to the
development of biosensors [ , ].
As the protein concentration increases, proteinprotein interactions become
i rea i l im or a . Pro ei i al ere o o ha e a a ro ei ol -
tionair interface as a function of concentration, indicating that interactions between
nearby proteins can be very important in determining the ordering of hydrophobic
side chains (as well as the entire structure) [69, ]. Various surfactants were like-
wise shown to affect protein aggregation [ ].
I e i ai ro ei r re a i er a e S G i al ro i e ire
in situ evidence that interfacial protein side-chain ordering can change when the
environment is altered. For simpler biomolecules, such as amino acids and peptides,
with well-defined structures, more detailed structural information may be inferred
from such studies; this is valuable for the study of structural changes related to bio-
compatibility and biofouling. However, the very large number of similar hydropho-
bic groups in many larger proteins makes it difficult to describe the structure in
e ail i i al alo e. I o o e la eli o e i ami o a i i o e a o
gain additional information from interfacial proteins.

8.3.2 Isotope Labeling

I o o e la eli a ro i e a a o e ermi e he orie a io o o e or a e i e

chain groups, by shifting the frequency of an individually labeled vibration to a new,
easily resolved position away from interfering or overlapping signals. The value of
hi a roa h ha ee emo ra e i li ear e ro o ha I . mero
labeling studies have demonstrated that isotope labeling can lead to a more accurate
 A.P. Boughton and Z. Chen

determination of molecular orientation [ ]. Labeling studies are particularly

useful when dealing with proteins, as it allows the signal from one specific group or
type of amino acid of interest to be isolated from other amino acids. For large and
complicated proteins, this provides an additional set of measurements that allow the
characterization of biomolecules in which multiple similar groups cannot be distin-
guished by spectral fitting alone. Thus, the use of isotope labeling, when combined
with the increased polarization-dependent measurements in SFG, may prove a very
valuable tool in the study of large proteins [ ].
I ha ee emo ra e ha S G i e i i e e o h o e e a i le la ele
ami o a i i a ro ei a a i er a e. I ha al o ee re or e ha S G a ele -
tively probe isotope-labeled segments in a protein, such as when deuteration of a
certain amino acid was introduced [ . I i ere e me o a ro ei a e
isotope-labeled in succession, SFG can be used to study the structure of a protein in
detail.
Another advantage of isotope labeling lies in the ability to separate signals con-
tributed by different molecules (e.g., labeled and unlabeled). SFG studies using
partially deuterated proteins and normal hydrogenated proteins have demonstrated
that protein adsorption to a polymer surface is a reversible process [ ] and that a
protein will induce changes in polymer structure as it adsorbs [ ]. The ability to
separate signals from different molecules at the surface can be quite useful in moni-
toring competitive adsorption, namely, the Vroman effect, in which one protein is
displaced by another with lower mobility but greater affinity for the surface [ ].
I o o e la ele ro ei ha e ee e o ho ha h ro ar o o ami a io o
the surface may gradually be replaced by a layer of adsorbed proteins over a period
of time [ ].

8.3.3 Studying Amide I Protein Signals with SFG

8.3.3.1 Advantages of the Amide I Band

he ami e I a o ro ei i a ra i e or i ra io al ie o ro ei . hi
band can be loosely described as a stretching mode of backbone carbonyls [ ], and
the exact peak center frequency depends on the secondary structure and conforma-
tion of the protein. These frequencies are roughly the same as those observed via
I S G ami e I e ra o o er he i er ere e rom ro a er
e i i al ha all o re he ami e I re io i I . e er
and more broadly tunable SFG instruments become increasingly available, it is
expected that this informative spectral region will be increasingly studied.
Further, the unique peak center frequencies of various secondary structural ele-
ments in a protein can be combined with the orientation information obtainable by
SFG to create a more complete picture of the interfacial orientation. Although
vibrational spectroscopy cannot directly inform us which segments of the protein
have adopted a given helical or -sheet conformation, the increasing availability of
 S m re e Ge era io i ra io al S e ro o

high-quality protein structures (via the Protein Databank, PDB) allows important
conclusions to be drawn about the structural changes that occur with slight changes
in the protein environment in situ. For example, if regions of the protein undergo
bending or deformation, changes in the spectrum will be observed. As discussed
a o e i al are o e i l o i er re i erm o a e aile mole lar
orie a io i o ra he ami e I a re e a i re o he a o e mole -
lar orientation that is well suited to the task.

8.3.3.2 Detection of SFG Amide I Signals

A method to probe the overall molecular orientation selectively at interfaces would

be of great utility. For example, many membrane-active proteins or peptides (such
a a imi ro ial e i e MP ) are i l o i heir a i e e iro me
yet knowing their orientation in membranes is critical for understanding the mode
of action and designing more effective antimicrobial agents [1, , ]. Because
these peptides are generally short and adopt characteristic secondary structures
(such as -helices and -sheets) that form the building blocks of larger proteins, the
orientation determination of small antimicrobial peptides also offers a proof-of-
concept and testing ground for larger proteins.
mi e I i al ha e ee e ami e e eral re ear h ro o a e. he r
S G ami e I i al ere olle e rom he ol mer ro ei i er a e i he
near-total-reflection geometry; a number of proteins were studied, demonstrating
that SFG is sensitive enough to distinguish among proteins with different secondary
structures [ ]. By using antimicrobial peptides such as heli al MSI a he
hee e i e a h le i re ear her ha e ho ha S G ami e I i al ori i-
nating from proteins can be detected from both -helices and -sheets distinguish-
ably [ , ]. Using the thin-film optical model for the interpretation of spectra,
Wang etal. proved that the signals obtained are from proteins adhered to the sur-
face, rather than only to part of a layer [ ].

8.3.3.3 Q
 ualitative Studies: Amide I Spectra Demonstrating
Orientation and Ordering

e i e a o e or he a e ral re io ro er i al are
obtained for greater molecular ordering or certain orientations though here the
signals obtained are a guide to the overall ordering and orientation of the protein
backbone. Thus, strong signals were observed for long and well-aligned poly--
benzyl-l-glutamate (PBLG) helices [ ], yet only very weak signals were obtained
from albumin despite a large number of -helical segments in the latter [ ]. As
with other chemical groups, this molecular ordering or orientation can be affected
by the composition of the surface in question, as was demonstrated for time-
e e e S G ami e I i al ha e o ri o e o a arie o ol mer r a e
[ ]. The surface charge can also alter the ordering. For example, the blood protein
 A.P. Boughton and Z. Chen

a or II II) a a oa i a e o e a i el har e r a e lea i o loo

oa la io . I a h o he i e ha hi a i a io i e o i rea e or eri a
a more favorable orientation when adsorbed on negatively charged surfaces, but this
ha ee i l o ire l e . i S G he e al. ha e ho ha e a-
tively charged) sulfonated polystyrene surfaces induce a net orientation and greater
or eri or a or e II om are o eha ior o e ral ol re e r a e
[91]. This view of proteins at surfaces provides useful structural insights for under-
standing biocompatibility.

8.3.3.4 Quantitative Orientation Studies from SFG Amide I Signals

he al e o M l i le Mea reme

S G ro i e a m er o a a a e or he o iomole le a i er a e I
is an inherently surface-sensitive technique with many of the advantages and appli-
a io o I e i i al o a a le o ro i i more mea reme o hara er-
ize the orientation of biological systems. Even a single-helix peptide in a well-defined
environment cannot always be described by assuming only a single orientation; for
larger proteins with multiple similar structural elements, obtaining more measure-
ments from additional vibrational spectroscopies is critical [ ].
Different vibrational spectroscopic techniques measure different structural
parameters. For example, if the tilt angle of a structural element (e.g., an -helix) is
er he r a e ormal I a e ermi e he a le mea ri o >.
or I hi i he o l arame er o ai e . here ore a orie a io a al-
ysis must assume that all chemically identical groups at the interface adopt the same
orientation (a i ri io ). I o ra a i e a o e S G a mea re
both <cos > and <cos >. Thus, SFG can be used to deduce two different struc-
tural parameters; in a Gaussian distribution, for example, this would correspond to
he a era e orie a io a i ri io i h. More om le i ri io re ire
a combined approach, and higher-order vibrational spectroscopic techniques can
measure more structural information.
or e am le o r a e mi i M) mea re o > and <cos > param-
eters; such additional measurements are useful for characterizing complex distribu-
tions involving multiple distinct orientations. Unlike SFG (a three-photon process),
M i a o r ho o ro e . er le el ia ram or S G a M are
shown in Fig. . or om ari o . e ial a e o M i ohere a i S o e
ama a eri S i . . ), in which two of the input beams are the same
re e . S e ro o a ima i ha e ee i el a lie o io-
logical samples [ ].
When the protein crystal structure is known, we can deduce the orientation of the
entire molecule at the interface by calculating the combined hyperpolarizability for
all -helical secondary structures in the protein. However, this requires the assump-
tion that the structure at the interface is the same as the crystal structure. When the
crystal structure is not known, or when structural changes in situ are suspected,
 S m re e Ge era io i ra io al S e ro o

Fig. 8.4Energy-level
diagrams for (a) SFG,
(b) M c) S

Fig. 8.5 The model protein

has three -helices and
several hee . I he
orientations of such
secondary structures can be
deduced, the rough structure
or conformation of the
protein can be determined

measuring the unique orientations of individual secondary structure elements may

help to understand the overall structure (Fig. . ). As discussed in Sect. . . . . , it
i i l o hara eri e m l i le i i orie a io i I alo e
however, in combination with additional measurements from isotope labeling and
o her i ra io al e ro o ie ha S Ga M) more e aile orie a io
distributions can be measured as a guide to the overall orientation of helical s egments
of the protein.
Where the physical constraints of crystallization may not be expected to alter the
native structure, the possibility of conformational changes (such as differences in
the relative orientations of secondary structures upon binding) still exists. These
changes may be difficult to observe through static techniques. Thus, if one collects
enough measurements to characterize the orientations of multiple distinct helical
segments individually, the in situ and time-resolved nature of SFG makes it possible
to study protein flexibility in a native environment [ ].
 A.P. Boughton and Z. Chen

As SFG is extended to the study of larger and more complicated systems, tools
for dealing with complicated structures become increasingly important. One
approach is to calculate the hyperpolarizability of different domains and secondary
structures of a large biomolecule as the sum of individual monomer hyperpolariz-
abilities based on ab initio calculations (taking into account perturbations or cou-
li ) hi i ea i he a i or he reel a aila le o are Pre i hi h
uses values for a model amino acid [95 ]. As these calculations become more
advanced, questions such as signal change due to flexibility or binding should
become easier to address.
e a e o he e i i i o he ami e I a o e o ar r re a orie a-
tion, the experimental validation of empirical fitting procedures also becomes
important for accurate orientation determination. By using a range of intermediate
polarizations in which signals from two separate polarization combinations are
observed (a technique known as polarization mapping), researchers are able to vali-
date the empirically fit parameters and determine the relative phases of different
vibrational modes, tasks that would be difficult to do using only the conventionally
measured polarizations. Further, Wang etal. have shown that empirical fitting alone
can produce misleading results and that the use of techniques such as polarization
mapping can improve reliability [ ].
I he ollo i hree e io e hall re e ae ie emo ra i he
al e o S G orie a io mea reme i ami e I i al i a arie o i ere
applications.

Orientation of eli al Meli i e ermi e mi e I Si al

Mem ra e o ro ei a mem ra e a i e e i e are im or a i er a ial

molecules of great interest, and SFG is well suited to their study in model lipid
bilayers. (For more details on the model membrane, see Sect. . . .) SFG has
recently been applied to study the effect of a variety of small, antimicrobial peptides
in cell membranes [ ]. The mechanism of activity for these peptides is of wide inter-
est, and here the ability of SFG to determine complicated orientation distributions
e ome al a le. i ere mo el e i I ha ee e e ha he e e i e
can either form a carpet parallel to the bilayer plane, or insert into the bilayer and
form pores to disrupt cell membranes.
om i i I i h S G he e al. mea re he orie a io o he
helical peptide melittin in a bilayer at the threshold concentration for antimicrobial
activity; they found that two distinct orientations of the peptide were present simul-
taneously [ . i I o l o e arame er a e mea re amel he
i e i ra io o he a olari e I a or a e . h he e re l
alo e a iel o l he a era e orie a io . I e er meli i mole le a o e he
ame orie a io I o l re i a il a le o rela i e o he r a e
normal. However, the unique <cos > and <cos > parameters measured by SFG
were not consistent with either a -distribution or a single Gaussian distribution
(Fig. . ).
 S m re e Ge era io i ra io al S e ro o

Fig. 8.6|zzz yyz| ratio as a

function of the -helix
orientation angle, assuming a
-distribution or Gaussian
distribution function. The
shaded area represents the
actual experimental result
(Figures . , . and . are
reproduced with permission
rom e . .
meri a hemi al So ie )

Fig. 8.7(a) Orientation distribution derived based on a dual -distribution; (b) orientation distri-
bution derived using the maximum-entropy function

Therefore, the orientation distribution must be more complicated. Under the

assumption that two orientations of melittin existed simultaneously, combined SFG
a I re l emo ra e ha a ro ima el o he meli i mol-
e le la o a a a le a he remai er oo i he ila er a a
angle (Fig. . a). This more detailed picture resolves a conflict between previous
I mea reme hi h e e ha he e e i e ere ei her er e -
dicular or arallel o he mem ra e r a e. I a both orientations are present,
shown schematically in Fig. . . To verify this, an additional method of data analy-
sis was applied that did not require assuming the nature of the distribution to start.
For an unknown distribution, it has been shown that the maximum-entropy func-
tion provides a mathematical way to use multiple measurements to characterize the
orientation while introducing the least bias. As shown in Fig. . , the resulting
distribution was quite similar to what would be derived under the assumption of a
 A.P. Boughton and Z. Chen

Fig. 8.8 Schematic of two

melittin orientations in the
lipid bilayer

Fig. 8.9(a) Schematic of the fibrinogen native structure; (b), (c), (e), and (f): possible conforma-
tion of fibrinogen after adsorption to PS; (d) ssp SFG spectrum collected from fibrinogen mole-
le a he PS ol io i er a e a e i h ermi io rom e . . meri a
hemi al So ie )

dual delta distribution. However, the maximum-entropy function requires no

assumptions. This approach to data analysis will be particularly important when
using SFG to characterize larger proteins with multiple distinct helical segments.

Orientation of eli al oile oil i i ri o e e ermi e

mi e I Si al

Fibrinogen is a blood protein that has been studied via a range of different tech-
niques; it is well known that the surface conformation plays an important role in the
formation of clots. A combination of vibrational spectroscopies was used to study
the orientation distribution of two -helical coiled-coils of fibrinogen at the polysty-
rene (PS)fibrinogen solution interface, Wang etal. found that adsorbed fibrinogen
molecules adopt a bent structure at the interface [ ].
 S m re e Ge era io i ra io al S e ro o

Fig. 8.10Orientation
distributions of fibrinogen
coiled-coils deduced by
(a)two and (b) three
mea reme e ro e
i h ermi io rom e .
[ . meri a
hemi al So ie )

Due to an approximate inversion symmetry in the native structure of fibrinogen

(Fig. . ), little signal would be expected if this was the adsorbed conformation.
Experimentally, very strong SFG signals (Fig. . ) were detected from -helical
coiled-coils at the PSfibrinogen solution interface, indicating that adsorbed fibrin-
ogen molecules must adopt a bent structure instead (Fig. . e ).
I or er o e ami e hi e r re i more e ail he orie a io i ri io o
fibrinogen on PS was studied in two ways: first, by assuming a Gaussian distribution
using SFG measurements alone (Fig. . a), and then by using a combination of SFG
a I mea reme o o r he orie a io i ri io i he ma -
imum-entropy function (Fig. . ), as outlined above. The two distributions were
similar, showing that the Gaussian distribution is a good assumption; however, they
were not identical, indicating that additional measurements would allow for a more
a ra e i re o he real i ri io . e ar le he i ri io i i e roa
suggesting that adsorbed fibrinogen molecules can adopt a variety of orientations and
o orma io o i l e o he o i orm ol mer r a e. I o ra meli i
in bilayers adopted a narrow distribution due to the strong ordering of the lipid bilayer
environment, which makes specific lipidprotein interactions more likely [ ].
 A.P. Boughton and Z. Chen

Fig. 8.11 SFG spectra collected from G a o ia e i h i ere li i ila er P P P P

P P P P P PG mi re) P P P PG. he ea er i al i ea h a el i olle e i
a olari a io om i a io he ro er o e i . I e i ra io o a S G i -
nals are varied in different lipid bilayers, indicating that G adopts different orientations. The
orientation of G i he P P P P ila er i ho i heupper left corner (Figures . , .
and . re ro e i h ermi io rom e . . meri a hemi al So ie )

rie a io o G Pro ei e e rom S G mi e I

Signals of -Helical Domains

As discussed above, the orientation of larger proteins may be determined by calcu-

lating the combined hyperpolarizability of all the helical segments in the molecule.
I a o he m l iheli mem ra e i ali ro ei G he e al. e he
orientation of -helical portions of the protein and the static crystal structure as a guide
to the overall molecular orientation of the protein bound in a membrane;, they demon-
strated that the lipid composition can modulate that orientation (Fig. . ) [ ]. This
subunit of an important signaling protein is anchored to the bilayer via a geranyl-
geranyl group with a high membrane affinity. Significantly weaker signal intensities
were observed in the absence of this group (Fig. . ): SFG experiments suggest
that the geranylgeranyl group also anchors the protein in a specific orientation in
preparation for binding (Fig. . ).

Shee iel hiral Si al or i io al Mea reme

Because -sheets may be quasi-centrosymmetric, there was originally some ques-

io a o he her i al o l e e e e or hi r re i S G. I a
 S m re e Ge era io i ra io al S e ro o

Fig. 8.12 S G ea er) a ro er) ami e I e ra o i er a ial G adsorbed onto a

P PG P PG ila er a) 5l o m ml era l era la e G was injected into the subphase,
and stronger SFG signals indicative of an -helix structure were observed; (b) l o m ml
of the soluble form of G was brought into contact with the bilayer. Only relatively weak signals
indicative of the -sheet were observed

Fig. 8.13 Schematics of G a or e o o a P PG P PG ila er e e rom S G e ra

(a) geranylgeranylated G; (b) G with no geranylgeranyl group

Wang etal. have shown that -sheets possess D symmetry; using a series of
intermediate polarizations and careful fitting procedures, weak signals could be
detected from a layer of adsorbed fibrinogen [ ]. By using a small peptide
a h le i I) i h a rea er ra io o -sheet structure and high surface affinity,
Wangs group could observe signals even without interference enhancement.
The symmetry of -sheets allows for the detection of strong chiral signals that
provide additional measurements [ ]. This makes it possible to characterize both
 A.P. Boughton and Z. Chen

tilt and twist angles at a complex interface (see Sect. . . . , Studying the
Orientation of a hee i h S G ). I ha ee ho ha he hiral S G i al
observed originated with the layer of protein adsorbed on the surface, and not
molecules in solution. Additional quantitative studies are expected as the study of
-sheets using SFG continues.

8.3.4 P
 lanar-Supported Lipid Bilayers: A Model
Membrane Environment

8.3.4.1 Introduction

As outlined earlier, SFG provides powerful new tools for characterizing the structure
of complicated proteins in the membrane environment; however, this same tech-
nique also makes it possible to understand changes in the lipid membrane itself.
I ar i lar a i a a a e o i o o e la ele li i i i o i le o mo i or
the orientation and ordering of lipid groups from both leaflets even as the protein
interacts with the bilayer. This information is of particular importance to the study
of interactions between antimicrobial peptides and cell membranes, for example,
where changes in the bilayer structure have important consequences for understand-
ing the mechanism, activity, and selectivity of antimicrobial peptides [1, ].

8.3.4.2 The Model System: Planar-Supported Lipid Bilayers (PSLB)

he mo el mem ra e mo ommo l e i I a S G ie i he
planar-supported lipid bilayer (PSLB) [99]. With the LangmuirBlodgett and
LangmuirSchaefer techniques a single lipid bilayer is created on an optically trans-
are oli or . M h or ha ee o e i S G a a e iall e i e
cell for temperature control to validate the model system as a useful mimic of the
real li i e iro me i a o o ha e re or e ha he e ila er mo el
show realistic behavior (such as phase-transition temperature) [ ] when
deposited on a thin optical substrate, providing evidence against the possibility of
significant substrate effects. These experiments are very sensitive to molecular
packing and ordering, including small changes in methyl and methylene group
orientation for the lipid tails [ , ].

8.3.4.3 Asymmetric Bilayers Allow Study of Both Leaflets

Using an asymmetric bilayer (such as with one leaflet deuterated and the other
hydrogenated) enables time-dependent spectra to be examined for each individual
 S m re e Ge era io i ra io al S e ro o

leaflet, yielding a detailed view of processes such as flip-flop [ ] and membrane

disruption by antimicrobial peptides [ ].
Due to SFGs unique selection rules, changes in signal over time may be inter-
preted in light of bilayer disruption in each leaflet individually. By preparing a sym-
me ri ila er or e am le he e al. e ime e e e i al o he
kinetics of membrane disruption via small antimicrobial peptides and oligomers
[ , . I i i i all i o o e la ele lea e are e i ea o e h ro e-
nated and the other deuterated), it becomes possible to monitor changes in both leaf-
lets at the same time [ , ], including the orientation of methyl groups in the lipid
tail [ , , . h i he a e o a mall a imi ro ial oli omer he
etal. showed that the molecule acted as a molecular knife by disrupting primarily
the outer leaflet of the bilayer and inserting at a certain critical concentration [ ].
This concentration was similar to that observed as the minimum inhibitory concen-
ra io MI ) i e lea a e e erime i e i le ho e er S G e erime
provide a greater level of detail regarding the oligomerbilayer interactions.
I a i io o mem ra e ila er S G ha al o ee a lie o he o li i
or fatty acid mono- and multilayers [19, , ]. From these studies, informa-
tion on chain packing and conformation can be obtained.

8.3.5 Other Biomolecules: SFG Studies of Nucleic Acids

S G ha re e l ee a lie o he o or ere la er o oli ra e

an area of potential interest in the development of high-throughput sensors and
hi or e e e e i 115 . I a i io o he o orma io al e i i -
ity of the technique, both single-stranded [116] and duplex [ oli o le-
otides may be studied in terms of molecular chirality.

8.4 Future Directions, Summary, and Conclusions

I he e a e i e S G a r a lie o iolo i al mole le a i e arie o

increasingly complex systems have been studied (including proteins, peptides, and
lipids). This trend is expected to continue, as improved methods of data analysis
open the possibility of studying larger proteins adopting multiple distributions.
A number of advances in experimental technique are also possible. For example,
o l re o a S G S G) ha ee o l limi e a li a io o iolo i al -
tems as yet, but this technique promises very large improvements in sensitivity for
er il e am le . he heore i al o a io or o le re o a e I i)
SFG has been studied previously [ , . i S G allo or more ele i e
spectroscopic information and better assignment of the vibrational modes; in the
re S G a e e elo e i o a o er l e h i e o iolo i al
molecules at interfaces.
 A.P. Boughton and Z. Chen

SFG microscopes have also recently emerged [ . I i ial or ha

focused on the study of patterned self-assembled monolayers [ ], demon-
ra i a a ial re ol io o e er ha m. We believe that SFG imaging will
be valuable in biology for the study of questions such as the homogeneity of protein
adsorption and how biomolecules respond to patterned surfaces.
Lastly (as discussed in Sect. . . . . ), SFG can be combined with measurements
from other techniques in order to characterize the orientation of more complicated
molecules in more detail.
SFG has only recently been applied to the study of biological interfaces, yet the
ro h o he el ha ee e remel ra i . e a a e i i r me a io
design and reliability, as well as improved methods of data analysis, have opened
he a or o a i e ra e o iolo i all i ere i am le . I ma ae
these early studies are already giving way to more detailed explorations of the com-
plex behavior of biomolecules at surfaces. The following areas are expected to see
increased activity in the future:
Mem ra e ro ei are ri i all im or a e i l o r alli e. I ha
been demonstrated that SFG, combined with other vibrational spectroscopic
techniques, can detect these proteins in situ and provide sufficient measurements
to characterize important structural details. Simultaneously, lipid signals can be
studied to determine the effect of the protein on the membrane environment.
Likewise, interactions between small peptides and cell membranes can be stud-
ied, with important consequences for drug design.
Preliminary work has shown that SFG can also be used to probe proteinmole-
cule interactions at the surface: Ligand binding is critical to drug design and also
important for the development of biosensors.
I ere i he io al eha ior o a ha ro i re e ear .
The binding of aptamers and molecular beacons are examples of sensing appli-
cations in which interfacial sensitivity will prove valuable.
Lastly, work on biocompatibility and biofouling is expected to continue, as SFG
provides an excellent and label-free probe for molecular ordering. By under-
standing the interactions that drive protein adsorption, researchers may develop
better coatings and materials.

Acknowledgments e ha he eo a al e ear h ) he a io al
S ie e o a io a ) he e ma o a io a he
i er i o Mi hi a or heir or .

References

. a lo M. imi ro ial e i e o m l i ell lar or a i m . a re. .

. ia ee he a a e e M hai e al. ra r re o a ol a e
e e e ha el. a re. .
. or e ra h e i or . Pro ei a i er a e II ame al a a li a io S
m o i m erie . a hi o meri a hemi al So ie . . .
S m re e Ge era io i ra io al S e ro o

. e ra M iil S am oha e . i o li e h olo a re e a re e

o ar e ie a e iro me all rie l a i o li oa i . Pro r oa .
.
. Gra . he i era io o ro ei i h oli r a e . rr i S r iol.
.
. a a i hi Sa i ama Imam ra . he a or io o ro ei o oli ra e a
ommo er om li a e he ome o . io i ioe . .
. he he . S G ie o i era io e ee a imi ro ial e i e a or e
li i ila er . io him io h a. .
. a lar e M he e M oh o he . S r S i. .
. he She Somor ai G . Mole lar ie o ro ei o orma io a ol mer li i
i er a e i m re e e era io i ra io al e ro o . e Ph hem.
.
. ai . S m re e i ra io al e ro o o he oli li i i er a e. hem So
ara a ra . .
. i e hal . i i i er a e ro e e o harmo i a m re e e ro -
o . hem e . .
. i hmo G . S r re a o i o mole le a a eo ra e. e Ph
hem. .
. She ro er ho . S m re e i ra io al e ro o o a er i er a e
olar orie a io o a er mole le a i er a e . hem e . .
. Go ala ri h a S i lle o M Sh l M . i ra io al e ro o i ie o
a eo i er a e al a i ae a a o ro . hem e . .
. ha G o i e hal .Sr ral ha e ra i io o mall mole le a air
a er i er a e . hem So ara a ra . .
. S er e a She . o li ear o i al ie o he re li i a or i er a e
i ra io al e ra a olar or eri . Ph e e . .
. S er e re She . i ra io al e ro o o a er a he a or a er
i er a e. Ph e e . .
. ro MG al er S a mo i hmo G . i ra io al m re e e ro -
o o al a e a er i er a e e erime a heore i al im la io . Ph hem .
.
. ar er lle . om e i io e ee PP a S S a he air a eo i er a e.
a m ir. .
. o o Me mer M i hmo G . I e i a io o r a a o orma io a or er
a he li i li i i er a e o al i er al re e io m re e i ra io al e ro -
o . Ph hem. .
. al elli S Mailho G o P She Somor ai G . Po e ial e e e orie a io o
a e o i rile o la i m ) ele ro e r a e ie m re e e era io . Ph
hem . .
. aha Mi ra o ri S ra ai o S aram hoi IS e al. re er i l i hi
r a e. S ie e. .
. ho im al elli S Somor ai G . i ra io al e ro o o ar o mo o i e
acetonitrile, and phenylalanine adsorbed on liquid electrode interfaces by sum frequency
e era io . le roa al hem. .
. he Gra ia Somor ai G . S m re e e era io S G) r a e i ra io al
spectroscopy studies of buried interfaces: catalytic reaction intermediates on transition metal
crystal surfaces at high reactant pressures; polymer surface structures at the solidgas and
oli li i i er a e . l Ph . .
. re h er G. r a e ie e a roa h o am ie re re a al i rea io . a al
o a . .
. Somor ai G . e mo el a al la i m a o ar i le ) a e e h i e S Ga
S M) or ie o rea io i erme ia e a r a e re r ri a hi h re re ri
a al i rea io . l S r S i. .
 A.P. Boughton and Z. Chen

. el i M She . o l re o a I m re e i ra io al e ro o o
mole lar hirali . Ph e e . .
. el i M la o r a S She . eo a m re e e era io
i hiral li i . Ma er. .
. i She . ami o li mo el or m re e hiral re o e rom li i
om o e o mole le i h a hiral i e hai a a a hiral hromo hore. m hem
So . .
. ri ma S e he o alla e i h er . ol e mole lar orie a io al
i ri io o he ol re e r a e. Ph hem . .
. Ga am S S h a hi o ala ha o al SM e a eh MS. Mole lar
r re o ol re e a air ol mer a oli ol mer i er a e . Ph e e .
.
. She . Pri i le o o li ear o i ile erie i re a a lie o i ). e or
ile . . .
. a e M he S hmaier ai e he . e e io o ami e I i al o
i er a ial ro ei i i i S G. m hem So . .
. h a Mira a P im She . Ma i mole lar orie a io a o orma io
a i er a e r a e o li ear o i . Ph e . .
. a Pa i e M he . Mea ri ol mer r a e or eri i ere e i air
a a er m re e e era io i ra io al e ro o . m hem So .
.
. am er G a ie P ei a . Im leme i he heor o m re e e era io
i ra io al e ro o a orial re ie . l S e ro e . .
. iro e ama ome . orm la or he a al i o he r a e S G e r m a
transformation coefficients of cartesian SFG tensor components. Appl Spectrosc.
.
. Moa Sim o G . i e rea me o ele io r le a mme r rela io or
m re e a e o harmo i e ro o ie . Ph hem . .
. ee S a rimm S he . Irre i le re re e a io a ro e io o era or a li a-
io o er a i o li ear o i al he ome a h er ama m re e e era-
io a o r a e mi i e ro o . Ph hem . .
. Pa li ore . omi oor i a e a r re a or or o heli al o ra io
o ol e i e hai . Pro a l a S i S . .
. Pa li ore . o ra io o ol e i e hai i h a ore orie a io aro
i le o o e lea e hee . Pro a l a S i S . .
. he a o h o P ri al he . M l i le orie a io o meli i i i e
a i le li i ila er e ermi e om i e i ra io al e ro o i ie . m hem
So . .
. Mar h M ller M S hmi . rie a io o he i rare ra i io mome or a al ha
heli . io h . .
. i o l ar er S e ar S re eri PM. era i orie a io i ool a ea her
olari e ama e ro o . io ol mer . .
. ee S rimm S. i i io a e i ra io al a al i o al ha ol ala i e). io ol mer .
.
. ee S rimm S. Polari e ama e ra o orie e lm o al ha heli al ol ala i e)
a i e era e a alo e. ama S e ro . .
. a ee S he . a i i he or eri o a or e ro ei i i . Ph hem
. .
. he o h o P e mer G he . I i i e i a io o he ero rimeri G ro ei
subunit binding and orientation on membrane bilayers. detection of chiral sum frequency
e era io i ra io al e ra o ro ei a e i e a i er a e i i . m hem So .
.
. a he lar e M he . e e io o hiral m re e e era io i ra-
io al e ra o ro ei a e i e a i er a e i i . Pro a l a S i S .
.
S m re e Ge era io i ra io al S e ro o

. he a S ia e i e M he . Pro i -helical and -sheet structures of

e i e a oli li i i er a e i h S G. a m ir. .
. a ie P ai . S m re e e ro o o mo ola er o al o ermi-
a e al a e hiol i o a i h li i . a m ir. .
. a he SM he . Mole lar hemi al r re o ol me h l me ha r -
la e) PMM ) r a e ie m re e e era io S G) i ra io al e ro o .
Ph hem . .
. e r al er . S r a e i ra io al r re a al a e li i a or i er a e .
hem Ph . .
. olma a ie P i hi a e S ei a . S m re e e era io rom
a m ir lo e m l ila er lm o me al a iele ri ra e . Ph hem .
.
. ea ie a o S ai . om ara i e o o e or a i mo ola er
ama a eri a m re e e ro o . i S e ro . .
. a She . heor o o l re o a i rare i i le m re e a i er -
e e re e e era io rom a or e mole le . Ph e . .
. a h eM a a hi M i S She . o l re o a m re e e era io
e ro o or r a e ie . hem Ph e . .
. a a hi M i S a h e M She . mole lar heor or o l re o a I
i m re e e era io . Ph hem . .
. ree e m er Sar e aer a a o ol e Ma i e al. le ro i a
molecular properties of an adsorbed protein monolayer probed by two-color sum-frequency
e era io e ro o . a m ir. .
. he ar ia Mali ia Gra ia She e al. I era io o ri o e
with surfaces of end-group-modified polyurethanes: a surface-specific sum-frequency-gener-
a io i ra io al e ro o . iome Ma er e . .
. Merm Philli or M rea ar S Somor ai G . I i a or io
ie o a ami o a i le i e l i e e i e o o h ro ho i ol re e a h ro-
philic silica surfaces using quartz crystal microbalance, atomic force microscopy, and sum
re e e era io i ra io al e ro o . m hem So . .
. Philli or Merm M rea ar S Somor ai G . Si e hai hai
length, and sequence effects on amphiphilic peptide adsorption at hydrophobic and hydro-
philic surfaces studied by sum-frequency generation vibrational spectroscopy and quartz
r al mi ro ala e. Ph hem . .
. ar M i hmo G . rie a io a o orma io o ami o a i i mo ola er
a or e a a oil a er i er a e a e ermi e i ra io al m re e e ro o .
Ph hem . .
. a e M ierer i he olle o o S hoe hM .
ha e i a or e ri o e o o er io o ri . a m ir. .
. im remer PS. l i a i ha e i i er a ial a er r re o ro ei a or io .
hem h hem. .
. im im G remer PS. I e i a io o a er r re a he oli li i i er a e i he
presence of supported lipid bilayers by vibrational sum frequency spectroscopy. Langmuir.
.
. a SM e M he . Mole lar re o e o ro ei a i ere i er a ial
e iro me ee e m re e e era io i ra io al e ro o . m hem
So . .
. a SM he . S m re e e era io i ra io al e ro o ie o
ro ei a or io . Ph hem . .
. im Somor ai G . Mole lar a i o l o me ri o e a o i e er m al mi
on hydrophilic and hydrophobic surfaces studied by infrared-visible sum frequency genera-
io a ore e e mi ro o . m hem So . .
. im G G ra M im remer PS. I e i a io o l o me a or io a he air a er
a ar a er i er a e i ra io al m re e e ro o . a m ir.
.
 A.P. Boughton and Z. Chen

. S im SM l er orio im G G ra M a e al. he roma e e a

mole lar le el e ri io o ri o e i la eme . m hem So .
.
. he Sa le remer PS. rea orie a io a ro ei ra e. m hem So .
.
. ree e Sar e aer m er Ma i M hi ier Pra ier e al. Pro i li a
protein recognition with sum-frequency generation spectroscopy: the avidin-biocytin case.
hem h hem. .
. ree e Sar e aer m er Ma i emaire M hi ier e al. S m re e
generation spectroscopy applied to model biosensors systems. Thin Solid Films.
.
. a SM he . he e e o r a e o era e o o orma io ha e o
bovine serum albumin molecules at the airsolution interface detected by sum frequency
e era io i ra io al e ro o . al . .
. im o a S a re e Somor ai G . S r a e r ral hara eri a io o ro ei
and polymer-modified polystyrene microspheres by infrared-visible sum frequency generation
i ra io al e ro o a a i or e mi ro o . a m ir. .
. r i I. I o o e e i e I e ro o or he o mem ra e ro ei . rr i
hem iol. .
. r i I Ma e ie r er . Si e ire e i hroi m a a me ho or o ai i
ro a io al a orie a io al o rai or orie e ol mer . m hem So .
.
. orre am P r i I . e o a e la el ) ) i he e ermi a io o a
structural model of phospholamban in a lipid bilayer. Spatial restraints resolve the ambiguity
ari i rom i er re a io o m a e e i a a. Mol iol. .
. a Pa i lar e M he he . e io o r ral i orma io o i er a ial
ro ei om i e i ra io al e ro o i me ho . Ph hem . .
. a lar e M ha he he . i i o o e la ele ro ei a m re-
quency generation vibrational spectroscopy to study protein adsorption. Langmuir.
.
. lar e M he . Pol mer r a e reorie a io a er ro ei a or io . a m ir.
.
. roma . e o a or e ro ei o he e a ili o h ro hili a h ro ho i
oli . a re. .
.P i a lar e M he . S m re e e era io i ra io al e ro o -
ie o ro ei a or io o o i e o ere i r a e . Ph hem . .
. Moore rimm S. ra i io i ole o li i mi e I mo e o e a ol e i e .
Pro a l a S i S . .
. Shai . rom i a e imm i o e o o e i e a imi ro ial e i e . rr Pharm e .
.
. a . io o a imi ro ial e i e o a e mo el. io hemi r . .
. a Pa i e M he . I er re a io o m re e e era io i ra io al
e ra o i er a ial ro ei he hi lm mo el. Ph hem . .
. oe e Pa al i S a M i e ee ra . S m re e e ro o
a ima i o ali e heli al ol e i e . I Sel o a . .
. lar e M a he . o orma io al ha e o ri o e a er a or io . Ph
hem . .
. he a Pa i a S hra e ara ara e al. r ere a or io o
oa la io a or II o e a i el har e ol mer r a e ro e m re e
e era io i ra io al e ro o . al ioa al hem. .
. he ia he G ie S. a er a i ohere a i S o e ama a eri
mi ro o a a li a io o ell iolo . io h . .
. he ie S. ohere a i S o e ama a eri mi ro o i r me a io
heor a a li a io . Ph hem . .
S m re e Ge era io i ra io al S e ro o

. ol mer . i ra io al ima i a mi ro e ro o ie a e o ohere a i S o e

ama a eri mi ro o . Ph l Ph . .
. Moa Moa Perr M am ler Goe e GS e e e al. Pre i i -
ali a io a a a a al i o are or o li ear o i . om hem. .
. Moa Sim o G . Sel o i e a roa h or im li i he mole lar i er re a io
o o li ear o i al a m l i ho o he ome a. Ph hem . .
. Perr M Moa e e am ler Sim o G . le ro i a i ra io al e -
o or er o li ear o i al ro er ie o ro ei e o ar r ral mo i . Ph hem .
.
. a lar e M he . Polari a io ma i a me ho o im ro e m re e
e era io e ral a al i . al hem. .
. amm M o ell M. S or e ho holi i ila er . io h . .
. i o o . ire mea reme o he ra ila er mo eme o ho holi i
m re e i ra io al e ro o . m hem So . .
. i o o . Pha e ra i io o a i le li i ila er mea re m re e i ra-
io al e ro o . m hem So . .
. i o o . ia l ho ha i l holi e i o mea re ire l m re-
e i ra io al e ro o . io h . .
. Me mer M o o i hmo G . er a io o mole lar or eri a he li i li -
i i er a e re o a m re e e era io . m hem So . .
. he a e M a e G he . er i a mole lar i e a or .
m hem So . .
. he a ri al he . eal ime r ral i e i a io o a li i ila er
during its interaction with melittin using sum frequency generation vibrational spectroscopy.
io h . .
. iro e ama ome . orm la or he a al i o r a e mfrequency gen-
era io e r m re hi mo e o me h l a me h le e ro . hem Ph .
.
. iro e amamo o ama ome . rie a io a al i im la io o i ra-
io al m re e e era io e r m re hi a o he me h l ro . Ph
hem. .
. Ma G lle . PP a m ir mo ola er a he air a er i er a e ro i he ail a hea
ro i ra io al m re e e era io e ro o . a m ir. .
. o a a M lle . om e i io e ee a mo heri all rele a a a i
mo ola er a he air a er i er a e. Ph hem . .
. M ller M S hi M o e S o M. I o hello o ell il o e i or .
Phase transitions in a lipid monolayer observed with vibrational sum-frequency generation.
SPI ol. . . .
. e S o a Mori a S o a i a a M. S r a e mole lar r re o a m ir
Blodgett films of stearic acid on solid substrates studied by sum frequency generation spec-
ro o . a m ir. .
. eS o a i hi a Mori a S a a M. i e i er a ial r ral ha e
o a m ir lo e lm o eari a i o oli ra e a m re e e era io
. a m ir. .
. o e S S hi Mller M o M. i ra io al e ro o i i e i a io o he ha e
ia ram o a iomime i li i mo ola er. Ph e e . .
. o M o e S er rli S amo li M ller M. mole lar ie o hole -
erol i e o e a io i a li i mo ola er. Ph hem . .
. am a G. hi a e o he ar . a io e h ol. .
. Sar e aer o rillo G ree e i Ma i hir P e al. S m re e e era io
e ro o o mo ola er . io e ioele ro . .
. S o e G Gi a i M oma Se o ie G e S e al. Ma i
e e o . m hem So . .
 A.P. Boughton and Z. Chen

. lor heimer M riller h . hemi al ima i o i er a e m re e

mi ro o . a m ir. .
. She S ia ie i i iar Mar o i P Pra a P . Se o harmo i a m
frequency imaging of organic nanocrystals with photon scanning tunneling microscope. Appl
Ph e . .
. m er Gra em r ea S a er M a e i e . S e o ar m re e
generation spectromicroscopy at a submicronic spatial resolution. Appl Phys Lett.
.
. S haller Sa all . ear el i rare m re e e era io ima i o hemi al
a or e o i e i ele i e. a m ir. .
. o ma MP h e er . S m re e e era io mi ro o e or o a e a
re e i am le . e S i I r m. .
. h e o ma MP i er M er . hemi al ima i o i er a e
sum-frequency generation microscopy: application to patterned self-assembled monolayers.
l Ph e . .
. ima al elli S. S m re e e era io mi ro o o mi ro o a ri e mi e
el a em le mo ola er . Ph hem . .
Chapter 9
Near-Field Scanning Optical Microscopy:
A New Tool for Exploring Structure
and Function in Biology

Nicholas E. Dickenson, Olivia L. Mooren, Elizabeth S. Erickson,

and Robert C. Dunn

Abstract Optical microscopy in the biological sciences has proven invaluable for
understanding structure and dynamics in these complicated systems. It is a noninva-
sive tool that is amenable with fragile samples and is easily implemented under
physiological conditions. It also benefits from a wide array of contrast mechanisms
that can be exploited to gain detailed insight into sample properties. One limitation,
however, concerns spatial resolution, which is limited to approximately half the
excitation wavelength by the diffraction of light. Overcoming this limitation has
motivated the development of near-field scanning optical microscopy (NSOM) or
scanning near-field optical microscopy (SNOM). NSOM uses specially fabricated
probes to deliver light down to the nanometric dimension, enabling optical micros-
copy with a spatial resolution of tens of nanometers. For the biological sciences,
NSOM can simultaneously map sample fluorescence and topography with high spa-
tial resolution and single-molecule detection limits. This makes NSOM a powerful
new tool for understanding biological structure at the nanometric dimension.

9.1 Introduction

The ability to view and study biological structures directly on the spatial dimension
on which they exist has proven invaluable for understanding structure function rela-
tionships. Historically, optical microscopy has proven particularly informative and
throughout the years has steadily enhanced our understanding of tissue and cellular
structure. Optical microscopy provides a relatively inexpensive, versatile platform
for studying biological samples, with a wide array of complementary contrast

. . i e o . . Moore .S. ri o . . *)
e ar me o hemi r al h . am I i e or ioa al i al hemi r
he i er i o a a e er ri e a re e S S
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 225

I . S ri er I er a io al P li hi S i erla
 N.E. Dickenson et al.

mechanisms. For example, optical techniques can provide high time resolution,
olari a io a e ro o i a a ili ie hi h e i i a i le mole le
fluorescence-detection limits. Moreover, these can all be applied to living systems
under biologically relevant conditions in a noninvasive manner.
One limitation of conventional optical microscopy, however, revolves around the
maximal spatial resolution. When lenses are used to focus light, the maximal spatial
resolution is limited by the diffraction of light and is generally restricted to half the
excitation wavelength. When working in the visible part of the spectrum, therefore
spatial resolution with aberration free optics is restricted to approximately 250
300nm for conventional microscopy. Since many biologically interesting structures
exist on smaller scales, other microscopy approaches have been developed.
Transmission electron microscopy (TEM) and scanning electron microscopy
(SEM) have proven very powerful for understanding biological structures beyond
the reach of optical techniques. These methods can resolve structures down to the
nanometer scale and have been exploited to understand tissue architecture, mem-
ra e or a i a io a ma romole lar r re o ame a e . he e e h i e
have combined to provide a unique window into biological structures from which a
wealth of functional data has been extracted. While these techniques offer vastly
improved spatial resolution over optical microscopy, they do require extensive sam-
ple preparation and necessitate that the samples be in vacuum. These requirements
can limit the applications and preclude the study of viable biological samples. With
he e elo me o a i ro e mi ro o SPM) e h i e h a a omi
or e mi ro o M) ho e er a hie i a ome ri a ial re ol io o ia-
ble biological tissues is now routine.
M e a mall l o i erro a e am le r a e a a e im leme e
o e iolo i al am le er ere o i io . m er o M me h-
ods have been developed to probe various aspects of the sample surface, but, in
general, topography information is the key parameter extracted from these measure-
ments. These methods have nanometric spatial resolution, require very little sample
preparation, and are, in general, minimally perturbative to the sample. However,
since these measurements do rely mostly on topography, they lack the specificity
inherent in immunofluorescence or in studies done with fluorescent proteins.
Optical microscopy, electron microscopy, and to a lesser extent scanning probe
microscopy have proven important tools for probing biological structure and
function. Near-field scanning optical microscopy (NSOM) is a technique that can
complement these more established methods. In general, the goal behind the devel-
opment of NSOM was to create a technique that combined the favorable aspects of
the more established microscopies mentioned earlier. NSOM is a light-based
mi ro o ha re ai he e i i e ral a olari a io a a ili ie a
high time resolution inherent in optical microscopy. NSOM, however, circumvents
the diffraction barrier that limits the spatial resolution in lens-based optical micros-
copy and thus provides a spatial resolution nearing that of electron microscopy and
other scanning probe techniques.
 ear iel S a i i al Mi ro o e ool

Here we briefly review the basic concepts of NSOM, focusing on aperture-

based methods. Following a brief overview of the technical aspects, several
applications to biological samples will be discussed. These examples will help
illustrate the capabilities of NSOM and also highlight some of the difficulties
involved in probing biological systems with NSOM. These difficulties have
largely limited NSOM applications to fixed biological systems. Finally, we
will discuss progress in extending these measurements to viable samples,
which has proven extremely difficult.

9.2 Near-Field Scanning Optical Microscopy (NSOM)

In conventional light microscopy, where a lens is used to focus light, spatial resolu-
tion is limited by diffraction from the limiting aperture in the optical path, which is
all e ermi e he iame er o he o i eleme . i h a i hro h
the focusing element interferes around the focal point, generating a diffraction
pattern, the two-dimensional section of which in the focal plane leads to the well-
o ir i a er 1 . he i e o he e ral o i he ir i a er
i a e he ma imal re ol io a hie a le i h he o i al em. mi ha
collimated, coherent light is directed through an aberration-free microscope objec-
i e he he o i e i i e
spot size = 0.77l vac /n sin q . )
where vac is the vacuum wavelength, n is the refractive index of the medium through
which the light travels, and is the half-angle through which the light is focused by
the objective. The collection of terms in the denominator of Eq. ( . ) determines the
mi imal o i e a i olle i el o a he meri al a er re ) o he
o e i e 2].
hile o i e a e ire l a i e re ol io i a more e i e a -
i a i all i e i erm o he a lei h ri erio hi h a e ha o
features are resolvable if they are separated by a distance greater than or equal to
that given by Eq. . . With good signal-to-noise, finer separations can be resolved,
while imperfections in the optics, poor signal-to-noise, or sample limitations tend to
reduce the resolving power. Given these tradeoffs, the resolving power using con-
ventional optics is approximately equal to . he o e or i he i i le re io
of the spectrum, therefore, spatial resolution is limited to approximately 250
m 3, ].
Near-field scanning optical microscopy (NSOM) is a scanning probe technique
that enables optical measurements to be conducted with a spatial resolution beyond
. hi me ho e ol e o o he e lo io i a i ro e e h olo ie h
a a omi or e mi ro o M). S M o er ome he i ra io limi
i li h a i hro h a a ele h i e a er re o eli er li h o o
 N.E. Dickenson et al.

Fig. 9.1Schematic
representation of light
passing through the
subwavelength-diameter
aperture of an opaque screen

he a ome ri ime io 5, . i h e i i he a er re o a S M ro e ill

diffract out, as shown schematically in Fig. . . o i io i he a er re ear he
sample surface, light exiting the aperture is forced to interact with the sample before
i ra i o a he a ial re ol io i o l limi e he i e o he a er re
and its proximity to the sample surface.
While the experimental details involved in implementing NSOM can be found
el e here 5, , ], we will briefly discuss aperture formation and implementing a
feedback system for tip-sample distance control. These present the main obstacles
that must be overcome for implementing NSOM measurements.
Several aperture-formation methods have been developed for NSOM, but those
based on fiber optics have been the most widely adopted. In general, a single-mode
optical fiber is heated and pulled to form a taper terminating at a small point that
e e all orm he S M a er re . e o me ho o ormi S M
probes uses chemical etching techniques to create the taper in the fiber. This method
uses hydrofluoric acid (HF) as the etchant and usually employs some type of organic
protection layer. The probe is dipped into the HF, and the meniscus formed by the
a or a i la er e ermi e he o erall ha e a o e a le o he i , 10].
While the pulling method generally provides a smoother taper in the probe, etching
methods can create much smaller aspect ratio tapers, which have advantages for
light throughput.
To confine the light within the tapered region, the sides of the probe are coated
with a reflective metal coating. For the visible region of the spectrum, aluminum is
the most reflective, with a skin depth of ~13 nm at a wavelength of 500nm. The high
reflectivity enables confinement of light within the taper region of the probe with
o l m o al mi m 11]. Figure . shows magnified views of a typical
fiber-optic NSOM probe. In the electron microscopy image shown in Fig. . a, the
granules from the aluminum coating are visible, as is the aperture at the very end of
the probe where the light emerges. The aperture is more apparent in Fig. . , which
ho he ame ro e a er milli i h a o e io eam I ) i r me . I he
optical image shown in Fig. . , light emerging from the nanometric aperture of
heated and pulled (a) and chemically etched (b) probes is visible.
 ear iel S a i i al Mi ro o e ool

Fig. 9.2(a) Scanning electron micrograph of a tapered, single-mode, fiber-optic, near-field probe
coated with aluminum to confine light. (b) The same probe after the end has been milled with the
e o a o e io eam I ) i r me . S ale ar are m i re a a e rom e . 12])

Fig. 9.3 Ma i e ie o l oa e S M ro e ma a re a) heating and pulling

and (b) the Turner method of chemical etching. The scale bars represent 30m. o h ro e
exhibit single apertures at their apexes, as indicated by the single spot of light exiting the probes.
The overall geometries become apparent in the inserts, which are expanded views of the same
probes (scale bars are 50m) i re a a e rom e . 13])

Once the NSOM probe is formed, high-resolution NSOM measurements rely on

the ability to position and maintain the probe within nanometers of the sample sur-
a e. m er o me ho ha e ee e elo e i l i a roa he a e o
hear or e ee a a i mo e , ]such as that developed for
a omi or e mi ro o a i or ee a . ll o he e me ho
generally rely on oscillating the tip and monitoring changes in amplitude as the tip
i era i h he r a e. ee a loo i im leme e ha mo e he i rela -
i e o he am le r a e o ee he i o illa io o a . ll he e me ho
230 N.E. Dickenson et al.

Fig. 9.4 . m1.25-

m NSOM fluorescence
image of single diI molecules
e ra e i hi a PP
monolayer. The full-width
half-maximum of the
intensity features is ~25nm,
hi h re e he i e o he
NSOM aperture used in the
imaging (Figure adapted from
e . 21])

enable nanometer control over the tip-sample gap and generate a force mapping of
he am le o o ra h m h li e M. S M here ore ro i e im l a eo
o i al a o o ra hi al i orma io a o am le ro er ie 20].
To illustrate the high spatial resolution and low detection limits of NSOM,
i . . ho a S M ore e e ima e o he e mole le iI (1,1-diocta-
e l e rame h li o ar o a i e er hlora e) o e i o a PP l--
dipalmitoylphosphatidycholine) lipid monolayer. Each bright spot in this image
represents the fluorescence from a single dye molecule in the lipid monolayer,
which illustrates the single-molecule detection limits of NSOM. Since the dye mol-
e le i m h maller ha he S M a er re he i e o ea h ore e e ea re
i i . . re eal he i e o he S M a er re e o ima e he am le. he
full-width half-maximum (FWHM) of the features is approximately 25nm, which
is roughly an order of magnitude better than the diffraction limit and illustrates the
high spatial resolution possible with NSOM. For the biological sciences, this high
resolution, along with the simultaneous force mapping of surface topography, pro-
vides a potentially powerful new tool. One concern, however, revolves around
o e ial am le hea i a e he S M ro e. li h i ire e hro h
the NSOM probe, heating occurs, which could potentially damage biological
am le . e ore i i a li a io here ore e ill rie mmari e li h
induced heating of NSOM probes.

9.2.1 Sample Heating from NSOM probes

Sample heating from the NSOM probe has raised some concerns for biological
applications. Metal-coated, fiber-optic NSOM probes increase in temperature as
more li h i o le i o he ro e. he iame er o he S M ro e e rea e
in the taper region, eventually no propagating modes of the light are supported and
 ear iel S a i i al Mi ro o e ool 231

Fig. 9.5 Plo o am le em era re a a io o o o er or e re re e a i e S M

probes fabricated using the pulling method. Each symbol corresponds to the heating curve of an
i i i al S M ro e. he i er ho he ormali e em era re e e e emi io e -
tra for the thermochromic polymer, perylene and N-allyl-N-methylaniline, used to obtain sample
hea i a a i re a a e rom e . 21])

only highly decaying evanescent fields reach the aperture. This makes NSOM
ro e hi hl i e ie a eli eri li h 22, 23]. Typical NSOM probes have
losses on the order of three orders of magnitude as the light travels toward the aper-
ture. With tens to hundreds of microwatts coupled into the fiber optic, only tens to
hundreds of nanowatts are delivered from the aperture, when measured in the far
field. The remaining light is lost to reflection back up the fiber and to absorption
losses into the metal coating surrounding the taper in the probe. The latter contrib-
utes to probe heating. Eventually, as more light is coupled into the NSOM probe,
heating causes catastrophic probe failure where the aluminum coating forming the
a er re i lo 3].
Mea re o ro e hea i are om li a e he mall i e he S M a er -
ture. Early studies used thermocouples positioned tens of microns from the tip aper-
re a o em era re a hi h a e ore e e al i ail re ].
Studies attempting to measure temperatures directly at the NSOM aperture yielded
re l ha ra e rom le ha o 25 . e e l he e mea re -
me ere e e e i a hermo hromi ol mer o ire l hara eri e he
hea i e erie e am le elo a S M a er re 13, 21].
These studies used a polymer consisting of perylene and N-allyl-N-methylaniline
) ha e hi i a o olor emi io e r m ha ha e i h em era re.
232 N.E. Dickenson et al.

Fig. 9.6 Se e e o ele ro mi ro ra h o a l oa e er o i S M ro e a e a he

laser power coupled into the probe was increased. (a) lo la er o er he al mi m oa i o
the tip remains intact. (b) he o er i i rea e a ri li e ra re i he l oa i a ear
a a ro e iame er o m and 35m from the apex. This results from the differential thermal
e a io o he l oa i a he ar er. c) Finally, at higher powers, the complete failure
of the probe is marked by the removal of the coating below the fracture, exposing the bare fiber
beneath. Scale bars are each 10m i re a a e rom e . 13])

he l emi io e r m o i o o ea lo a e a ma m.
o hea i he ra io e ee he ma m ea i rea e i . . ),
which provides a marker of sample heating that is not sensitive to light intensity or
am le o e ra io , . h ra iome ri mea reme o he ea
i e i ie e a le re i e hara eri a io o ro e hea i ire l elo he
NSOM aperture.
Sample heating as a function of output power from a sampling of NSOM probes
is shown in Fig. . . or all ie he o o a m io e la er Po er
Technology) was coupled into the NSOM probes. In general, all probes showed an
initial rapid rise in sample heating that leveled out at elevated output power.
o o er o a e a o a hea i a he am le ra e rom
o a le ele o a ma im m o a ro ima el a S M ro e o
o er ra i rom o 21 . Sli h aria io i i a er re i e a
taper geometry contribute to the range of heating results observed in Fig. . .
Interestingly, similar sample temperature profiles were measured with NSOM
ro e a ri a e i h he hemi al e hi e h i e 13]. Despite being much
more efficient than probes fabricated using the heating method, sample heating with
output power remains unchanged. In either case, the sample heating is modest,
especially at low powers, and should not pose a major difficulty for biological appli-
cations. Moreover, the sample heating is expected to be lower under aqueous condi-
tions, where heat can be more efficiently released into the surrounding media.
The leveling off in sample heating was ascribed to the heat-induced expansion of
the NSOM probe, which increases the distance between the probe aperture and the
sample surface. The large difference between the thermal expansion coefficients
of the fiber optic and surrounding aluminum coating ( l . 5
1 and
ar =0.510 1
) leads to differential elongation in the probe as power is
i rea e , ]. This effectively increases the distance between the aperture and
the sample as power is increased.
 ear iel S a i i al Mi ro o e ool 233

The large difference in thermal expansion coefficients for the fiber optic and
s urrounding metal coating of the NSOM probe also explains the catastrophic damage
experienced at high powers. Figure . displays a series of high-resolution electron
mi ro ra h a e o he ame S M ro e a he i o er a i rea e . lo
powers, no damage is observed in the metal coating of the probe and the aperture at
the end of the probe exposing the fiber dielectric appears dark (Fig. . a). o er
is increased in Fig. . , a dark ring representing a fracture in the aluminum coating
is first observed approximately 35m rom he a er re. o er i r her
increased, this initial fracture eventually leads to the complete loss of aluminum from
the region below the fracture, as seen in Fig. . . This loss of aluminum from around
the taper region leads to the loss of the NSOM aperture and a dramatic increase in
li h hro h ha i i i a i e o S M ro e ail re a hi h o er 13].

9.3 Biological Applications of NSOM

NSOM is generally used for studying surfaces and has been widely employed for
i iolo i al lm a mem ra e . or re ie ee 5 a i i ].
These studies have taken advantage of the simultaneously collected fluorescence
a o o ra h i orma io i h S M or hara eri i he a e o he mem-
brane systems. Small domains in biological membranes have long been of interest;
the high resolution of NSOM provides a tool that now allows these domains to be
studied directly.

9.3.1 Membranes

9.3.1.1 Monolayers

The complex composition and structure of biological membranes have led to the
development of methods for studying model membranes under more controlled con-
i io . he e o a m ir lo e lm allo or he rea io o mo ola er or
multilayer films, where temperature, composition, and phase can be precisely con-
rolle . a m ir lo e lm ha e here ore ee i el a o e o li i
ha e a om o e i ri io i mo el o a ral mem ra e 30, 31].
The phases and component distribution in model membranes have been studied
extensively with fluorescence microscopy. Many studies have shown that the inclu-
io o ore e li i a alo h a iI , into the lipid films will preferen-
tially partition into the less-ordered regions of the film. For lipid monolayers
ra erre o o ra e i he li i e a e ) li i o e e )
phase-coexistence region of the pressure isotherm, therefore, the condensed regions
appear dark and the expanded regions appear bright in the fluorescence image.
For example, Fig. . a ho a ore e e ima e o a PP li i mo ola er
 N.E. Dickenson et al.

Fig. 9.7(a) o o al ore e e a b) M ima e o a PP mo ola er e o i e a

m m o o lea e mi a i he a m ir lo e e h i e. hi r a e re re he
li i o e e ) a li i e a e ) ha e o he mo ola er oe i . a) In the fluo-
re e e ima e ri h re io i or ora e a ore e li i a alo a mar he omai .
he ar omai are he ha e. b) I M mea reme he omai a ear hi her
ha he rro i mi ro omai . o e ha he a eara e o mi ro omai i hi he
e a e re io o he M ima e i o re ol e i he o o al ima e i re a a e rom
ollar a 32])

o e ih . mol o iI ra erre o o a mi a r a e i he
ha e oe i e e re io . I he ima e emi ir lar ar omai are r -
ro e ri h re io ha o ai he ore e li i a alo 32].
The phase structure in lipid monolayers can be studied with higher spatial
re ol io i e h i e ha M. he mall . . m hei h i ere e
e ee he a ha e i PP a e mea re i h M o ma he
phase partitioning with nanometric spatial resolution. Figure . ho a M
ima e o a PP mo ola er ra erre er he ame o i io a ho e ho
in Fig. . a. Semi ir lar omai o a hi her o o ra h mar he re io o
he lm. he lo er o o ra h re io re eal a mi ro r re o re ol e i
the lower-resolution fluorescence image shown in Fig. . a.
NSOM measurements offer the advantage of measuring both high-resolution
fluorescence and topography simultaneously, which are particularly informative for
studies of model membranes. Figure . shows the NSOM fluorescence (a) and
o o ra h ) ima e o a PP mo ola er o e i h . mol iI , trans-
ferred under the same conditions as the films shown in Figs. . a . The higher
a ial re ol io i he ore e e ima e re eal he mi ro r re i he
regions not observed with far-field fluorescence measurements. These structures
can be directly correlated with areas of lower topography seen in the NSOM topog-
ra h ima e o rmi he ari e rom omai 32].
The high spatial resolution of NSOM fluorescence should be useful for under-
a i li i ra . i i ra are mall omai lo a e i hi he ell mem ra e
that are rich in cholesterol, glycolipids, and sphingolipids. These elusive structures
 ear iel S a i i al Mi ro o e ool 235

Fig. 9.8(a) i h re ol io S M ore e e ima e o he ame PP mo ola er e i e

inFig. . . he ri h omai re eal he ha e re io i or ora i he e hile he
ar area re re e omai . b) NSOM topography image collected simultaneously with the
fluorescence shown in (a). he mall omai o er e i he M ima e are i i le i
both the NSOM fluorescence and topography measurements (Figure adapted from Hollars and
32])

are ho h o e or a i i eleme o he mem ra e e o mo i he io

o i e ral mem ra e om o e . a ere r la i e a ell lar mem ra e
fractions that are resistant to cold, nonionic detergents such as Triton X-100. While
there have been extensive biochemical studies of these detergent-resistant fractions,
the believed small dimensions of rafts and their mobility within biological mem-
ra e ha e ma e hem e ee i l i l o ire l 33, ] .
e e l he e o S M o mo el mem ra e ha le o i i h i o he
distribution and characteristics of submicron domains in lipid films. For example,
Johnston and coworkers investigated the distribution of a commonly used lipid raft
mar er a leo i e GM i a er ar li i mi re 35 . or he e ie a
mo ola er om o e o a mol ra io mi re o P hole erol a
sphingomyelin was prepared containing 1% GM1, a gangleoside that preferentially
ar i io i o li i ra . mall amo o a ore e l la ele li i e a
e PP a . IP GM ro i e o oro hore ha are o h
e ie ih m ra ia io . i re . a shows the NSOM fluorescence image of
he mo ola er ollo i m e i a io i hi h o h la el are e i e . ar
emi ir lar omai are rro e ri h re io i he S M ore -
cence image. Of particular interest are the small, ~150-nm bright microdomains
o er e i he ar re io o he lm. o er a he ori i o he e
omai he e i a io a ele h a ha e o m hi h o l e i e he
e a e PP .
Figure . shows the same region of the membrane, with excitation wavelength
a m. hi e i a io e er o l he e a e PP la el i e i e a
the fluorescence reveals the phase separation in the monolayer. The bright microdo-
mai re io l ee i he re io are o a e i i a i ha he e
 N.E. Dickenson et al.

Fig. 9.9 These are 25-m25-m NSOM fluorescence images of a 1:1:1 ternary mixture of
hi om eli hole erol P mo ola er ra erre o o mi a a m m i he
me ho . he e lm al o o ai e a e PP a leo i e GM a .
o i la ele GM . Ima e a) re re e he ore e e rom o h he GM o i ra
mar er a he PP . b) he ame area o he lm i h o l he PP oro hore
e i e . Similar area o he o a are o li e i h o e e ri e i h ermi io rom
r o e al. . o ri h he meri a hemi al So ie 35])

domains contain the labeled GM1 raft marker. These domains are consistent with
he orma io o mall ra omai hi h a ree i h e ree Mmea reme
ho i imilar mall omai i he re io ].
The high-resolution fluorescence and topography information from NSOM has
al o ee e loi e o hel er a re ira or i re rome S). Si
associated with the lung surfactant, a monolayer of lipids and proteins that lines the
l a re e he r a e e io o ear ero. Prema rel or i a are
o e i i e ih S hi h re l rom i ie l e elo e l ra -
tant. These infants have poor oxygen transport and labored breathing; left untreated,
S a e a al . S a e rea e hro h leme i he e ie l
surfactant with artificial or natural lung surfactant mixtures. Survanta is one treat-
me or S ha o i o o i e l r a a or i e i h PP . S M
has been used to help understand the phase structure and collapse of monolayers of
S r a a ].
or he e ie mo ola er o S r a a o e i h . mol o he o-
re e li i a alo iI were transferred onto a freshly cleaved mica surface at a
erie o r a e re re . a r a e re re o m m elo he ol-
la e re re o o al ore e e ima i i e i e li i he lm.
shown in Fig. . , the fluorescence image reveals a broad area where the phase
structure overlaps, indicating buckling in the film.
 ear iel S a i i al Mi ro o e ool

Fig. 9.10 m100-m

confocal fluorescence image
of a Survanta monolayer
deposited onto mica with the
e h i e. a ra er
re re o m m he
film is near collapse and the
overlapping phase structure
in the fluorescence indicates
the film has buckled. The
scale bar is 10m (Figure
a a e rom e . ])

Fig. 9.11 Two 20-m20-m simultaneously collected NSOM fluorescence (left) and topogra-
phy (right) ima e o a ol i a S r a a lm e o i e he e h i ea m m. o h
images indicate a discontinuity in the monolayer consistent with the buckling suggested by far-field
ima e i re a a e rom e . ])

Figure . shows NSOM fluorescence (left) and force (right) images of a

Survanta monolayer transferred onto mica at the same surface pressure. The NSOM
fluorescence image reveals a similar phase structure and overlapping structures
i i ai eo li . om ari o o he S M ore e e i h he im l a e -
ous force images confirms the formation of a multilayer region in the film as the
membrane buckled and folded over itself. The simultaneously collected topography
ho a ri e o a ro ima el m i he re io o he ol o rmi ha he
membrane has partially folded over itself. This is consistent with the discontinuity
in the phase information seen in the fluorescence image. Upon raising the surface
re re e o he olla e re re o he mo ola er o m m he mo ola er
no longer buckles but appears to break off from the rest of the membrane. The
 N.E. Dickenson et al.

Fig. 9.12 hema i upper left) of the electric fields near a metal-coated NSOM aperture as
mo ele e he o am heor . he z-component in the electric field near the edge of the
aperture can excite molecules with transition dipoles pointing at the probe. This leads to excitation
when the molecule is located near the aperture edges and no excitation when centered below the
aperture. The single-molecule fluorescence image reveals shapes indicative of molecules with this
orie a io i re a a e rom e . ])

S M aa e ha mem ra e li ma e o e me ha i m hi h S
i a le o re i olla e a hi h re re ].
NSOM also has single-molecule fluorescence detection limits, which have been
e loi e o he mole lar le el r re i li i mo ola er . I e i
a hi he er li he a emi al S M a er ha o o l re or e he i le
molecule detection limits of NSOM, but also showed that the fields present at the
NSOM aperture lead to patterns in the single-molecule fluorescence that reflect the
hree ime io al orie a io o he mole le ]. Using carbocyanine dyes dis-
er e i a ol me h l me ha r la e PMM ) ma ri he o a erie o emi -
sion patterns that could be understood by modeling the fields at the NSOM aperture
i e he o am heor ]. The boundary condition between the metal
coating and the dielectric fiber optic leads to a curvature in the fields at the NSOM
aperture, as shown in Fig. . . Molecules with transition dipoles vertically oriented
will be excited when near the edge of the aperture, while those lying in the plane will
be excited when centered under the aperture. This difference leads to patterns in the
emi io ima e hi h a e a al e o e ra e orie a io i orma io .
These single-molecule emission patterns were used in single-molecule NSOM
ie o ro e he li i mem ra e r re a he mole lar le el ]. The goal
a o a al e he i le mole le S M emi io a er rom ore e li i
analogs doped into the lipid monolayer as probes of local structure. Figure .
 ear iel S a i i al Mi ro o e ool

Fig. 9.13 omi or e

micrograph of a
. m . m PP
ila er. o h la er ere
e o i e a m m i he
phase-coexistence region of
the pressure isotherm. The
second layer was transferred
he a m ir S hae er
method, creating the bilayer.
he hree a i e hei h
levels in the bilayer differ by
a a re l o he ha e
overlap in the two layers
(Figure adapted from Hollars
a 32])

ho a S M ore e e ima e o a PP mo ola er o e i h he ore -

e li i a alo I P . a al i o he ea re o ha he ra i io
i ole mome o I P a orie e a ro ima el rom he ormal
a a o e i i e o ha e i r a e re re rom . o . m m. hi
o o e i e e or I P orie e arallel o he r a e
o he o er e i al a a re l o a alo orie e a a a le o or e o
from the normal.
In theory, these types of measurements can reveal the orientation of molecules
embedded in a matrix and provide a molecular-level marker of orientation. However,
other studies have found evidence of tip-induced changes in emission patterns that
a i e e e ra e orie a io , ]. Further work understanding the mag-
nitude of this effect is needed before confident orientations are extracted from
single-molecule NSOM fluorescence measurements.

9.3.1.2 Bilayers and Multilayers

NSOM measurements have also proven useful for studies on multilayer membrane
systems. The complexity of these systems quickly increases with the number of lay-
ers. For example, Fig. . ho a M ima e o a PP ila er here ea h
leaflet of the bilayer was transferred in the phase-coexistence region of the pressure
i o herm. or he e am le he r la er a ra erre o o mi a i he
e h i e a he e o la er i he a m ir S hae er a roa h. hree
a i e hei h le el are o er e i he M ima e o he ila er 32]. These
hei h ha e orre o o he o i le om i a io o a i e ee he
a ha e re e i he i i i al lea e o he ila er. or e am le he lo -
e o o ra h ea re i he ima e orre o o he a i o o he
e le el orre o o o or o a all he alle r re
 N.E. Dickenson et al.

Fig. 9.14 m10-m S M a o he PP ila er e i e i i . . . The fluores-

cent diI lipid analog is included only in the bottom layer, while the upper layer is comprised of pure
PP . he hi h re ol io ore e e ha el i ho i a) and the simultaneously collected
sample topography is depicted in (b) i re a a e rom ollar a )

orre o o o . So hile i i lear ha ha e r re i re e i ea h

lea e o he ila er a i i hi r re o ea h lea e i o o i le i h M.
NSOM, however, can be used to study the microdomain structure present in each
side of the bilayer by controlling which leaflet of the bilayer contains the fluorescent
lipid analog. Figure . ho he S M ore e e a o o ra h o a PP
bilayer similar to that shown in Fig. . . In this bilayer, the lower leaflet was doped
i h he l ore e li i a alo iI hi h ar i io i o he ha e.
he ore e e ima e ho a ha e r re imilar o ha o er e or PP
mo ola er i h ar re io rro e mi ro omai . Mea reme
with dye doped into the top layer reveal a similar structure. These measurements
show that the phase structure is maintained in forming the bilayers and suggest that
microdomains and lipid asymmetry can be probed with high resolution in these
more om le em 32].
In biological membranes, asymmetry is maintained across the bilayer and is
ri i al or ro er ell lar io . i i a ro ei om o i io i i ere
across the two leaflets of the plasma membrane of cells and loss of this asymmetry
i a i o a o o i . Pro i hi a mme r ire l i i l i e he mem -
ra e hi e i o l a ro ima el m. l ore e e re o a e e er ra -
er ) ho e er i a ore e e e h i e ha i e i i e o i a e o
hi le h ale a ha ee e e e i el o ro e mem ra e or a i a io .
a e a a a eo r er e er ra er e ee a o or a a e or
dye that are in close proximity to each other. Nonradiative energy transfer between
dyes with the proper spectral properties is very sensitive to changes in the distance
e ee he o e ma i a o er l ool or mo i ori mall i la e-
me i iolo i al em , ]. Several groups have combined the distance
e e e eo i h S M o rea e a e e h i e ha ha hi h la eral a
 ear iel S a i i al Mi ro o e ool

ell a a ial re ol io hi h ho l e a a le o hara eri i mem ra e

a mme r ].
i i ial emo ra io o hi e h i e ili e a a ere er o i ha
lacked the metal coating used in NSOM probes to confine the light within the probe.
he e o he ro e a i ea oa e i he e h i e i h a PP
mo ola er o e i h a rho ami e la ele li i ]. The sample used in these
ie o i e o a m l ila er lm o i i o a ore ei o e PP
monolayer, three layers of arachidic acid that did not contain any dye, and finally
a o her la er o e o e PP . hema i o he e erime al arra eme i
shown in Fig. . . or he e ie he ore ei a e a he o or e
a he rho ami e a a he o he S M ro e a e a he a e or.
Figure . shows the NSOM fluorescence images of the multilayer film follow-
i e i a io o he o or a m. he o ima e ho re re e he ore -
cence from the donor (Fig. . a) and acceptor (Fig. . ). om ari he
o ima e ill ra e he e io i a a ili ie o om i i i h S M.
The features in the donor emission shown in Fig. . a reveal the domains in both
the top and bottom layers of the film that contain the donor dye though it is not pos-
sible from this image to assign from which layer the emission originates. The acceptor
emission shown in Fig. . , however, arises from energy transfer from the excited
o or i he am le o he i o a e or. h he e er ra er e -
ciency from the top layer of the film is much higher than that from the bottom layer.
In Fig. . , some domains become much weaker, indicating they reside in the
bottom layer while other domains become brighter since they reside in the top layer,
closer to the NSOM probe. The resolution of these domains also sharpens as
expected since only dyes in closest approach contribute to the energy-transfer signal.
ho i i . . hi emo ra e ha he S M e h i e a
ro e am le i h a la eral re ol io o m a z-sectioning capabilities of
le ha m.
o her a a a e o om i i i h S M i ha he e h i e oe
not require the fiber-optic probes normally used in NSOM to deliver light. This was
emo ra e here io ali e M ro e ere e i he o ra-
io o ma e hi h re ol io o i al mea reme . e ore a a e or e a
a a he o a ili o i ri e M ro e a he o or a i ri e i he am le.
Far-field excitation was used to excite the donor, and emission from the tip-bound
a e or a mo i ore o e era e he hi h re ol io ima e. e ore hi h re o-
lution is obtained because of the strong distance dependence of the energy-transfer
me ha i m ].
I Sh ei a a o or er e e e he S M e h i e ].
In these studies, the acceptor dye was attached to an uncoated fiber-optic NSOM
ro e i a m o i all ra are PMM la er o ai i he e.
This creates a dye reservoir that extends the life of the measurement as mechanical
wear depletes the acceptor from the probe. This creates a more robust system for
ai e S M mea reme .
 N.E. Dickenson et al.

Fig. 9.15 Schematic (top) o he e erime al e e i emo ra i he S M

me ho . he er o i ro e i oa e i h a mo ola er o ai i a a e or oro hore.
The sampled consisted of a multilayer film where only the top and bottom layers contained the
donor dye. (a) This image shows the fluorescence from the donor dye in the sample. (b) e or
emission that arises from energy transfer from the excited donor dye in the sample to the tip-bound
a e or. om ari o o he o ima e a e i he ame area ho ha ome ea re i he
S M ima e lo e i e i i i a i he ari e rom omai i he o om la er hile
others becoming brighter and more resolved, indicating they are from the top layer (Figure adapted
rom e . ])
 ear iel S a i i al Mi ro o e ool

Fig. 9.16 m10-m S M a o he ame mem ra e ho i i . . . The

tip-bound acceptor fluorescence was collected, while the excitation wavelength was resonant with
he am le o o or oro hore. he re l i ore e e ima e re eal m ea re
i re a a e rom e . ])

Fig. 9.17 Two 20-m20-m NSOM fluorescence (a) and topography (b) images of a human
ar erial moo h m le SM) ell i hi h he a i ha ee la ele i h a ore e hal-
loidin conjugate

9.3.2 Cells

The extension of NSOM measurements into intact biological samples such as viable
cells has been difficult. Working with fragile samples under buffered conditions
presents several challenges for NSOM measurements that are proving very difficult
 N.E. Dickenson et al.

Fig. 9.18 Near-field images of a fibroblast cell incorporating a fluorescently labeled phosphatidyl-
holi e i o he la ma mem ra e. he r a e re e i la I ro ei ere al o imm -
nolabeled with a tetramethylrhodamine-conjugated antibody. The near-field fluorescence image
(a) reveals the location and distribution of the fluorescent lipid analog, while (b) shows the anti-
o la eli o he ro ei . c) Simultaneously collected topography image of the cell
e ro e i h ermi io rom e . 50 . o ri h he io h i al So ie )

to overcome. The limited success in this area will be discussed in the next section.
Fixed biological tissues, however, are amenable with NSOM measurements both
dry and under buffered conditions, opening vast new areas of research. For studies
of fixed cells, the simultaneous collection of both high-resolution fluorescence and
or e i orma io i ar i larl e l a i orma i e. a e am le i . .
shows the NSOM topography (a) and fluorescence (b) images of a fixed human
ar erial moo h m le SM) ell i hi h he a i ha ee ore e l
la ele i h halloi i o a e o he ore e ale a or . ro
correlation is observed between the F-actin fibers identified in both the fluorescence
and force channels. The smallest fibers observed in the fluorescence have an FWHM
o m hi h re e he i e o he S M a er re e i he ima i .
In addition, many of the fibers observed in the fluorescence image are not resolvable
i he or e ima e a o l o e ee i h e h i e ha M.
Edidin and coworkers exploited the high resolution of NSOM to probe lipid
ra i h ma ro la ell 50]. The fibroblasts were dual labeled with
IP ho ho holi e IP P ) a a mo o lo al a i o e i or
he la ro ei o he ma or hi o om a i ili om le M ). he la er
was labeled with fluorescent tetramethylrhodamine. Figure . a shows the high-
re ol io S M ore e e ima e o he IP P i he ro la mem -
brane. Small, ~200-nm dark domains that exclude the fluorescent probe are visible
in the upper left region of the cell. Interestingly, the NSOM fluorescence of the
M ho i i . . reveals fluorescence from the areas seen in Fig. . a
to exclude the membrane probe. This supports the idea that small domains or rafts
are or e i he mem ra e a e o or a i e mem ra e om o e .
These measurements are also consistent with those done previously on model
mem ra e em 35]. The topography measurements shown in Fig. .
reveal no significant correlation between the membrane topography and the loca-
tion of the microdomains.
 ear iel S a i i al Mi ro o e ool

Fig. 9.19(a) S hema i o he e erime al e ili e or olle io o o o al a S M

images. (b) m20-m confocal fluorescence image of a dendritic cell with fluorescently
imm ola ele SIG e re e o he mem ra e. c) o erla o im l a eo l olle e
topography (gray) and NSOM fluorescence (color) of the region of the dendritic cell outlined in
(b). he S M ore e e ha ee olor o e a or i o emi io olari a io . Polari a io
a i o e red emi io a i green, and emission in between is coded as shown by the
gradient scale

More re e l e a er a o or er 51] investigated the distribution and

mole lar or a i a io o SIG a e le i e o e o he r a e o e -
ri i ell ). o e l er o SIG are re o i le or he i e
recognition of a variety of pathogens through binding to specific haptens. Though
he role o he e le i i o he me ha i m hro h hi h he re o i e
specific pathogens is not well understood.
Figure . a shows a schematic of the instrumental setup used to collect the
confocal and NSOM data. Figure . shows a confocal fluorescence image of a
e rie i hi h he SIG le i are la ele i h mo o lo al a i o-
ie a ai he e o e e i o e o SIG . o a e e o ar a i -
body is used as the fluorescent marker of the antibody labeling. Figure .
displays a 12-m m S M mea reme rom a e io o he ho i
Fig. . . The NSOM results are presented as an overlay of the fluorescence and
o o ra h i orma io . Polari a io i orma io a al o olle e im l a e -
ously onto two detectors and color-coded as a gradient scale where red represents a
olari a io o a ree re re e olari a io .
he S M a a re eal mall l er o SIG ha are a ro ima el
m i iame er. he olari a io a a e ha mem er o he l er are
ra oml orie e al ho h ome i le mole le e e re l e i re olari a-
io a i e i e he ir le . Gar ia Para o a o or er ere a le o om i e
hi i orma io o e ermi e SIG e i hro h ali ra e i e i mea -
reme a al o e lore he mole lar or a i a io o e eral l er hro h
e e ial S M ima i i h i le mole le ho o lea hi . ili i a
Gaussian fit, the authors were able to determine the position of the fluorophores to
i hi m. he ima e a al i o he hi h re ol io S M a a re eale
SIG l er or a i a io ha i i e he ero e eo a rai ha ma e e i
he a ili o imma re e ri i ell o re o i e a lar e arie o a ho e .
 N.E. Dickenson et al.

9.3.3 Tip-Sample Interactions

The previous examples on model membranes and fixed biological tissues help to
illustrate the utility of NSOM in the biological sciences. The high-resolution fluo-
rescence and topography information, single-molecule detection limits, along with
he ho o e ro o i ime re ol e a olari a io a a ili ie o er a o er-
ful set of tools. It should also be noted that since resolution does not depend on
wavelength, as it does for far-field measurements, moving to longer wavelengths to
avoid autofluorescence will not sacrifice resolution. However, extending NSOM
measurements into unfixed biological tissues under buffered environments has
proven extremely difficult.
One of the biggest obstacles for implementing NSOM measurements on fragile
biological samples revolves around the forces generated in the feedback mechanism
e o o rol he i am le a . e a e i al er o i S M ro e ha e
lar e ri o a he om are ih M ro e he ha e ee le
successful at imaging unfixed biological tissues under buffered conditions. While
certain geometries have been shown capable of noninvasive imaging of unfixed
samples, these often rely on complicated fabrication procedures that have reduced
their utility. Future progress in this area will likely require a dramatic departure
from current probe designs or feedback approaches.
Early NSOM approaches on fragile samples circumvented the force problem by
scanning samples in constant height mode, with the tip feedback turned off. While
some information can be gained from such studies, generally this is an unsatisfac-
or mo e o olle S M ima e . eal h o ie ha ho ha ima e
o ra i S M a e ro l a e e ha e i i am le a 35].
Moreover, samples such as cultured cells can have large, micron-level height
changes across the cell body. For high-resolution NSOM information, the probe
must remain within nanometers of the sample surface during the measurements.
Therefore, without an active feedback mechanism, resolution is not only lost but
can change within an image, leading to difficulties in image interpretation.
While the problems associated with implementing a feedback scheme under aque-
o e iro me ere i l ol e i a m er o a roa he 5255], reduc-
i he or e i ol e ha ee more ro lema i . em o re e he or e
have largely revolved around using a feedback signal not based on force, reducing the
spring constant of the fiber-optic NSOM tip, or changing the probe itself.
o o a i er erome ri a roa h or o rolli he S M i am le a
was shown to be gentle enough to image the airliquid interface while in feedback
]. This approached took advantage of the differences in path length for photons
directly exiting the NSOM probe and those reflected off the sample surface as the
probe approached the surface. Detecting the interferometric signal from the side
resulted in oscillatory changes in the signal amplitude as the tip approached the
am le ]. The constructive and destructive interferences were observable when
the probe was approximately one micron from the sample surface, providing a truly
noncontact signal to implement a feedback scheme. While the technique was shown
 ear iel S a i i al Mi ro o e ool

to successfully provide feedback on dry samples in air or samples at the airliquid

interface, extending this arrangement into samples under aqueous conditions has
proven difficult. The collection of the interferometric signal from the side is not eas-
ily implemented for samples under buffered conditions, and other collection
geometries result in poor and unstable feedback signals.
ha i he ri o a o he er o i S M ro e a or e i i
new probes with low spring constants are the most straightforward ways of circum-
e i he or e ro lem. hemi all e hi o he iame er o he er o i
S M ro e i h ere lea o maller ri o a ]. However,
when used either in the shear-force approach with linear probe geometries or in the
tapping-mode arrangement with cantilevered NSOM probes, the smaller spring
constant leads to a dramatic dampening of the oscillation under aqueous conditions.
This reduces both the amplitude and stability of the feedback signal, which makes
live cell imaging difficult.
he ili o i ri e ro e e i a omi or e mi ro o M) ha e ri
constants that are amenable with live cell imaging. This has led to various attempts
o i or ora e a ear el li h o r e i o he e i 3]. In one application, a solid
immer io le a i or ora e i h a M a ile er ]. These probes were
manufactured for solid-state applications, however, and were not meant to break the
diffraction barrier or image biological samples. Thus, the spatial resolution was
limited to approximately 250nm, and the forces associated with these tips were not
appropriate for imaging fragile samples.
More re e e ea or o o le M ro e i h S M a li a io ha e
e o e io eam I ) mo i a io o o e io al M ro e ]. In
the most straightforward example, a small hole is drilled at the apex of the pyramid
ha a a he M l . hi hole a a he a er re or li h i S M a li-
cations. These probes have been used mostly in transmission measurements, from
which it is difficult to extract reliable measures of spatial resolution due to the pos-
sible presence of tip-sample coupling artifacts. More reliable spatial information
comes from NSOM fluorescence measurements. To date, fluorescence measure-
ments taken with these probes have not reported subdiffraction limited spatial reso-
lution; further experiments are needed to show that high resolution is possible with
these probes.
I rela e ie I mo i a io o M ro e ha le o he e elo -
ment of hybrid probes that incorporate elements that act as NSOM light sources.
I o e e am le a I a e o rill a m hole in the end of a conventional
M a ile er a he rami al l ormall e i M ima i a
removed, as shown in Fig. . . m high-index glass sphere was glued into the
hole a r her ma hi e i o a rami ha e i h he I . he i e ere he
coated with 50100nm of aluminum to confine the excitation light, and a small
a er re a o e e a he e he rami i h he I . hi ro e re l e i
a M i i or ora i a eleme ha a a a S M li h o r e.
imilar ro e a a ri a e i er o i S M ro e i or ora e i o
ili o i ri e M a ile er . ai i h he e o I e h olo a mall
hole i rille i o he e o a M ro e. al mi m oa e er o i
 N.E. Dickenson et al.

Fig. 9.20 erie o ima e e i i he e e i i or ora i a hi h re ra i e i e la

here i o a M a ile er. o e io eam I ) i r me i e o ar e a hole i o
he M a ile er a ho i a). m glass bead is then secured within the hole using
UV-curable glue (b). he I i e o are ll ar e he la ea i o a rami . he ro e
is then coated with a thin layer of aluminum to confine light. The completed probe is shown in (c)
and a magnified view of the waveguide is shown in (d) i re a a e rom e . ])

NSOM probe is carefully lowered into the machined hole and glued into place.
e l e he e e er i rom he a i e o he M i i h he I
as shown in Fig. . . he e i ha e he a a a e o i ell hara eri e
fiber-optic NSOM probes to deliver the light combined with the low spring c onstants
o ommer ial M a ile er .
Figure . shows NSOM topography (a) and fluorescence (b) images taken
with these probes on a sample containing small, fluorescent spheres embedded in a
matrix. The fluorescence image reveals small, subdiffraction limited features that
are not correlated with features in the topography image. This confirms the spatial
resolution capabilities of these hybrid probes. Of more interest, however, are the
NSOM fluorescence, force, and amplitude images shown in Fig. . . These
S M ima e ere a e o a li i SM ell er ere o i io . he
a re er i re e or ere ore e l la ele i h ra o i IP a
re ea e a o he ame ell o o e i e e o ama e a e he S M
Fig. 9.21 erie o ima e o li i he ma a re ro e or lo or e S M ro e .
ommer ial M a ile er a) i la e i i e a o e io eam I ) i r me a a
m-diameter hole is carved into the cantilever as shown in (b). er o i S M ro e i
inserted into the hole (c) and secured with UV-curable glue. The fiber is then cut flush with the
back of the cantilever (d) i re a a e rom e . ])

Fig. 9.22 m10-m S M ima e a e i a h ri M S M ro e li e ha

shown in Fig. . . (a) and (b) display the simultaneously collected NSOM force and fluores-
cence images, respectively. The sample consisted of 50-nm fluorescent latex spheres embedded in
a hi a e a e ma ri i re a a e rom e . ])
250 N.E. Dickenson et al.

Fig. 9.2330-m30-m NSOM force (a), deflection (b), and fluorescence (c) images of a live
human arterial smooth muscle cell. The hybrid probe shown in Fig. . was scanned in contact
feedback mode under buffered conditions. The fluorescence image results from the labeling of
a re er i re e or i h ra o i IP i re a a e rom e . ])

M h ri ro e. hile he e mea reme ill ra e ha S M mea re-

ments on living cells are possible, the laborious nature of probe fabrication and the
ee or e e i e I e h olo limi he a li a ili o hi a roa h. learl
more developments are needed to make routine NSOM measurements of living
tissues useful.

9.4 Summary

Pro re i i S M o ro e iolo i al am le ha ee lo e he
technical difficulties involved, but NSOM now seems poised to take its place among
other complementary techniques. The high-resolution fluorescence and force infor-
mation seems particularly well suited for biological applications where both con-
ra me ha i m a a i e i i h . om i e i h he i le mole le
detection limits and spectroscopic capabilities, this offers a powerful set of tools for
the biological sciences.

Acknowledgments e ra e ll a o le e he or o I GM )a he Ma i o
a ila Sel o a io . . . . a o le e he or rom he I ami e o
hemi al iolo rai i ro ram GM ).

References

. e . ei ra e r heorie e Mi ro o er Mi ro o i he ahr ehm . r h

Mi ro a. .
. or M ol . Pri i le o o i . h e . am ri e am ri e i er i Pre .
. o o e h . Pri i le o a o o i . e . e or am ri e i er i
Pre .
 ear iel S a i i al Mi ro o e ool 251

. Pohl . i a he a ome re ale. Philo ra So o o Ser Ma h Ph S i.

.
. . ear el a i o i al mi ro o . hem e . .
. o o Pohl e h .S a i ear el o i al ro e i h l ra mall o i e.
e . .
. e i ra ma arri ei er S o ela . rea i he i ra io arrier
o i al mi ro o o a a ome ri ale. S ie e. .
. i i M. ear el a i o i al mi ro o a ire all o iolo . ra .
.
. o ma P oi Sala he P. om ari o o me ha i all ra a ro e io la er
hemi all e he o i al er i . l rami ro o . .
. r er . .S. Pa e o. .
. ollar . al a io o hermal e a ora io o i io e i oa i al mi-
m o ear el er o i ro e . e S i I r m. .
. Moore ri o S i e o . e i ear el a i o i al
mi ro o or iolo i al ie . . .
. i e o ri o S Moore . hara eri a io o o er i e
hea i a ama e i er o i ro e or ear el a i o i al mi ro o . e S i
I r m. .
. e i i P ei er S. om i e hear or e a ear el a i o i al mi ro -
o . l Ph e . .
. Shalom S ie erma e i ohe S . mi ro i e e or e ro e i a le or ear el
a i o i al mi ro o . e S i I r m. .
. ole o ro a P he ae ira a i M. ear el i ere ial a i o i al
mi ro o e i h a omi or e re la io . l Ph e . .
. M rama hi a omma a a ima a a haS mi ihira M.
ear el o i al mi ro o i li i . l Ph e . .
. alle oo e G . i h re ol io ore e e ima i i h a ile ere
ear el er o i ro e . l Ph e . .
. a er Maa a a i i her . ami or e i a e o rol i e o ario
ro e or a i ear el o i al mi ro o . e S i I r m. .
. Pae ler M Mo er P . ear el o i heor i r me a io a a li a io . e .
e or oh ile .
. ri o S . Sam le hea i i ear el a i o i al mi ro o . l Ph
e . .
. e h Si il P e er e o i Mar i Pohl .S a i ear el
o i al mi ro o i h a er re ro e ame al a a li a io . hem Ph .
.
. o o a er . i h ro a a io i a li ri al a e i e i h a om le me alli
iele ri io . Ph e S a Ph Pla ma . .
. S aheli M o M arra h G Mei er ho e Gra a her I. em era re
ro le o er i e i a i ear el o i al mi ro o . l Ph e .
.
. oa a o o I alle . ri i a e e o hermal ro e e o ear el
o i al ro e . l Ph e . .
. ai iG o e e i m ro io lle ri i M a iel Sil a G iar i PG
Pa a e S ai ea a ialli . i al ro i o am le hea i i a i ear
el e erime i h a er re ro e . l Ph e . .
. ai iG o e e i m ro io lle ri i M G iar i PG Pa a e S a iel
Sil a ai ea a ialli . I e i a io o hea i e e i ear el e erime
i h l mi e e or a i emi o or . S h Me . .
. G iar i PG olo i M a ar i M lle ri i M. hermal e a io e e i ear el
o i al mi ro o er ro e i e la er li h a or io . l Ph e .
.
252 N.E. Dickenson et al.

. ie a i h er l ae er . i h i e e a io o er i i ear el a i
o i al mi ro o . l Ph e . .
. M o ell M. S r re a ra i io i li i mo ola er a he air a er i er a e.
e Ph hem. .
. Moh al . Pho holi i a ho holi i ro ei mo ola er a he air a er i er a e.
e Ph hem. .
. ollar . S mi ro r re i L-a-dipalmitoylphosphatidylcholine
monolayers and bilayers probed with confocal, atomic force, and near-field microscopy.
io h . .
. i i M. he a e o li i ra rom mo el mem ra e o ell . e io h iomol
Sr . .
. Simo a . Mo el em li i ra a ell mem ra e . e io h
iomol S r . .
. r o P a irio M oh o . o olor ear el ore e e mi ro o
ie o mi ro omai ra ) i mo el mem ra e . a m ir. .
. a rlo r o P oh o . he i e o li i ra a a omi or e mi ro o
o a lio i e GM omai i hi om eli P hole erol mem ra e . io h
. .
. S h ar M M S a lo ello . e o r a a o mor i i mor ali
a re o r e e i e or i a ei hi o . e l Me .
.
. Si a a . i h re ol io ie o l r a a olla e. Pho o hem
Pho o iol. .
. e i hi he er . Si le mole le o er e ear el a i o i al mi ro o .
Seie e .
. ollar . Pro i i le mole le orie a io i mo el li i mem ra e i h
ear el a i o i al mi ro o . hem Ph . .
. Ger e Gar ia Para o M o o eerma i er a l .I e i
he a lar emi io o a i le mole le. Ph e e . .
. Ger e Gar ia Para o M o o eerma i er a l . ear el
e e i i le mole le emi io . Mi ro or ). .
. are ri ma o i M. ima i . a io e h ol. .
. Sel i P . he re ai a e o ore e e re o a e e er ra er. a S r iol.
.
. Se a ii S . l ore e e re o a e e er ra er a i ear el o i al mi ro o .
Philo ra So o o Ser Ma h Ph S i. .
. Sh ei a G Se a ii S ie ler G e o ho S. o al ore e ro e or he ore -
e e re o a e e er ra er a i ear el o i al mi ro o . l Ph e .
.
. i er S .S a i ear el ore e e re o a e e er ra er mi ro o .
io h . .
. i er S . om i i Ma or hi h re ol io ore e e mi ro -
o . Mi ro or ). .
. i e o . li he ). li he a a.
. a Ghe er Mar oli i i M. omai i ell la ma mem ra e i e i a e
ear el a i o i al mi ro o . io h . .
. e a er I e a e am i or eri P a i M a l i or G
Gar ia Para o M . a o ale or a i a io o he a ho e re e or SIG ma e
i le mole le hi h re ol io ore e e mi ro o . hem h hem. .
. r o P Ia o l ao irio M oh o a lor S. ear el
scanning optical microscopy probes: a comparison of pulled and double-etched bent NSOM
ro e or ore e e ima i o iolo i al am le . Mi ro . .
. oo ma M e a er I Gar ia Para o M a l . Shear or e ima i o o
am le i li i i a i i ell o e . l Ph e . .
 ear iel S a i i al Mi ro o e ool 253

. e er h l e o Mihal ea S hol er er S. a ile er ro e or S M

a li a io i h i le a o le a er re i . l rami ro o . .
. alle ee M . Si le mole le e e io a er a er ore e e ima i
i h a ile ere ear el er o i ro e . l Ph e . .
. Shi . ear el a i o i al mi ro o ie o L-alpha-dipalmitoyl
phosphatidylcholine mo ola er a he air li i i er a e. Mi ro or ).
.
. Shi ro meier . o o a ear el a i o i al mi ro o ima i
i a i er erome ri o i al ee a me ha i m. a m ir. .
. ee M alle i er S ro meier ollar Shi . Pro re
o ar ima i iolo i al am le i h S M. Pro SPI I l So S a i or e
Mi ro iome l. .
. Par o i Peale i M M rra i G o i al i o o S
o aa har hi he er eh . i h o er la er li h o r e or ear el
o i a i a li a io o hi h e i o i al a a ora e. l Ph e .
.
. iom a a e ri ehrer re S l a h hl o . i h re ol io
constant-height imaging with apertured silicon cantilever probes. J Microsc (Oxford). 2001;
.
. er re la M Ger e ei elma S h rma G oell S a er e
ooi . ear el ore e e ima i ih m re ol io a e o mi ro a ri a e
a ile ere ro e . l Ph e . .
. ehrer re Pe er e S S l a h hl o iom a a e ri el .
Fabrication of silicon aperture probes for scanning near-field optical microscopy by focused
io eam a o ma hi i . Mi roele ro . .
. a oe i Gra ie P. om i e S M M mi ro o i h mi roma hi e a oa er-
re . Ma er S i Pola ). .
. ro meier . o e io eam mo i a io o a omi or e mi ro o i or
ear el a i o i al mi ro o . l Ph e . .
. a iai Moore i hol ar ell ro meier . ri ear el a -
i o i al mi ro o i or li e ell mea reme . l Ph e . .
Chapter 10
Atomic Force Microscopy: Applications
in theField of Biology

J.K. Heinrich Hoerber

Abstract The invention of telescopes and microscopes about 400years ago

revolutionized our perception of the world, extending our sense of seeing. Extending
it further and further has since been the driving force for major scientific develop-
ments. Local probe techniques extend our sense of touching into the micro- and
nanoworld and in this way provide complementary new insight into these worlds we
see with microscopic techniques. Furthermore, touching things is an essential
prerequisite to manipulating things, and the ability to feel and manipulate single
molecules and atoms for sure marks another of these revolutionizing steps in our
relation to the world in which we live.
Local probes are small-sized objects, such as the very end of sharp tips, which
interact with a sample, or better, the surface of a sample at selected positions.
Proximity to or contact with the sample is required to have a high spatial resolution.
This, in principle, is an old idea that appeared in the literature from time to time in
context with bringing a source of electromagnetic radiation in close contact with a
sample (Synge, Philos Mag 6:356, 1928; OKeefe, J Opt Soc 46:359, 1956; Ash and
Nicolls, Nature 237:510, 1972). It found no resonance and therefore was not pur-
sued until the early 1980s. Nanoscale local probes require atomically stable tips and
high-precision manipulation devices. The latter are based on mechanical deforma-
tions of spring-like structures by piezoelectric, electrostatic, or magnetic forces to
ensure continuous and reproducible displacements with precision down to the
picometer level. They also require very good vibration isolation. The resolution that
can be achieved with local probes is given mainly by the effective probe size, its
distance from the sample, and the distance dependence of the interaction between
the probes and the sample measured. The last can be considered creating an effec-
tive aperture by selecting a small feature of the overall geometry of the probe tip,
which then corresponds to the effective probe. One of the great advantages of local

J.K.H. Hoerber (*)

HH Wills Physics Laboratory, University of Bristol, Tyndall Ave., Bristol BS8 1TL, UK
e-mail: [email protected]

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 255

DOI 10.1007/978-3-319-01360-2_10, Springer International Publishing Switzerland 2014
256 J.K.H. Hoerber

probes is that they can work in any environment; this way, they provide the possibil-
ity to study live biological processes similar to optical microscopy, but at a resolu-
tion similar to electron microscopy (EM).

10.1 Introduction

The atomic force microscope (AFM) belongs to the large family of instruments
called local or scanning probe microscopes. All of these instruments are based on
the principle of measuring the interaction of a small tip structure with a sample
surface at close distances. In the case of the first instrument of the family, the scan-
ning tunneling microscope (STM), developed in the early 1980s by Binnig and
Rohrer [4] at the IBM Research Laboratory in Zurich, the interacting tip is a single
atom (see Fig.10.1). This is possible because of the very short-range electronic
interaction used with a decay length on the size of atoms. This interaction between
the tip and surface, which allows electrons to cross a small gap before an electrical
contact is made, is related to the quantum-mechanical process of electron tunneling
reflected by the name. Such a tunneling current through a vacuum gap is in the range
of pico- to nanoamps and drops by an order of magnitude if the distance changes
from 0.4 to 0.5nm. With atoms being the same size, it is possible that the main
contribution to this tunneling current is related to the interaction of the foremost tip
atom with a single atom at the surface closest to this tip atom. If the surface and the
tip are moved with respect to one another at a constant average distance, the varia-
tions seen in the current reflect the topography of the surface with atomic resolution
if the surface is homogeneous. In the case of different types of atoms present at the
surface, the signal measured is a mixture of topography and differences between the
atom species in their electronic interaction with the tip atom.
With this incredible resolution, the most important part of such a microscope is
the mechanism that causes the tip to approach the surface and moves it across with
the necessary precision and stability. The mechanism Binnig and Rohrer came up
with is based on the piezoelectric effect discovered in 1880 by the brothers Pierre

Fig. 10.1 STM imaging with

atomic resolution. The tip is
scanned at a constant height
over the surface, and the
current measured between the
tip and surface is modulated
by the distribution of surface
atoms
10 Atomic Force Microscopy: Applications in the Field of Biology 257

Fig. 10.2(a) An electric field applied across a quartz crystal interacts with the crystalline struc-
ture through its polarization, which affects the crystal structure and leads to a change in size.
(b)The piezotube changes its length if a voltage difference between the inside electrode and all
four outer electrodes is applied. The tube bends sideways if a voltage difference between the inside
electrode and one segment of the outside electrodes is applied. In this way, a tip mounted in the
direction of the z-coordinate can approach a surface very precisely and be scanned across with the
same precision, which is mainly related to the stability of the voltage source used

and Jacques Curie. They noticed that mechanical stress affects the polarization of
crystalline structures like quartz and that this can be used in reverse with an electric
field applied across a quartz crystal to change its size. In the case of quartz, a very
high electric field is necessary for only a small change in size. In the 1950s, ceramic
materials such as perovskites were discovered, which change their size in the
micrometer range with only a few hundred volts applied. Three stacks of ceramic
discs were used in the first STMs to position the tip in all three dimensions with the
necessary precision that allowed a precise approach of the tip to the surface and
scanning it across. Meanwhile, more often a ceramic tube with an electrode on the
inside and a segmented electrode on the outside is used, which allows tip movement
in all three directions in a controlled way by applying a voltage between the inner
electrode and all outside electrodes to approach and retract the tip and between the
inner and individual outer electrodes to bend the tube sideways to do the scanning
(Fig.10.2).
To prevent the tip from crashing into the surface while scanning, due to roughness
of the sample surface or when mechanical vibrations of the instruments occur, the
electronics that controls the movement of the tip includes a feedback loop that
limits the tipsurface interaction by moving the tip away from the surface if necessary
to reduce the current below an upper limit. The feedback can keep the interaction
constant during scanning if it is fast enough, which means the tip will follow the
contour of a sample with homogeneous electronic properties. This imaging mode is
called constant current imaging. The maximum speed of the feedback control,
mainly determined by mechanical resonances, is the main limitation in scanning
probe microscopy that determines how fast an image (picture) of an area with a
certain size can be produced. Faster imaging can be done in so-called constant-height
mode, but this is only possible on sample surfaces that are flat at the atomic scale to
avoid the tips crashing into the sample at protrusions.
258 J.K.H. Hoerber

10.2 Atomic Force Microscopy

Due to the electronic interaction measured in the case of the STM, both the tip and
the surface have to be conducting. This significantly limits the materials that can be
investigated. In 1986, Binnig, together with Quate and Gerber, overcame this limita-
tion with the invention of the next member of the probe microscope family, called
the atomic force microscope (AFM) [5]. The idea that allows also the investigation
of nonconducting materials was to use the van der Waals forces between atoms as
a type of interaction present in all types of atoms. The corresponding interaction
potential (LennardJones potential) has a very weak attractive regime between 0.7
and .35nm and becomes strongly repulsive with a 1/r12 distance dependence below
0.35nm. A microscope using these interactions between the tip and the surface
atoms needs to detect forces in the nano- to pico-Newton range, thus asking for a
very sensitive force-detection system. The most straightforward way to measure
forces is by using a spring, which can be characterized by its spring constant con-
necting the spring extension (or bending in the case of a cantilever) and force by
Hookes law: F=kx. With such a linear forcedistance relationship, a spring with
a stiffness of 1N/m needs a precision in the distance measurement in the nanometer
range to achieve a force resolution in the nano-Newton range. The STM, invented
just five years before by Binnig and collaborators, can easily measure distances with
such a precision; therefore, it was the obvious choice to be used in the first AFM to
detect the bending of a small cantilever with a tiny piece of diamond as a tip at the
end. A schematic of the first AFM instrument is provided in Fig.10.3.
Initially, there was doubt about whether the AFM could achieve a true atomic
resolution and whether the van der Waals interaction of one tip atom with just one
surface atom on crystalline surfaces is actually measured. Some researchers sug-
gested that clusters of atoms on both sides interact with one another and that the
structures seen are interference patterns created while scanning across the surface
that have more or less a relationship to its atomic structure. It was much later that
Giessibl could show that under certain conditions, even atomic substructures can be
imaged with an AFM [7]. However, the normal case, especially on molecular

a b 25 m
B
.25 mm
SCANNERS, A C FEEDBACK F
F FEEDBACK 1cm DIAMOND
D STM
AFM E TIP
.8 mm
x
LEVER
BLOCK (ALUMINUM) z (Au-FOIL)
y

Fig. 10.3 Schematic of the first AFM from the IBM patent application in 1986 (Adapted from
Ref. [6]). In (a) the cantilever is not drawn to scale; the dimensions are shown in (b). A: sample, B:
AFM diamond tip, C: STM tip, D: metal cantilever as sample for the STM detection, E: modulat-
ing piezo, F: viton vibration insulation
10 Atomic Force Microscopy: Applications in the Field of Biology 259

Fig. 10.4 Optical detection scheme showing the laser beam deflected by the cantilever toward the
quadrant photo detector

structures, turned out to be that more than one atom of the tip interacts due to
various types of surface forces with a surface area of a corresponding size and in
this way reduces the AFM resolution to the nanometer range.
An important development providing the basis for the successful use of the AFM
in biology was the replacement of the STM detection with an optical detection
scheme (Fig.10.4), as this allowed measurements even in salt-containing solutions.
In 1988, Meyer and Amer demonstrated that a reflected laser beam at the tip of a
cantilever observed in some distance with a split photodiode can provide a sensitiv-
ity similar to that of the STM detection [8]. The standard detection scheme in most
commercial AFMs is based meanwhile on a diode laser focused at the end of the
cantilever directly above its tip. The bending of the cantilever changes the direction
of the reflected laser beam and thus its position on the detector. The change in the
intensity distribution between the two halves of the split diode leads to a change in
the electric signals produced by both halves. If a sensitive amplifier is used to mea-
sure the signal difference between the two halves, it is possible to detect movements
of the cantilever in the range below 1nm, which, with a small cantilever spring
constant of 0.01 N/m, leads to a force sensitivity in the range of 10 pN.
The cantilevers used are micro-machined from silicon or silicon nitride wafers
with a length between 50 and 250m and a thickness between 0.5 and 2m, leading
260 J.K.H. Hoerber

Fig. 10.5 EM image of an AFM cantilever from below, with the tip visible at the end as a small
5-m square

to spring constants between 100 and 0.01/m (Fig.10.5). The tips are made by
different procedures and in some cases have a roof-shaped end due to the crystalline
structure of the material used. In such a case, the resolution depends not only on the
length of the ridge, but also on its tilt and the scanning direction with respect to the
direction of the ridge. A so-called oxide-sharpening technique, which leads to final
tip radii of only a few nanometers at a total tip length of 510m, makes the small-
est tip radii. The shape of cantilevers quite often is triangular, to provide a higher
stability against lateral forces, which can become quite high due to friction effects
while scanning the tip across a surface. These lateral or friction forces twist the
cantilever sideward and lead to a movement of the laser beam on the detector per-
pendicular to the normal deflection, which can be measured by using a quadrant
photodiode instead of a split one (Fig.10.4).
With a device that can detect the bending of the cantilever, it becomes difficult to
move the tip, which is usually done in an STM. Therefore, many AFMs have the sam-
ple mounted on the piezotube. As the sample normally has a larger mass than the
cantilever, the systems resonance frequency decreases in such a setup, decreasing
the scan speed. However, with a specially designed laser optic, the problem can be
overcome and a scanning head can be made that can be placed over a fixed sample.
Nevertheless, usual imaging speeds are in the range of minutes; the efforts to build
fast-scanning AFMs are still ongoing, mainly based on reducing the cantilever size.

10.2.1 Application Modes

As in the case of an STM, it is important to control the tip interaction while scanning
with the AFM in order not to destroy the tip or sample due to surface roughness or
10 Atomic Force Microscopy: Applications in the Field of Biology 261

external vibrations. The electronic feedback in the case of an AFM tries to keep
the force between the tip and the sample within limits. This way, if the reaction time
of the feedback is fast compared to the scanning speed, the tip will follow the
surface contour, applying the same force to the sample everywhere when working
in the repulsive van der Waals regime. This mode is called constant-force mode.
The feedback normally is created as fast as the system resonances allow when
weighted linear, differential, and integral components of the signal are being used.
Nevertheless, over slopes, the feedback with reasonable scan speeds always becomes
too slow and characteristic image artifacts occur depending on whether the tip is not
retracted or pushed down fast enough. For very flat samples, a constant-height mode
is also possible, with forces between the tip and the sample changing during scan-
ning, as higher sample structures will push up the cantilever tip in this case. The tip
is kept in contact due to an average preadjusted loading force. In this mode, the
cantilevers spring constant and the corresponding resonance frequency determine
the reaction time and, therefore, the maximal scan speed:

k
F~ ,
m

where F is the resonance frequency, k is the spring constant, and m is the mass of the
cantilever.
If a tip with a certain interaction force is scanned over a sample, lateral forces act
on the tip due to the friction between the tip and the sample. Measuring various
lateral friction forces is a way to characterize inhomogeneous samples, where dif-
ferent components have different frictions. For such friction-mode measurements,
the AFM has to be equipped with a quadrant photodiode instead of a split one, as
mentioned earlier. If adjusted in the right way, the quadrant diode detects in one
direction the normal bending of the cantilever and perpendicular its sideward twist-
ing due to the lateral friction force. Friction and high forces on rough surfaces,
together with a slow feedback, easily destroy soft samples, resulting very often in
parts of the sample sticking to the tip. A way to reduce this risk is to use the AFM
in a dynamic mode. In this mode, the cantilever is vibrating close to its natural
resonance frequency driven by a small piezo mounted where the cantilever is fixed.
With small oscillation amplitudes in the ngstrom range in the dynamic AFM mode,
force gradients are measured, allowing measurements even in the narrow and shallow
attractive range of the van der Waals forces.
Some instruments drive the cantilever oscillation with a magnetic coating above
the tip, which allows for much better control of the tip motion in this dynamic mode,
especially if the cantilever works in a fluid chamber. If a piezo is used, it has to
be mounted outside the fluid chamber to avoid electrical shorting. In this case, the
vibrating piezo excites acoustic waves, which drive the cantilever in a more indirect
way that makes the control of the actual tip motion difficult.
Working in air under ambient conditions has a serious issue with water, as all
materials have thin layers of water present on the outer surface, with the water
thickness depending on the humidity, the temperature, as well as the actual material.
During the approach of the cantilever tip to the surface, the water layer on the tip
262 J.K.H. Hoerber

Fig. 10.6 Principle of a dynamic AFM that uses an oscillating cantilever to measure tip interac-
tions by amplitude and phase changes of the oscillation driven with a feedback using a phase-
shifted signal to enhance the cantilevers quality factor, which is significantly reduced in liquid due
to the viscosity

and that on the surface coalesce at one point during the approach and form a water
meniscus. The forces generated by this meniscus are in the micro-Newton range;
scanning this meniscus across a surface will wipe away everything that does not
adhere well to the surface. With high enough amplitudes of the cantilever oscilla-
tions, in the so-called tapping mode, the water meniscus between the tip and the
surface will break during each retraction cycle and no lateral forces of the meniscus
will be produced.
In dynamic mode, the amplitude and phase changes of the vibrating cantilever
reflect elastic and inelastic tipsample interactions and can be used to characterize
these material properties at the nanometer scale (Fig.10.6). Images of inhomoge-
neous samples created using either the amplitude or the phase changes nicely depict
distributions of different materials. Unfortunately, with the cantilever immersed into a
liquid, the viscosity leads to a strong damping of the cantilever oscillations and a sig-
nificant broadening of its resonance frequency curve, reducing the resolution in both
the amplitude and the phase-shift signal. With a special type of feedback using a 90
phase-shifted input signal, it is possible to restore to some extend the so called quality
factor of the cantilever, which was greatly reduced by the damping of the liquid.

10.2.2 Working in Liquid

The already-mentioned possibility to work with an AFM in liquid and especially in

salt solutions extends the possibility of light microscopy to study the time course
of processes from the micro- into the nanometer range. For biological applications,
this opened a new era of studies on cellular structures and their dynamics. But as
mentioned earlier, the cantilever motion in liquids is damped by the liquids viscos-
ity, and with the maximum speed the cantilever can be moved, the maximal imaging
speed for high-resolution imaging is reduced. The viscous drag directly affects the
mechanical properties of the cantilever immersed into a medium. How the
10 Atomic Force Microscopy: Applications in the Field of Biology 263

Fig. 10.7 Changes to the frequency spectrum of a cantilever were measured while the ratio of the
concentration of glycol to water is increased from 20% (a) to 80% (d) in 20% increments

mechanical properties of a cantilever change with increasing viscosity is shown in

Fig.10.7, where the glycol concentration in water is increased systematically, lead-
ing to a significant increase in viscosity. The resonance curve is substantially broad-
ened, and its maximum shifted from nearly 24kHz at a 20% glycol concentration
down to 7kHz at an 80% concentration. These resonance curves were recorded as
the difference between a cantilever moving in liquid driven by thermal fluctuations
and the same cantilever in contact with the liquid chamber wall. This measuring
protocol enables one to extract the instruments mechanical and electrical noise
(from the signal) to a level where the resonance curve excited by thermal fluctua-
tions even in a very viscous medium can be measured.
The response of a mechanical cantilever to time-dependent external forces is
described by a differential equation that characterizes the cantilever by its mass and
its spring constant and the surrounding medium by its viscosity , leading to a
velocity-dependent damping :

F ( t ) = m d 2 z ( t ) / d t 2 + g d z ( t ) / d t + kz ( t ) ,

F = time-dependent driving force,
m = moving mass,
= velocity-dependent damping,
k = spring constant.
With this equation, we can derive the maximum of the resonance curve of the
cantilever as

k g2
fmax = .
m 2m 2

264 J.K.H. Hoerber

Fig. 10.8 Schematic of the

different mass components
necessary to describe the
motion of a cantilever in a
viscous medium

Fig. 10.9 Fit to the maximum values of the resonance curves at different mixtures of water and
glycol, which allows the values of m1 and m2 to be determined

To describe the way the resonance maximum changes with changing viscosity,
as depicted in Fig.10.8, the viscosity and other parameters have to change to
produce the monotonous decrease in the resonance frequency with increasing vis-
cosity. There is no reason to assume that the spring constant of the cantilever may
change, as it is only related to the material properties of the cantilever itself, which
do not change. Another possibility is to assume that some mass of the medium is
added to the cantilever mass. When a cantilever moves in a medium, some of the
medium moves with the cantilever and some of the medium flows around these
moving components to make way in front and to fill the space left behind.
Therefore, to account for the hydrodynamic effects, we need to add two addi-
tional components of mass into the equation (see Fig.10.9):

F ( t ) = ( m + m1 + m2 ) d 2 z ( t ) / d t 2 + g d z ( t ) / d t + Kz ( t ) ,

m=cantilever mass,
m1=fluid mass moving with the cantilever,
m2=fluid mass moving in the opposite direction.
With this equation, we can fit the observed changes of fmax and calculate the
different mass components. Assuming that the shape of the cantilever is causing
10 Atomic Force Microscopy: Applications in the Field of Biology 265

only secondary effects, we can approximate the actual rectangular cantilever with a
cylinder of the same mass and length, which allows us to solve the differential equation.
The results show that in pure water twice the mass of the cantilever moves with the
cantilever, and a little bit less than half the mass moves in the opposite direction.
With an addition of 80% glycol, and the corresponding high viscosity, 140 times
the mass of the cantilever is moved together with it, but only 20% more than in pure
water is moving in the opposite direction. In the case of a cantilever with a diameter
of 6m, the layer of liquid moving with it changes from 2.4m in pure water to
80m if 80% glycol is added. In the latter case with 80% glycol, the layer thickness
of the liquid moving against the cantilever is 200nm, and in pure water it is 10nm.
This calculation shows that in water the hydrodynamic effects are limited to a range
below 3m around the cantilever. With an actual tip length of 5m for most
commercial cantilevers, this will not influence the sample interaction.

10.3 Applications in Biological Systems

Interactions of molecular structures are traditionally characterized by binding

constants, on and off rates, and corresponding binding energies. The relevant ener-
gies for single binding events range from thermal energy up to some hundred times
that of the thermal energy when covalent bonding is involved. Besides covalent
bonds, hydrogen bonds, with energies between 4 and 16 kT, play an important role
for macromolecular structures and their spatial organization. At interfaces of these
structures, coulomb and dipole forces determine the interaction with the aqueous
environment and thus between molecules. An important question for the biological
function of these molecular interactions is their distance dependence in the actual
environment. This question can be addressed using AFM force spectroscopy on the
single-molecule level, giving a new, more physical view of the molecular interac-
tions in terms of forces and their distance dependence.
An important aspect of measuring forces at the molecular level is the dependence
of the force value on the timescale in which the force is applied. This connection can
be understood by the deformation to the molecular interaction potentials caused by
an applied force. This deformation leads to an effective lowering of the binding
energy if the force tries to pull the molecular structures apart. Consider thermal
fluctuations with a pulling force applied: It becomes more likely that the bond will
break during a certain observation time. The off rate for the biotinavidin binding,
for instance, at room temperature is on the order of six months. If a small force of
about 80 pN is applied, the lowering of the binding potential reduces the off rate to
about 9s. Doubling this force further decreases the lifetime of the binding by three
orders of magnitude. Unfortunately, these experiments cannot be modeled with
molecular dynamic computer simulations, as the timeframe of these simulations
with the available computer power is limited to the pico- up to the nanosecond time
range. Simulations of the rupture forces of biotinavidin at this fast timescale lead
to calculated forces of 600 pN. For AFM measurements completed at the time scale
of 100ms, the actual measured forces are 100200pN [9].
266 J.K.H. Hoerber

10.3.1 Intermolecular Forces

Protein adsorption is a very important topic for many biomedical and biotechno-
logical applications. For instance, many chromatographic separations, such as
hydrophobic, displacement, and ion-exchange chromatography, are based on dif-
ferences in the binding affinities of proteins to surfaces. In addition, in vitro cell
cultures require cell-surface adhesion, which is mediated by a sublayer of adsorbed
proteins. The molecular basis of cell adhesion is, in general, an important process in
tissue development, tissue engineering, and tissue tolerance to implants.
Protein adsorption to a surface is a net result of various complex interactions
between and within all components, including the chemistry and topology of the
solid surface, the protein itself, and the medium in which it takes place. The interac-
tion forces involved include dipole and induced dipole moments, hydrogen bond
formation, and electrostatic forces. All inter- and intramolecular forces contribute to
a decrease in the Gibbs free energy during adsorption, defining the binding energy.
An important question for protein-adsorption processes is the reversibility of the
adsorption process. One approach to this question is an analysis of the time course
of adsorption. As the adsorption is a multiple-step process in most cases, this relates
to the more specific question of until which step is the process reversible? The most
common way to quantify adsorption is by using the adsorption isotherm, character-
izing at a constant temperature the number of molecules adsorbed in relation to the
steady-state concentration of the same molecules in the bulk solution. Adsorption
isotherms provide a convenient method for determining whether or not an adsorp-
tion process is reversible. Reversibility is commonly observed with the adsorption
of small molecules on solids, but only rarely in the case of more complex random
coil polymers.
Proteins are polymers, but globular proteins do not have a random coil structure.
The native state of these proteins in aqueous solution is highly ordered; most of the
polypeptide backbone has little or no rotational freedom. Therefore, significant
denaturing processes have to occur to form numerous contacts with any surface.
Structural rearrangements may occur in a way that the internal stability of globular
proteins prevents them from completely unfolding on a surface into a loose loop
and tail type of structure. Thus, a subtle balance between intermolecular and intra-
molecular forces determines the number of protein-segment to surface contacts
formed at steady state.
The thermodynamic description of the protein-adsorption process is based on the
laws of irreversible thermodynamics [10]. The process is strongly time-dependent;
some of the involved steps of molecular rearrangement are remarkably slow and
probably lead to significant binding of proteins only on the timescale of seconds.
With different timescales for the various interactions, the adsorption process can be
divided into fast stepswhich can be reversibleand slow steps, where protein
structure rearrangements can occur controlled by the adhering surface and the rest
of the environment. The latter processes are irreversible in most cases.
With the AFM, it is possible to measure the adhesion forces established by single
proteins, such as protein A and tubulin molecules, within contact times of
10 Atomic Force Microscopy: Applications in the Field of Biology 267

milliseconds to seconds. Protein A and tubulin are both globular proteins and can be
seen as examples of different types of protein binding. Protein A is often used to
bind antibodies to solid substrates, and tubulin forms the filament structure of
microtubules as part of the cells cytoskeleton. The molecules can be attached to the
cantilever tip and then brought into contact with different solid surfaces, which
might be covered with layers of other molecular structures. To study implant com-
patibilities, bare metal surfaces such as gold or titanium are of interest. In addition,
the optically transparent indium tin-oxide (ITO) is of relevance for the development
of interfaces between biological molecules and electro-optic devices.
The procedure for measuring adhesion forces between proteins and surfaces using
an AFM needs to define the approach speed, maximal applied force, contact time,
retraction speed, retraction distance, and time away from the surface (see Fig.10.10).
With these parameters fixed, one can measure the interaction force by approaching
gold, titanium, or ITO surfaces, for instance, with protein-coated tips. It turns out that
for the first contact, there is a specific interaction characteristic for the different mol-
ecules and the different metals [11]. With the observed reproducibility of a certain
value for an adhesion force within a series of measurements, it has to be assumed that
in these experiments a well-defined interaction between a specific amino acid group
located at the surface of the protein and the metal determines the first step of the
adhesion process. The experiments performed by Eckert et al. [11] demonstrated
that this technique, combined with an adequate preparation procedure, can be used
to measure these first steps in the protein-adhesion process at the single-molecule
level. Furthermore, these experiments demonstrated that the technique can be used to
study the dependence of these interactions on the environment.

10.3.2 Intramolecular Forces

The modular structure of proteins in natural fibers and the cytoskeleton seems to be a
general strategy for resistance against mechanical stress. One of the most abundant
modular proteins in the cytoskeleton is spectrin. In erythrocytes, spectrin molecules
are part of a two-dimensional network that provides the red blood cells with their
special elastic features. The basic constituent of spectrin subunits is a structural repeat,
which has 106 amino acids forming three antiparallel -helices with a left-handed
coiled-coil structure. Helical linkers of 1020 amino acids most likely connect the
repeats. The mechanical properties of several modular proteins have been investi-
gated by AFM so far. The first study was done on titin [12], which exhibits a -sheet
secondary structure. The experiments have demonstrated that the elongation events
observed during stretching of single proteins may be attributed to the unfolding of
individual domains; experiments with optical tweezers have corroborated these
results [13, 14]. These studies suggest that single domains unfold one at a time in an
all-or-none fashion when subjected to directional mechanical stress.
A major technological step forward in protein-folding studies with the AFM was
the development of a double-tip detection system. The concept uses one tip that
268 J.K.H. Hoerber

Fig. 10.10 AFM approach and retraction scheme for molecular adhesion measurements. The image
sequence to the left shows a cantilever-tip approach/retract cycle with a molecule linked to the
tip by a molecular tether. The top right image depicts the changes during the cycle in the vertical
position of the cantilever base; it is bending, which corresponds to the force acting on the cantile-
ver tip. The bottom right image shows the movement of the cantilever tip during the approach.
The jump indicates an adhesion force pulling the cantilever toward the surface due to attractive van
der Waals forces. The following linear change corresponds to the cantilever response while pushed
against the surface; its slope reflects the cantilever spring constant

stays in contact with the surface during the force and distance measurements, to
stabilize the distance between the second force-measuring tip and the surface.
This distance normally has drift speeds of several nanometers per second in AFM
due to the long mechanical connection between the cantilever and the sample holder.
The double-tip scheme (Fig.10.11) allows long-term stabilization of this distance
with a fraction-of-an-ngstrom precision over minutes and makes so-called force-
clamp measurements over such timescales possible. With this technique, it became
possible to perform unfolding and refolding experiments on extremely small protein
structures such as spectrin, with only four repeats [15]. In these experiments
(Fig. 10.12), the proteins were attached to clean, freshly prepared gold surfaces.
While the clean tip of the second force-measuring cantilever approaches such a
surface, a single molecule can become attached to its tip, and forces in the range of
pico to nano-Newtons can be applied to the molecular structure. With forces of
about 80pN, the molecules are stretched at a speed of 3nm/s to more than 10 times
their folded length, reaching almost the total contour length. The forceextension
10 Atomic Force Microscopy: Applications in the Field of Biology 269

Fig. 10.11 Schematic of double-cantilever AFM with one lever stabilizing the other cantilevers
tip and surface distance, which allows it to be used exclusively measure forces

curves show characteristic sawtooth-like patterns. The reaction coordinate of unfolding

is imposed by the direction of pulling, and unfolding events occurring in a single
protein can be studied this way. Each peak in the bottom trace of Fig.10.12 can be
attributed to a breakage of a main stabilizing connection of a folded structure.
Recombinant DNA techniques were used in the experiments to construct chains
of repeats from identical spectrin domains. The method extends the monomer (R16)
at both ends, so that the polymeric protein product contains a 13-residue linker
between the consecutive R16 units, with two cysteine residues at the C-terminal
270 J.K.H. Hoerber

Fig. 10.12 Stretching spectrin repeats between a gold surface and a cantilever tip

end. The forceextension relationship was measured on a four-repeat construct

fixed to the gold surface with the cysteine residues introduced at the C-terminal end.
In the sawtooth-like pattern of the forceextension curves (Fig.10.12, bottom), each
peak represents an unfolding event of one subunit. After each peak, the force drops
back as additional length of the protein chain unfolds, is made available, and starts
to extend. The maximal extension in these experiments is variable because the poly-
mer is not always picked up at the end, and sometimes the tip can be attached to the
surface at multiple locations. Nevertheless, the observed extensions never exceeded
the contour length of four repeats.
The so-called worm-like chain (WLC) model describes the measured force
extension curves well. It predicts the isothermal restoring forces of a flexible
10 Atomic Force Microscopy: Applications in the Field of Biology 271

Fig. 10.13 Distribution of

elongation length observed
by stretching the -helical
spectrin with the AFM. The
left peak relates to partial
unfolding, as shown in
Fig.10.14, and the right peak
relates to the total unfolding
of the structure

polymer, acting as an entropic spring during extension. It also defines a persistence

length to the polymer chain, which turns out to be 0.6nm for the spectrin chain.
Two consecutive force peaks are spaced by a distance that represents the gain in
length produced by a single unfolding event. These distances were measured for
several hundred unfolding events. The histogram (Fig.10.13) of the measured dis-
tances reveals a length distribution with two statistically relevant peaks. When the
histogram is fitted with two Gaussian distributions, the maxima are at 15.5 and
31nm. The force distributions associated with the short and long elongation events
can also clearly be separated. The mean force for the 15.5- and 31-nm elongation
events are 60 and 80pN, respectively, at a pulling speed of 3nm/ms.
These experiments demonstrated for the first time at the single-protein level that
-helical domains can unfold in several discrete steps in a much more complex way
than previously suggested by molecular dynamic simulations. The appearance of
different elongation events suggests that at least one intermediate state for the
spectrin repeat exists between the folded state and the completely unfolded state
(see Fig.10.14). In thermodynamic studies, such an intermediate state has not been
found. A statistical analysis of the AFM force curves shows that the short elongation
events are half the length of the total unfolding length of one domain. The molecular
pathway of unfolding of a spectrin domain can be simulated based on the AFM
measurements and shows that the first steps involve a partial opening of the bundle
and the loss of the secondary structure at the repeats connecting helices. The central
part reorganizes around a hydrophobic core resembling a shorter version of the
repeat structure.
Protein-folding and -unfolding pathways are very complex. However, by apply-
ing a force, the unfolding pathways can become very strongly restricted. The results
with only two possible unfolding pathways can be explained along this line,
272 J.K.H. Hoerber

Fig. 10.14 Normal spectrin

repeat folding of the three
-helices forming a slightly
twisted N in space, with
connection possibilities to
other repeats at both ends.
Below the stable intermediate
structure is shown with the
connecting a-helices
extended, but with the
structure around the
hydrophobic core still stable

Fig. 10.15 Energy scheme for the pathway leading to an intermediate state, and the energy scheme
that shows how the process leads directly to the completely unfolded state when the force is
applied faster and the energy landscape is further deformed

as outlined in Fig.10.15. One leads directly from the folded to the unfolded state,
the other first to an intermediate state. The intermediate state is conceptually avail-
able along both pathways, but each pathway by itself leads from the native folded
state 0 either to the partially unfolded state 1 or to the completely unfolded state 2.
The relative difference in the height of the free energy barrier along either of the two
pathways determines whether state 1 or state 2 is attained. The advantage of such a
modeling approach is that only one free parameter is needed for differentiating
between the two pathways.
In the native folded state, the protein is in state 0 at the bottom of the potential
well. The mechanical stress applied by the AFM tip not only decreases the barrier
height for thermally activated unfolding but also reduces the options of the protein
to follow either path 1 or 2 during unfolding (Fig.10.15). A certain protein will fol-
low only one path, leading to the observed bimodal probability distribution with
35% and 65% probability for path 1 and 2, respectively.
10 Atomic Force Microscopy: Applications in the Field of Biology 273

For such a simple model, a Monte Carlo simulation can be used to test the reaction
kinetics simultaneously for the short and long elongation events, which means paths
1 and 2, respectively. The kinetics can be characterized by two parameters: the
width of the first barrier, and an effective so-called attempt frequency related to the
barrier height as an exponential multiplication factor normalized by the thermal
energy. In such a simulation, there is no reason to assume a change in the width of
the first barrier, but the attempt frequency should vary and needs to be adjusted to
agree with the relative difference in barrier height. The force and elongation his-
tograms obtained from 5,000 simulations reproduce the observed general features
well, with a barrier width of 0.4nm and a difference in the barrier height between
both pathways of twice the thermal energy.
The experiments shed new light on the mechanical behavior of spectrin,
which is an essential cytoskeletal protein with unique elastic properties. They
demonstrated that the unfolding of spectrin repeats occurs in a stepwise fashion
during stretching with stable intermediates. The stable stretched states under the
applied forces show, on the one hand, that protein structures are determined by
the environment and, on the other hand, how the elastic properties of spectrin
come about. Actually, spectrin behaves within a certain force range as a quan-
tized but perfectly elastic structure: Applied forces stretch the molecule to a
defined length; after the forces are turned off, it goes back to the original state.
These details about the unfolding of single domains revealed by precise AFM
measurements show that force spectroscopy can be used to determine forces that
stabilize protein structures and also to analyze the energy landscape and the
transition probabilities between different conformational states.

10.3.3 AFM and Optical Microscopy

From the very beginning, the AFM was thought to be an ideal tool for biological
research. Imaging living cells under physiological conditions and studying dynamic
processes at the plasma membrane were envisioned although it was clear that such
experiments are quite difficult, as the AFM cantilever is by far more rigid than cel-
lular structures. Therefore, the way cells are supported becomes quite important.
In addition, a parallel optical observation is necessary to control the cantilever-tip
approach to distinct cellular features. In order to address these problems, the IBM
Physics Group in Munich started the development of a special AFM built into an
inverted optical microscope [16]. This instrument finally provided the first repro-
ducible images of the outer membrane of a living cell held by a pipette in its normal
growth medium. A conventional piezotube scanner moved this pipette. The detec-
tion system for the cantilever movement was, in principle, the normal optical-
detection scheme, but a glass fiber was used as a light source very close to the
cantilever to avoid optical distortions of the light while going through the water
surface. This setup allowed a very fast scanning speed of up to a picture per second
for imaging cells in constant-height mode. This is possible because the parts
274 J.K.H. Hoerber

moving in the liquid are very small andquite different from the imaging of cells
adsorbed to a microscope slideno significant excitation of disturbing waves or
convection in the liquid occurs. It is possible to keep the cell alive and well for days
within such a liquid cell. This makes studies of live activities and kinematics possible
in addition to the application of other cell physiologymeasuring techniques. With
this step in the development of scanning probe instruments, the capability of optical
microscopy to investigate the dynamics of biological processes of cell membranes
under physiological conditions was extended into the nanometer range. With imag-
ing rates of up to one frame per second, structures as small as 10nm could be
resolved and their dynamics studied. This made it possible to visualize processes
such as the binding of labeled antibodies, endo- and exocytosis, pore formation, and
the dynamics of surface and membrane cytoskeletal structures in general.
With the cantilevers integrated tip, forces in the range from pico- to nano-
Newtons are applied to the investigated cell; the mechanical properties of cell sur-
face structures therefore dominate the imaging process. On the one hand, this fact
mixes topographic and elastic properties of the sample in the images; on the other
hand, it provides additional information about the underlying membranes cytoskel-
eton and its dynamics in various situations during the life of a cell. To separate the
elastic and topographic properties, we need further information. The additional
information can be provided by using AFM modulation techniques. The pipette
AFM concept is well suited for such modulation measurements, as the convection
or excitation of waves in the solution caused by movement of the pipette is negli-
gible and the modulation can be done very quickly. Nevertheless, for a thorough
analysis of a cell wall elasticity map, we would have to record pixel by pixel a
complete frequency spectrum of the cantilever response to derive image data in
various frequency regimes. This would take too long for a highly dynamic system
like a living cell; therefore, such analyses are restricted to certain small areas.
Experiments performed on living cells with this setup showed in some cases a
certain weak mechanical resonance in the regime of several kHz. In 2007, Cross and
coworkers rediscovered this [17] and used it to characterize cancer cells. It might
lead to new methods of medical diagnosis at the cellular level and to new techniques
in drug development.
Figure10.16 shows the schematic arrangement of the first AFM set up to study
live cells. It was built on a highly stable sample holder stage within an inverted opti-
cal microscope. The sample area can be observed from below through a planar
surface defined by a glass plate with a magnification of 6001200 and from
above by a stereo microscope with a magnification of 40200. The illumination
is from the top through the less well-defined surface of the aqueous solution. In
order not to block the illumination, the manipulator for the optical fiber and for the
micropipette point toward the focal plane at a 45 angle. The lever is mounted in a
fixed position within the liquid slightly above the glass plate tilted also by 45.
A single-mode optical fiber is used to avoid beam distortions at the water surface
by bringing the end as close as possible to the lever within the liquid. To do so
requires removing the last several millimeters of the fibers protective jacket. The
minimum distance is determined by the diameter of the fiber cladding and the
10 Atomic Force Microscopy: Applications in the Field of Biology 275

Fig. 10.16 Schematic of the first AFM built in1988 to image live cells under physiological condi-
tions [16]

geometry of the lever. In the original experiments, fibers for 633-nm light with a
nominal cladding diameter of 125m and levers with a length of 200m have been
used. Holding the fiber at a 45 angle with respect to the lever means that the fiber
core can be brought as close as 150m to the lever. A 4-m-diameter core has a
numerical aperture of 0.1, and the light emerging from the fiber therefore expands
with an apex angle of 6. For the geometry given above, the smallest spot size
achievable is 50m, approximately the size of the triangular region at the end of the
cantilever. Due to the construction of the mechanical pieces holding the pipette and
the fiber, the closest possible distance of the position-sensitive quadrant photodetec-
tor to the lever is about 2cm. This implies a minimum spot size of 2.1mm, within
the 3-mm3-mm borders of the detector. Using a 2-mW HeNe laser under normal
operational conditions produces a displacement sensitivity of better than 0.01nm,
with a signal-to-noise ratio of 10, which compares well with the sensitivity of other
optical detection techniques. The advantage of this method is that there are no lenses
and there is no airliquid or airsolid interface for the incoming light beam to cross.
Only the outgoing light crosses the waterair interface, as the detector cannot be
immersed in the water, but this does not influence the information about the changes
in the cantilevers bending.
The pipette used for holding the cell in the setup was made from a 0.8-mm boro-
silicate glass capillary, which was pulled to about a 24-m opening in three pulling
steps. It is mounted on the piezotube scanner and coupled to a fine and flexible
Teflon tube, through which the pressure in the pipette can be adjusted by a piston or
water pump. The pipette is fixed at an angle that allows imaging of the cell without
276 J.K.H. Hoerber

the danger of touching the pipette with the lever. All these components are located
in a 50-ml container. The glass plate above the objective of the optical microscope
forms the bottom of this container.
After several microliters of the cell suspension have been added, a single cell can
be sucked onto the pipette and fixed there by maintaining low pressure in the pipette.
Other cells are removed through the pumping system. The cell fixed at the pipettes
end is placed close to the AFM lever by a rough approach with screws and finally
positioned by the piezo scanner. When the cell is in close contact with the levers
tip, scanning of the capillary with the cell attached leads to position-dependent
deflections of the lever. The levers used were micro-fabricated silicon nitrite trian-
gles with a length of 200 m and a spring constant of 0.12N/m. The forces that can
be applied with these levers and the sensitivity of the detection system can be as
good as 0.1nN.
With such forces acting on the plasma membrane of a cell, the stiffness of the
structures dominates the images. Therefore, the scaffolds of the cortical layer of actin
filaments and actin-binding proteins, which are cross-linked into a three-dimensional
network and closely connected to the surface membrane, are prominent in the images.
The relatively stable arrangements of the actin filaments are responsible for their rela-
tively persistent structure. However, these surface actin filaments are not permanent.
During phagocytosis or cell movement, rapid changes of their shape can be observed.
These changes depend on the transient and regulated polymerization of cytoplasmic
free actin or the depolarization during the breakdown of the actin filament complex.
The timescale of cell surface changes on a larger scale observed with the AFM was
about 12h at room temperature. Except for small structures some 10nm in size,
everything is quite stable on this timescale. More and larger structures are rear-
ranged within 1015min after these periods of stability.
In a series of experiments, Hrber and coworkers studied the membrane dynam-
ics after infection of the cells with pox viruses [18]. The first effects were observed
seconds to minutes after adding the virus solution to the chamber. After this time,
the cell surface became smooth and so soft that the tip tended to penetrate the cell
surface even with a loading force far below 1 nN. After a few minutes, the cells
became rigid again; with stable imaging, the same structures as before became vis-
ible. During the first hour after infection, no significant variation in the membrane
structures was observed. It is known that within 48h, viruses are reproduced inside
the cell and are passed through the cell membrane via exocytosis [19]. However,
approximately 2.5h after infection, with the AFM a series of processes started at the
membrane. Single clear protrusions became visible, which were growing in size.
The objects quickly disappeared and the original structure of the membrane was
restored. Such processes sometimes occurred several times in the same area and
lasted about 90s for small protrusions of about 20nm and up to 10min for larger
ones (up to 100nm in size). Each process proceeded distinctly and apparently inde-
pendently of the others, was never observed with uninfected cells, and was never
observed before 2h after infection.
The fact that the growing protrusions abruptly disappeared makes it likely that
the exocytosis of particles related to the starting virus reproduction was observed,
10 Atomic Force Microscopy: Applications in the Field of Biology 277

Fig. 10.17 EM images of viral particles at the end of microvilli structures on the left taken from
the publication by Stokes etal. (Adapted from Ref. [19]) compare well with processes observed by
AFM [18] showing the process of microvillus development from the lower right, with the release
of a particle at the end before the last image on the upper left

but not viruses. The first progeny viruses should appear 48h after infection and
are clearly bigger than the structures observed. Nevertheless, after 23h, the early
stage of virus reproduction is finished and the final virus assembly just begins. The
observation of protrusions after this characteristic timespan supports the assump-
tion that they are related to the exocytosis of protein agglomerates originating from
the virus assembly.
Significantly more than 6h after infection, even more dramatic changes at the
cell membrane can be observed. Large protrusions, with cross sections of 200
300nm, grow out of the membrane near deep folds. These events occur much less
frequently than those observed after 2h. These large protrusions also abruptly dis-
appear, leaving behind small scars at the cell surface. Considering the timing and
their size, it is very likely that these protrusions are now progeny viruses exiting the
cell after the assembly is finished. With approximately 20100 virus particles exit-
ing an infected cell, and with roughly 1/40 of the cell surface accessible to the AFM
tip, the observation of one or two of those events should be possible. In the experi-
ments, the number of processes exhibiting the right size and timing observed varied
between none and two. During the first experiment of 46h of continuous observa-
tion of a single infected cell, such events were seen 19 and 35h after infection.
Stokes could show with electron microscopy that individual viruses sometimes exit
the cell at the end of microvilli [19]. Such a process might have been observed in
one case with the AFM when the cantilever tip was pushed up by the growing
microvillus and the release of a hard structure with the right size for a viral particle
was detected. The release of other hard structures of the right size occurred in flatter
regions, but the number of microvilli observed always increased significantly after
infection. The striking similarity between the EM and AFM images shown in
Fig.10.17 makes us believe that indeed the exocytosis of a progeny virus through
the membrane of a living infected cell was observed and that the AFM can be used
to study such processes.
These experiments make two advantages of the AFM obvious: First, it provides
a resolution similar to that of the EM without the need for fixation and coating of the
278 J.K.H. Hoerber

Fig. 10.18 Schematic of the AFM/patch-clamp setup (top) with an image of the cantilever pipette
approach controlled through an optical microscope (left)

sample; and, second, it can produce a movie of a biological process under natural
conditions for the cells studied similar to what optical microscopy can provide at a
lower resolution.

10.3.4 AFM Application in Electrophysiology

The method of fixing a cell to a pipette and scanning it across an AFM cantilever tip
is flexible enough to allow the integration of and combination with well-established
cell physiological techniques of manipulation and investigation of single cells.
The structures observed with the AFM can be related to known features of mem-
branes, but more important than just the imaging of structures with this technique is
the possibility of observing their time evolution. In this way, the AFM can provide
information that brings us closer to understanding not only the being of these
structures, but also their becoming.

10.3.4.1 Single Cells and Excised Membrane Patches

In principle, the AFM setup with the cell fixed on the end of a pipette allows simul-
taneous patch-clamp measurements for studying ion channels in the membrane of
whole cells. Therefore, a logical step in the development of combinations of the
AFM technique with established cell biological techniques was a combined patch-
clamp /AFM setup that could be used to investigate excised membrane patches for
single-ion-channel recording (see Fig.10.18) [20]. In particular, the study of mech-
anosensitive ion channels in the membrane, which become activated by mechanical
stress in the membrane, is an obvious application for a force-measuring device such
10 Atomic Force Microscopy: Applications in the Field of Biology 279

as an AFM. The importance of such ion channels is very obvious in our senses
of touching and hearing. During the development that started in 1991, a new type of
patch-clamp setup was developed that was much more stable to satisfy the needs
of AFM applications. It was designed to accommodate the usual electrophysiologi-
cal and optical components besides the AFM. The chamberwhere a constant flow
of buffer solution guarantees the right conditions for the experimentsconsists of
two optical transparent plastic plates, one at the top and one at the bottom, keeping
the water inside just by its surface tension. In this way, this type of flow cell is acces-
sible from two sides. The chamber is mounted on an xyz-stage together with the
optical detection of the AFM lever movement and a double-barrel application
pipette. The latter was integrated to use the setup also for standard patch-clamp
measurements, with the application of chemicals and different kinds of ions at dif-
ferent concentrations to characterize the sample further. The patch-clamp pipette in
this setup is again mounted on a piezotube scanner fixed within the field of view of
an inverse optical microscope necessary to control the approach of the pipette to the
cell and to the AFM cantilever tip. In patch-clamp experiments, small membrane
pieces are excised from a cell inside the chamber, which can contain none, one, or a
few of the ion channels to be studied. This allows the measurement of currents through
a single ion channel that is only open on the timescale of micro- to milliseconds,
resulting in currents in the picoampere range.
In such a setup, with a membrane patch at the end of the pipette, the AFM can
image the tip of the pipette with the patch on top. Furthermore, changes in the patch
according to the changing pressure in the patch pipette can be monitored, along with
the reaction of the patch to the change of the electric potential across the membrane
[20]. In the images, cytoskeleton structures are clearly seen excised together with
the membrane. On these stabilizing structures of the membrane and on the rim of
the pipette, resolutions down to 1020nm can be achieved, showing reproducible
structures and changes that can be induced by the application of force.
Voltage-sensitive ion channels change shape in electrical fields, leading eventu-
ally to the opening of the ion-permeable pore. To investigate the size of this electro-
mechanical transduction, the AFM setup was used in studies of cells from a cancer
cell line (HEK 293) that were kept at a certain membrane potential (voltage-
clamped) [21]. Cells transfected with Shaker K+ ion channels were used as controls.
In these control cells, movements of 0.55nm normal to the plane of the membrane
were measured, tracking a 10-mV peak-to-peak AC carrier stimulus to frequencies
above 1kHz with a phase shift of 90120, as expected of a displacement current.
The movement was outward with depolarization, and the amplitude of the movement
only weakly influenced the holding potential. In contrast, cells transfected with a
non-inactivating mutant of Shaker K+ channels showed movements that were sensi-
tive to the holding potential, decreasing with depolarization between 80 and 0mV.
Further control experiments used open or sealed pipettes and cantilever place-
ments just above the cells. The results suggested that the observed movement is
produced by the cell membrane rather than by artificial movement of the patch
pipette and/or acoustic or electrical interaction of the membrane and the AFM tip.
The large amplitude of the movements and the fact that they also occurred in control
280 J.K.H. Hoerber

Fig. 10.19 EM images of the mammalian hair cells of the inner ear with the V-shaped bundles of
stereocilia. The right image shows a close-up where the lateral linkers between the stereocilia are
marked by arrows

cells with a low density of voltage-sensitive ion channels imply the presence of
multiple electromechanical motors. These experiments open up a route to exploit
the voltage-dependent movements as a source of contrast for imaging membrane
proteins with an AFM.

10.3.4.2 Tissue Sections

Cochlear hair cells of the inner ear are responsible for the detection of sound
(Fig.10.19). They encode the magnitude and time course of an acoustic stimulus as
an electric receptor potential, which is generated by a still-unknown interaction of
cellular components. In the literature, different models for the mechanoelectrical
transduction of hair cells are discussed [22, 23]. All hypotheses have in common
that a force applied to the so-called hair bundlewhich are specialized stereocilia
at the apical end of the cellsin the positive direction toward the tallest stereocilia
opens transduction channels, whereas negative deflection closes them. For a better
understanding of the transduction process, it is important to know what elements of
the hair bundle contribute to the opening of transduction channels and to study their
mechanical properties. The morphology of the hair cells is described precisely by
scanning and transmission electron microscopy. Unfortunately, this method is
restricted to fixed and dehydrated specimen. Therefore, the development of an AFM
setup to enable studies of cochlear hair cells in physiological solution was thought
to provide further information on the dynamic properties of the system.
To control the approach to the tissue section, the AFM was again built into a
patch-clamp setup onto an upright differential interference contrast (DIC) light
microscope [24]. A water-immersion objective was used to achieve a high
10 Atomic Force Microscopy: Applications in the Field of Biology 281

resolution of about 0.5m even on organotypic cell cultures with a thickness of

about 300m. With this setup, it is possible to see ciliary bundles of inner and outer
cochlear hair cells extracted from rats several days old and to approach these struc-
tures in a controlled way with the AFM cantilever tip. AFM images can be obtained
by measuring the local force interaction between the AFM tip and the specimen
surface while scanning the tip. The question of whether or not morphological arti-
facts occurred at the hair bundles during AFM investigation can be clarified by
preparing the cell cultures for the electron microscope directly after the AFM mea-
surements. Forces up to 1.5nN applied in the direction of the stereocilia axis did not
change the structure of the hair bundles. The resolution achieved with the AFM in
this situation is high enough to image the tips of individual stereocilia and the typical
shape of the ciliar bundle of inner and outer hair cells.
From AFM scan traces, the position of the individual stereocilium as well as its
stiffness can be derived. The force interaction in the excitatory direction shows a
stiffness increase with bending from 0.73103N/m, reaching a steady-state level
at about 2103N/m. The mean stiffness value of stereocilia in the excitatory direc-
tion, calculated at the steady-state level, was determined to be (2.50.6)103N/m
and (3.11.5) 103N/m in the inhibitory direction. Taking into account the stan-
dard deviation of 0.6103N/m, the average force constant showed only a weak
dependence on the position of the stereocilium in the excitatory direction. However,
some stereocilia located at the center had an exceptionally high stiffness, and some
located at the outer region revealed a very low stiffness. In the inhibitory direction,
the mean stiffness in general is slightly higher, with a standard deviation 2.5 times
above the excitatory direction. The higher stiffness might be explained by a direct
contact between the taller and the adjacent shorter stereocilia, while in the excit-
atory direction, shorter stereocilia are pulled by elastic tip links. These links connect
the rows of taller and shorter stereocilia [23]. The high standard deviation might be
an effect of the variation in the spatial interaction between taller and shorter stereo-
cilia. Depending on the angle between the scan direction and the direction defined
by the centers of adjacent taller and shorter stereocilia, mechanical compliance can
vary in a wide range. For the excitatory direction, two stiffness distributions are
distinguishable. One group of values is located around 2.2103N/m, and a smaller
one around 3.1103N/m, which corresponds to stereocilia located in the central
region of the investigated hair bundles. Comparing the stiffness to the correspond-
ing AFM images, one can find a correlation between the arrangement and stiffness
of stereocilia. Those standing a little bit apart from their neighbors clearly show a
higher stiffness, with a value of 3.2103N/m. A possible hypothesis may be that
lateral links connecting these stereocilia with their direct neighbors are oriented in
the direction of the exerted force. This will allow the transmission of additional
force along these links to their neighbors.
The hypothesis above implies that all stereocilia standing closer display much
less interaction with their neighbors mediated by side links. It can be directly tested
by another AFM experiment. The force transmission between adjacent stereocilia in
this experiment is examined by using a lock-in amplifier for detecting the transmit-
ted forces stimulated by a glass fiber touching individual pairs of stereocilia.
282 J.K.H. Hoerber

AFM imaging again allows an exact localization of stereocilia and the stimulating
glass fiber. The AFM force signal is detected by a lock-in amplifier in phase with the
vertical oscillation of the AFM cantilever at 357Hz. The lateral force transmitted by
lateral links can be calculated from the output signal and related to the stereociliums
position. Only that half of the hair bundle can be taken into account, where the fiber
is in contact with the directly stimulated stereocilium across its entire width but
doesnt touch the nearest adjacent stereocilium at all. During the experiment,
the arrangement of stereocilia and stimulating fibers was regularly controlled by
AFM imaging. Force interaction between the stereocilia and the AFM tip leads to
displacements of stereocilia from 0nm to about 250nm. Therefore, the relative
displacement between the directly stimulated stereocilium and the stereocilium
pushed by the AFM tip is expected to result in the continuous stretching of lateral
links located in between. This allows detection of transmitted forces by lateral links
for different states of stretching. For us to be able to compare the results more accu-
rately, the forces have to be normalized with respect to the corresponding maximum
force detected at the directly stimulated stereocilium. Normalized forces rapidly
decrease from the directly stimulated to the first adjacent stereocilium. During the
fiber tips approach to the stereocilium, the relative position is detected, but not the
preload force. Obviously, forces transmitted by lateral links rapidly decrease from a
directly stimulated stereocilium to an adjacent stereocilium in the experiments done
with rats at postnatal age day 4. This result supports the hypothesis of a weak inter-
action between stereocilia by lateral links.
In these experiments, the noise level made it difficult to distinguish between slight
couplings by lateral links from mechanically independent stereocilia. Besides the
modulation signal at 357Hz, the same AFM curves of the identical outer hair cells
contain information about the stiffness of stereocilia as described above. This infor-
mation can be used to detect the mechanical effect of the touching glass fiber on the
stiffness of adjacent stereocilia. If lateral links contribute to the stiffness measured at
individual stereocilia, one would expect to see an increase in the stiffness of adjacent
stereocilia compared to data shown for stereocilia investigated without the fiber
attached. Stiffness data were derived only for the excitatory direction, where the AFM
tip displaces the stereocilia toward the fiber tip. The results of these experiments show
that not only does the stiffness of the directly touched stereocilium increase, but also
the stiffness of the five stereocilia next to it increases. The mean stiffness in the
excitatory direction turns out to be (4.81.8)103N/m, which is about 1.9 times
higher than the mean stiffness of stereocilia not touched by the fiber.
In the experiments described, the AFM allows local stiffness measurements on
the level of individual stereocilia. The results represent the local elastic properties of
the directly touched stereocilium and its nearest neighbors. It can be seen that stiffness
depends on the orientation of links with regard to the direction of the stimulus.
A stimulating fiber had only little mechanical effect on adjacent stereocilia if not
in direct contact with the fiber. For a partial decoupling of the tectorial membrane
from hair bundles of outer hair cells, it therefore can be assumed that only stereo-
cilia still in contact with the tectorial membrane and their nearest neighbors are
displaced by an incoming mechanical stimulus. Lateral links may not compensate
for loss in contact with the tallest stereocilia of the outer hair cells. Decoupling of
10 Atomic Force Microscopy: Applications in the Field of Biology 283

the tectorial membrane following exposure to pure tones at high sound-pressure

levels is supposed to protect hair cells and to avoid damage [25]. A strong interac-
tion between the tectorial membrane and hair bundles of the outer hair cells seems
to be essential for the efficacy of the cochlear amplifier and the transduction of
sound into an electrical signal.
In the experiments described above, the AFM was used only to examine the elas-
tic properties of stereocilia on hair cells, similar to all the micromechanical measure-
ments that already have been performed on the entire stereocilia bundles of sensory
hair cells using thin glass fibers directly attached to the bundle or by using fluid jets.
The receptor potential or transduction current in these experiments was measured in
response to the displacement of stereocilia. To study the kinetics of a single transduc-
tion channel over the whole range of its open probability requires a technique that
allows stimulating a single stereocilium [26]. As shown in the previous section, the
AFM is such a technique, allowing a force to be exerted very locally to an individual
stereocilium. To patch the hair cells with a pipette necessitates removing the support-
ing cells with a cleaning pipette. After cleaning, a patch pipette filled with intracel-
lular solution (concentrations in mM: KCl 135, MgCl2 3.5, CaCl2 0.1, EGTA 5,
HEPES 5, ATP 2.5; at pH7.4) can be attached to the lateral wall of the outermost row
of outer hair cells. This so-called cell-attached configuration is the precursor of the
whole-cell configuration where the microelectrode is in direct electrical contact with
the inside of the cell. For low-noise measurements of single ion channels, the seal
resistance should typically be in the range >1G. A pulse of suction applied to the
pipette breaks the patch, creating a hole in the plasma membrane, and provides
access to the cell interior. During recording, the electrical resistance between the
inside of the pipette and the hair cell should be very small. Many voltage-activated
K+ ion channels are embedded in the lipid membrane of outer hair cells. The opening
and closing of these channels increases the background noise level during transduc-
tion current measurements. The current response of outward-rectifying K+ ion chan-
nels can be controlled by applying 10-mV voltage steps across the cell membrane
(progressively increasing from 100 to +40mV). The outward currents mainly cor-
respond to potassium ion currents of voltage-gated K+ channels. During transduction
current measurements, the holding potential of the hair cell is set to 80mV, corre-
sponding to the reversal potential of K+ ion channels. After forming a seal, the AFM
tip is moved to the top of the corresponding hair bundle under optical control. It then
is used to successively displace each stereocilium within a hair bundle as described
before, but in contrast to the force-transmission measurements, a sinusoidal voltage
is now added to the normal AFM scan signal, modulating the AFM tip in horizontal
directions with an amplitude of 190nm at 98Hz. In this way, the hair bundle is
slightly displaced several times while interacting with the lateral side of the AFM tip.
The tip is repeatedly scanning along the same line while approaching the hair bundle.
As expected, an inward current cannot be detected until a stereocilium is displaced in
the excitatory direction. In the experiments described above, a weak transmission of
force from the directly stimulated stereocilium to adjacent stereocilia was detected,
implying that only a few channels are opened. The results of the elasticity measure-
ments are nicely complemented by the electrophysiological results of experiments
where the tip of the AFM cantilever is repeatedly scanned across the same
284 J.K.H. Hoerber

Fig. 10.20 AFM force signal (bottom) and ion-current response (top) while stimulating a single
stereocilium with the cantilever tip

stereocilium of an outer hair cell from the medial turn of a postnatal rat at day 3. With
applied horizontal forces of up to 800 pN resulting in stereocilia displacements of
about 350nm in the excitatory direction and 250nm in the inhibitory direction, the
AFM tip opens up transduction channels in the excitatory direction for about
90130ms. The current amplitudes reflect the displacements and show for small
forces the expected characteristics of single-channel currents. With these experiments,
Langer and colleagues [2634] demonstrated for the first time single-channel record-
ing of mechanosensitive ion channels (Fig. 10.20).

10.4 Conclusion

Since its invention in 1986, the AFM was continuously developed to suit many differ-
ent applications. In biological research, it provides high-resolution, three-dimensional
imaging not only of single molecules and macromolecular assemblies, but also of
10 Atomic Force Microscopy: Applications in the Field of Biology 285

intact, living cells in physiological solutions and gives a detailed view of surface fea-
tures of large biological structures such as grains and seeds. In addition to imaging,
the AFM has the capability to simultaneously measure biophysical properties, such as
the stiffness of molecular structures, but it also provides the means to analyze binding
interactions down into the range of only a few pico-Newtons and to analyze the effect
of the environment (pH, salt, temperature, etc.) on the interaction. The implications of
this are far-reaching for many biomedical applications, including the development of
new drugs, targeted drug delivery, biocompatible materials for implants, as well as in
vitro and in vivo sensors, and in the agro-food area for quality control.
For the study of many biological samples, a parallel optical observation is essen-
tial to control the approach to the sample and to have available certain standard
controls for the actual sample. However, an AFM built on top of an optical micro-
scope loses part of its stability. A major technological breakthrough in this respect
was the development of a multiple-detection scheme using two cantilevers in paral-
lel to separate force and distance measurement. This development allows long-term
stabilization of the distance between the cantilever tip and the surface with picom-
eter precision, thereby enabling so-called force-clamp measurements. With such an
approach, it is possible to perform unfolding and refolding experiments of even small
protein structures, for instance. The details about the unfolding of single-protein
domains revealed by precise AFM measurements have demonstrated that force
spectroscopy can be used to determine the forces that stabilize protein structures and
to analyze the energy landscape and the transition probabilities between different
conformational states. In this way, the AFM can make essential contributions to the
understanding of the connection between protein structure and function.
The AFM used as an imaging tool and as a force-measuring tool simultaneously
allows the localization of molecular structures and the determination of their bio-
physical properties such as elasticity. Unfortunately, the AFMs that are currently
available commercially are still quite large and bulky, expensive, and difficult to
use, requiring a heavy vibration-stabilization platform. The future development of a
mass-producible micro-AFM with a stability and price that enable field applica-
tions will extend the application range dramatically not only in biomedical and bio-
engineering laboratories, but also in many industrial areas, especially for quality
control. Such a development will automatically include a much higher imaging
speed, allowing the analysis of processes in the millisecond range. It will also elimi-
nate many operational problems, which then can be fully managed by software
developments.
The future AFM will be (1) compact, (2) rapid-scanning, (3) user-friendly, (4)
mass-producible and, consequently, cheap.

References

1. Synge EH. A suggested method for extending microscopic resolution into the ultra-micro-
scopic region. Philos Mag. 1928;6:356.
2. OKeefe JA. Resolving power of visible light. J Opt Soc. 1956;46:359.
286 J.K.H. Hoerber

3. Ash EA, Nicolls G. Super-resolution aperture scanning microscope. Nature. 1972;237:510.

4. Binnig G, Rohrer H, Gerber C, Weibel E. Tunneling through a controllable vaccuum gap. Appl
Phys Lett. 1982;40:178.
5. Binnig G, Quate CF, Gerber C. Atomic force microscope. Phys Rev Lett. 1986;56:930.
6. Atomic force microscope and method for imaging surfaces with atomic resolution, issued to
Gerd K. Binnig of IBMs Zurich Research Laboratory in 1988, U.S. Patent 4724318.
7. Giessibl FJ, Hembacher S, Bielefeldt H, Mannhardt J. Subatomic features on the sili-
con(111)-(77) surface observed by atomic force microscopy. Science. 2000;289(5478):422.
8. Meyer E, Amer NM. Novel optical approach to atomic force microscopy. Appl Phys Lett.
1988;53:2400.
9. Florin E-L, Moy VT, Gaub HE. Adhesion forces between individual ligand-receptor pairs.
Science. 1994;264:415.
10. Haynes CA, Norde W. Globular proteins at solid-liquid interfaces. Colloids Surf B

Biointerfaces. 1994;2:51766.
11. Eckert R, Jeney S, Hrber JKH. Understanding intercellular interactions and cell adhesion:
lessons from studies on protein-metal interactions. Cell Biol Int. 1998;21:707.
12. Rief M, Gautel M, Oesterhelt F, Fernandez JM, Gaub HE. Reversible Unfolding of Individual
Titin Immunoglobulin Domains by AFM. Science. 1997;276:1109.
13. Kellermayer MSZ, Smith SB, Granzier HL, Bustamante C. Folding-unfolding transitions in
single titin molecules characterized with force-measuring laser tweezers. Science.
1997;276:1112.
14. Tskhovrebova L, Trinick J, Sleep JA, Simmons RM. Elasticity and unfolding of single mole-
cules of the giant muscle protein titin. Nature. 1997;387:308.
15. Altmann SM, Grunberg RG, Lenne PF, Ylanne J, Raae A, Herbert K, Saraste M, Nilges M,
Hrber JKH. Pathways and intermediates in forced unfolding of spectrin repeats. Structure.
2002;10:108596.
16. Hberle W, Hrber JKH, Binnig G. Force microscopy on living cells. J Vac Sci Technol.
1991;B9:1210.
17. Cross SE, Yu-Sheng Jin, Jianyu Rao, Gimzewski JK. Nanomechanics of human metastatic
cancer cells in clinical pleural effusions. Nat Nanotechnol. 2007;2:7803.
18. Hrber JKH, Hberle W, Ohnesorge F, Binnig G, Liebich HG, Czerny CP, Mahnel H, Mayr A.
Investigation of living cells in the nanometer regime with the scanning force microscope.
Scanning Microsc. 1992;6:919.
19. Stokes GV. High-voltage electron microscope study of the release of vaccinia virus from
whole cells. J Virol. 1976;18:636.
20. Hrber JKH, Mosbacher J, Hberle W, Ruppersberg P, Sakmann B. A look at membrane
patches with a scanning force microscope. Biophys J. 1995;68:1687.
21. Mosbacher J, Langer M, Hrber JKH, Sachs F. Voltage-dependent membrane displacements
measured by atomic force microscopy. J Gen Physiol. 1998;111:6574.
22. Corey DP, Hudspeth AJ. Analysis of the microphonic potential of the bullfrogs sacculus. J
Neurosci. 1983;3:962.
23. Furness DN, Hackney CM, Benos DJ. The binding site on cochlear stereocilia for antisera
raised against renal Na channels is blocked by amiloride and dihydrostreptomycin. Hear Res.
1996;93:136.
24. Langer MG, ffner W, Wittmann H, Flsser H, Schaar H, Hberle W, Pralle A, Ruppersberg
JP, Hrber JKH. A scanning force microscope for simultaneous force and patch-clamp mea-
surements on living cell tissues. Rev Sci Instrum. 1997;68:2583.
25. Adler HJ, Poje CP, Saunders JC. Recovery of auditory function and structure in the chick after
two intense pure tone exposures. Hear Res. 1993;71:214.
26. Langer MG, Koitschev A, Haase H, Rexhausen U, Hrber JKH, Ruppersberg JP. Mechanical
stimulation of individual stereocilia of living cochlear hair cells by atomic force microscopy.
Ultramicroscopy. 2000;82:269.
10 Atomic Force Microscopy: Applications in the Field of Biology 287

2 7. Goldsbury C, Scheuring S. Curr Protoc Protein Sci. 2002;17.7.117

28. Hrber JKH, Miles M. Science. 2003;302:10025.
29. Braga PC, Ricci D. Atomic force microscopy: biomedical methods and applications. Totowa:
Humana Press; 2004.
30. Connell SDA, Smith DAM. Mol Membr Biol. 2006;23(1):1728.
31. Forman JR, Clarke J. Curr Opin Struct Biol. 2007;17:58.
32. Yang H, Wang Y, Lai S, An H, Li Y, Chen F. J Food Sci. 2007;72(4):655.
33. Gonalves RP, Buzhysnskyy N, Scheuring S. J Bioenerg Biomembr. 2008;40:1338.
34. Liu H, Fu S, Zhu JY, Li H, Zhan H. Enzyme Microb Technol. 2009;45:27481.
Chapter 11
Multi-technique Characterization
of DNA-Modified Surfaces for Biosensing
and Diagnostic Applications

Chi-Ying Lee, Lara J. Gamble, Gregory M. Harbers, Ping Gong,

David W. Grainger, and David G. Castner

Abstract Complementary surface analysis techniquesX-ray photoelectron

spectroscopy (XPS), near-edge X-ray absorption fine structure (NEXAFS), time-of-
flight secondary ion mass spectrometry (ToF-SIMS), and surface plasmon resonance
(SPR)were combined to characterize the structure and composition of DNA-
modified surfaces. Both model systems [thiolated single-stranded DNA (ssDNA) on
gold surfaces] and commercial systems (ssDNA spotted onto microarray slides)
were investigated. Pure thiolated 20-mer ssDNA assembles onto gold, with the
ssDNA binding spontaneously to the surface via the thiol groups and nitrogen atoms

C.-Y. Lee
National ESCA and Surface Analysis Center for Biomedical Problems,
University of Washington, Box 351653, Seattle, WA 98195-1653, USA
Department of Chemical Engineering, University of Washington,
Box 351653, Seattle, WA 98195-1653, USA
3M Health Care Business Group, Infection Prevention Division,
3M China R&D Center, 222 Tian Lin Road, Shanghai 200233, China
L.J. Gamble
National ESCA and Surface Analysis Center for Biomedical Problems,
University of Washington, Box 351653, Seattle, WA 98195-1653, USA
Department of Bioengineering, University of Washington,
Box 351653, Seattle, WA 98195-1653, USA
G.M. Harbers
Replenish, Inc., 2645 Nina St., Pasadena, CA 91107, USA
Department of Pharmaceutics and Pharmaceutical Chemistry,
University of Utah, Salt Lake City, UT 84112-5820, USA
P. Gong
Department of Pharmaceutics and Pharmaceutical Chemistry,
University of Utah, Salt Lake City, UT 84112-5820, USA
Seventh Sense Biosystems, 286 Cardinal Medeiros Ave,
Cambridge, MA 02141, USA

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 289

DOI 10.1007/978-3-319-01360-2_11, Springer International Publishing Switzerland 2014
290 C.-Y. Lee et al.

in the DNA bases, resulting in a monolayer with limited ssDNA chain order. XPS,
NEXAFS, SPR, and radiotracer studies showed that when the pure ssDNA mono-
layers were exposed to short-chain functionalized alkyl thiol diluents [either
11-mercapto-1-undecanol (MCU) or oligo(ethylene glycol) (OEG)], the diluents
initially displaced the weaker goldnitrogen DNA interactions, reorienting the
ssDNA chains to a more upright configuration. After longer exposures to diluent
thiols, some ssDNA chains were displaced from the gold surface. The efficiency of
the target hybridization to complementary DNA from solution depended on the
structure and composition of the immobilized probe ssDNA surface. As the upright
orientation of the ssDNA chains increased, the amount of hybridization increased.
As ssDNA chains were displaced from the surface, the amount of hybridization
decreased. Incorporating the diluent thiol eliminated the small amount of nonspe-
cific binding from noncomplementary target DNA observed on the pure ssDNA
monolayers. The DNA hybridization kinetics were significantly more rapid on
the mixed ssDNA/MCU and ssDNA/OEG surfaces compared to the pure ssDNA
surface. The OEG diluent was more effective than the MCU diluent at reducing
nonspecific protein adsorption during DNA hybridization from blood serum.
Nanoliter drops of amine-terminated ssDNA were robotically spotted onto a glass
microscope slide with an amine-reactive microarray polymer coating; the resulting
150-m spots were imaged with XPS and ToF-SIMS. Imaging XPS provided single-
spot phosphorous, nitrogen, sodium, and silicon elemental images. Small-spot XPS
data were then used to quantify DNA hybridization efficiencies in each microspot
as a function of the ssDNA concentration in the probe printing solution. The DNA
microspots were also readily visualized in the negative ToF-SIMS images of key
fragments from the DNA backbone (e.g., POx), DNA bases (AH, TH, GH,
CH, etc.), and the substrate (e.g., Si). Principal component analysis (PCA) of the
ToF-SIMS images was used to distinguish heterogeneities within the DNA
microspots due to the variations in printing process and solution additives (salts,
sodium dodecyl sulfate, etc.). The different types of data available from combining
these complementary surface analytical methods provide new information essential

D.W. Grainger
Department of Bioengineering, University of Utah, Salt Lake City, UT 84112-5820, USA
Department of Pharmaceutics and Pharmaceutical Chemistry,
University of Utah, Salt Lake City, UT 84112-5820, USA
D.G. Castner (*)
National ESCA and Surface Analysis Center for Biomedical Problems,
University of Washington, Box 351653, Seattle, WA 98195-1653, USA
Department of Bioengineering, University of Washington,
Box 351653, Seattle, WA 98195-1653, USA
Department of Chemical Engineering, University of Washington,
Box 351653, Seattle, WA 98195-1653, USA
e-mail: [email protected]
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 291

to understanding aspects of DNA on surfaces. Such information is important for

designing and improving new technologies that employ nucleic acids on surfaces,
including bioassays, diagnostics, molecular computing, self-assembling materials,
and miniaturized separations.

11.1 Introduction

The current intensive interest in nucleic acids (e.g., DNA, RNA) on surfaces is driven
both by the inherent interest in understanding the interfacial physics and chemistry of
these molecules and by their importance for biosensing and advanced medical diag-
nostics. DNAs primary biological function involves information storage and transfer,
imparted by its unique complementary recognition pairing between the base pairs
adeninethymine (AT) and cytosineguanine (CG) (Fig. 11.1) as the foundation for
genetic coding. The ability of a single-stranded DNA (ssDNA) molecule to seek out
and complex or hybridize its complementary strand in a sample and the ease with
which nucleic acids can be custom synthesized to comprise different lengths,
sequences, and functional groups that allow attachment to various solid supports have
led to the development of DNA-based detection systems, such as DNA microarrays
[15] and biosensors [2, 68]. Microarrays, comprising thousands of hybridization
reactions performed in parallel, are commonly used for genotyping and expression

Fig. 11.1 The DNA double-stranded helix recognition required for many DNA surface-capture
assays. Each DNA molecular strand comprises a series of sugars, polyphosphate backbone, and
attached purine or pyrimidine bases. Two complementary DNA strands complex into duplexes by
specific WatsonCrick and/or Hoogsten hydrogen bonding between paired bases, adenine (A) with
thymine (T), and guanine (G) with cytosine (C). Adenine and thymine interact with two hydrogen
bonds per pairing, while guanine and cytosine interact with three hydrogen bonds per pairing
292 C.-Y. Lee et al.

profiling. Several potential clinical applications have also begun to emerge as methods
and capabilities for DNA microarrays and the data they generate improve. Biosensors
tend to be dedicated to the detection of a small number of analyte sequences in a
sample, with data acquisition often performed in real time. New biosensing tech-
nologies continue to be attractive, with perceived major impacts on the fields of
medicine, forensics, environmental studies, and food safety.
Detection and sequence identification of DNA analyte molecules using DNA
microarrays or biosensors rely on the fidelity of the hybridization of DNA target
strands in solution samples to surface-immobilized complementary probe strands of
known sequences. This bioassay seeks a specific response only for perfect matches
between surface-bound ssDNA probes and complementary strands, with sensitive
distinctions for one-base-pair mismatches. However, the nucleic acid hybridization
behavior observed between complementary probe and target DNA molecules in
bulk solution differs from identical hybridization at a solidliquid interface. While
factors that control hybridization between complementary DNA molecules in solu-
tion have been extensively investigated [9], fewer studies have focused on under-
standing differences in these interactions between surface-immobilized ssDNA
probes and solution-phase targets, where molecular-level processes are more com-
plex and interfacial behavior confounds the reaction. In surface hybridization, non-
specific probesurface interactions, electrostatic forces, and steric issues between
adjacent immobilized DNA probes influence target hybridization and capture
efficiency, kinetics, and capacity. For example, nucleotide primary amines on nonhy-
bridized DNA segments can interact (e.g., covalently [10], or by acidbase adsorp-
tion) with the surface, becoming unavailable to hybridize with target DNA molecules.
Thus, once an ssDNA probe is immobilized to a surface, several important parame-
ters, including surface coverage (density) and orientation, must be determined to
control target interactions. Since these properties determine the bioassay perfor-
mance at several levels (e.g., molecular, macroscopic, signal generation, kinetic),
the abilities to assess and control DNA molecular disposition at surfaces are critical
for improving these formats. This need will undoubtedly push the limits of current
surface analysis techniques to provide molecular specificity, extreme sensitivity,
spatial information, and the ability to determine the chemistry and structure of the
immobilized DNA molecules.
The need to design increasingly selective and sensitive DNA biosensors has stimu-
lated a great deal of interest in understanding the basic physicochemical behavior of
DNA molecules on various technological surfaces, including glass, silicon oxide,
metal oxides, gold, and diamond-like carbon. Surface science has played a key role in
furthering these studies. For example, X-ray photoelectron spectroscopy (XPS),
time-of-flight secondary ion mass spectrometry (ToF-SIMS), near-edge X-ray
absorption fine structure (NEXAFS), infrared spectroscopy (IR), and scanning
probe microscopies (SPMs) represent often-employed surface analytical techniques
for characterizing DNA-modified surfaces. While these techniques are powerful for
investigating the DNA immobilization process, they generally encounter difficulties
in monitoring surface-adsorption processes spatially, in real time and in liquid envi-
ronments, which are critical aspects for studying DNA hybridization on surfaces.
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 293

Currently, in most commonly employed microarray systems, the detection and

monitoring of DNA hybridization rely on the use of extrinsic fluorescent labeling of
select DNA bases. Although fluorescence-based target detection has a number of
attractive features (e.g., broad availability and ease of use), the complex reagent
labeling processes present several problems (e.g., poor reproducibility, lack of abso-
lute molecular detection) and are time-consuming [5]. Additionally, surface-derived
fluorescence signal is subject to many measurement issues and sources of variation
that preclude the accurate quantitation of analyte. Alternatively, label-free detection
techniques, such as surface plasmon resonance (SPR) spectroscopy and SPR imag-
ing [1114], quartz crystal microbalance (QCM) [15, 16] and electrochemical DNA
biosensors [8, 17], provide more direct methods of DNA assay while avoiding fluo-
rescence metric issues. However, these techniques by themselves are not chemically
selective methods and therefore must be used in combination with other analytical
techniques, such as XPS, ToF-SIMS, NEXAFS, and IR, to provide accurate surface-
sensitive information.
The aim of this chapter is to demonstrate the capabilities of several complemen-
tary surface analytical techniques, including XPS, ToF-SIMS, NEXAFS, and SPR,
for investigating the surface and hybridization properties of immobilized ssDNA on
microarray and biosensor surfaces. The information gained from these studies has
helped provide a new understanding of ssDNA interfacial behavior by providing
quantitative correlations between immobilized DNA surface properties (i.e., com-
position, density, and orientation) and the resultant hybridization efficiency from
complex biological samples. The following sections summarize recent findings
from ongoing experiments in our laboratories and present a description of ssDNA
behavior on gold and commercial microarray surfaces provided by these various
analytical techniques.

11.2 Assessing the Structure and Hybridization Properties

of Mixed DNA/Alkylthiol Monolayers on Gold Using
XPS, NEXAFS, and SPR

The immobilization of ssDNA on a transducer surface is frequently the first step in

DNA biosensor design. Several different strategies, including surface entrapment,
physical adsorption, and chemical binding, are employed. Despite the large body of
work recently devoted to this topic, the development of methods for surface-
immobilizing ssDNA that preserve their original hybridization specificity with min-
imal nonspecific interactions remains elusive and hence is an important goal for
improving the performance of DNA microarrays and biosensors. Thus, information
about the structure and coverage of immobilized ssDNA and how these properties
impact subsequent DNA target hybridization would be invaluable for optimizing
the performance of DNA biosensors and other applications using surface-bound
DNA oligomers.
294 C.-Y. Lee et al.

The process of interfacial self-assembly and adlayer formation on surfaces

provides a powerful tool to generate molecular films of biological molecules on
several relevant substrates [18]. The convenience and flexibility of employing
self-assembled monolayers (SAMs) and possibilities for controlling biomolecule
density and surface orientation allow SAMs to play important roles in designing
artificial biomolecular recognition devices. Chemisorbed adlayers of alkylthiols on
gold [18] provide a high degree of control over surface chemical and physical prop-
erties, providing useful model systems for examining relationships between ssDNA
surface composition and orientation, as well as subsequent hybridization behavior
with solution-phase targets. Tarlov et al. [19] pioneered a technique for DNA immo-
bilization that utilizes the sequential chemisorptive self-assembly of first a thiol-
terminated single-stranded DNA (HS-ssDNA) monolayer onto gold surfaces,
followed by exposure to a short hydroxyl-terminated alkylthiol surface diluent [e.g.,
mercaptohexanol (MCH)]. XPS [19, 20], neutron reflectivity [21], and SPR [22]
have been used to study this mixed SAM system; the MCH has been shown to
disrupt nonspecific interactions of DNA bases with the gold surface and enhance
specific attachment through thiolDNA chemistry. These studies have provided
important quantitative information on possible ssDNA probe conformational
changes in self-assembled mixed monolayers resulting from the MCH surface dilu-
ent. Immobilized ssDNA oligomers appear to behave differently when their packing
density or other properties (e.g., DNA length) are varied in these films. To develop
a deeper understanding of surface properties in these mixed DNA monolayers at the
molecular level, the packing density and nonfouling properties of ssDNA were sys-
tematically changed by incorporating other alkylthiol diluents with different func-
tional head groups [e.g., 11-mercapto-1-undecanol (MCU) and oligo(ethylene
glycol) (OEG)] into ssDNA monolayers. These strategies for DNA immobilization
onto gold surfaces are depicted schematically in Figs. 11.2a, b.
The effects of diluent backfilling time on the mixed adlayer composition, surface
density, DNA orientation, and target hybridization efficiency of ssDNA oligomers
were studied using complementary surface analytical techniques including XPS,
NEXAFS, and SPR. XPS is used to quantify the composition of the DNA films.
Polarization-dependent NEXAFS was used to probe the molecular orientation and
ordering of surface-bound ssDNA molecules. SPR was used to measure the target
DNA hybridization in situ.
XPS, also known as electron spectroscopy for chemical analysis (ESCA), is a
quantitative surface analytical tool sensitive to the atomic composition of the outer
210 nm of a material (the XPS technique is discussed in more detail in Chap. 2 of
this book). Compositional changes of mixed ssDNA/alkylthiol diluent monolayers
on gold were followed as a function of the MCU or OEG diluent backfill time
(0.518 h) starting from an initial ssDNA monolayer. Figure 11.3 shows XPS
elemental compositions from pure ssDNA and mixed ssDNA/MCU monolayers.
(Note: XPS analysis was also performed on the ssDNA/OEG monolayers [23]; data
not shown.) For the surface chemistries used here, elements P and N are unique to
DNA and are therefore good indicators of the relative amounts of ssDNA present on
the surface under each condition. After short-term exposure of the DNA monolayer
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 295

Fig. 11.2 Scheme for thiolated, single-stranded DNA (HS-ssDNA) immobilization onto gold.
Mixed ssDNA/mercapto-1-undecanol (MCU) (a) and ssDNA/oligo(ethylene glycol) (OEG) (b)
monolayers of varying DNA surface coverage were assembled by a known, sequential two-step
process onto gold surfaces. First, pure DNA monolayers were prepared by immersing freshly gold-
coated substrates in HS-ssDNA solutions. After HS-ssDNA assembly, samples were rinsed thor-
oughly and then immersed in MCU or OEG diluent thiol solution for backfill times ranging from
30 min to 18 h. After the specified diluent thiol backfill time, samples were removed from the
solutions, rinsed thoroughly, and stored under N2 until analysis

to the MCU diluent thiol (<1 h), the relative atomic percents of C and S increased,
while the relative atomic percents of Au, P, N, and O all decreased (see Fig. 11.3).
Increased C and S content is consistent with the presence of MCU, as the smaller
MCU diluent thiols take up unoccupied gold sites surrounding the loosely packed
DNA on the surface [19, 23, 24]. The corresponding Au signal decreased during this
same short-term exposure of the DNA monolayer to MCU, consistent with this
interpretation. MCU surface reactions could also displace other weaker, nonspecific
interactions between nitrogen-containing DNA bases and gold, promoting single-
point tethering of DNA oligomers on gold via thiolate bonds, with a greater ten-
dency to orient away from the surface. NEXAFS data further support this contention.
296 C.-Y. Lee et al.

Fig. 11.3 XPS compositional data for pure ssDNA and mixed ssDNA/MCU monolayers on gold
as a function of MCU back-filling time. XPS results indicate that during short-term (1 h, green
bars, upper) MCU diluent thiol backfill, the MCU molecules incorporate into the unoccupied gold
surface sites surrounding the loosely packed low-density thiolated-ssDNA. While the DNA dis-
placement by MCU is initially slow, upon extended MCU backfill time (>1 h, blue bars, lower), the
ssDNA surface coverage steadily decreases. This change is supported by the observed reduction in
XPS P2p and N1s, and increase in S2p and Au4f signals at later times, reflecting thinner layers
with higher density MCU sulfur and reduced DNA signals

XPS analysis of sequentially adsorbed ssDNA/MCU and ssDNA/OEG monolayers

on gold indicated that both MCU and OEG molecules initially incorporate into the
unoccupied surface sites surrounding the ssDNA [19, 23, 24]. While DNA displace-
ment by the MCU and OEG diluents is initially slow, upon extended exposure or
backfill time (>1 h), these diluents eventually also displace adsorbed ssDNA mole-
cules from the gold surface. As seen in Fig. 11.3, the relative Au substrate atomic
percent increased, while the relative atomic percent of P and N decreased further for
the longer MCU exposure times. Thus, ssDNA surface coverage steadily decreased
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 297

Fig. 11.4 Nitrogen K-edge

NEXAFS spectra from pure
ssDNA and mixed ssDNA/
MCU monolayers on gold at
normal (90) and glancing
(20) incident X-ray angles
(t = MCU backfill time in
hours). The increase in the
polarization dependence of
nitrogen K-edge NEXAFS
spectra indicates that DNA
bases are oriented more
parallel to the surface than
bases in the pure ssDNA
monolayer, and that ssDNA
oligomers reorient on average
toward a more upright
orientation on the surface
upon MCU addition (Adapted
with permission from Ref.
[24] . Copyright 2006
American Chemical Society)

with longer MCU and OEG backfill times. These mixed DNA monolayers on gold
displacement trends over time have been fully characterized in a companion work
utilizing a 32P radiolabel assay [25]. The maximum probe coverage was observed to
be 4.4 (0.4) 1013 probes/cm2 for pure ssDNA adlayers without MCU backfill.
With increasing MCU backfill time, the DNA surface coverage eventually reached
a density of ~1.7 1013 probes/cm2 after 18 h [23, 25, 26].
Polarization-dependent NEXAFS, also known as X-ray absorption near-edge
structure (XANES) [27], was used to determine the orientation changes of these
DNA monolayers as a function of diluent thiol exposure, co-adsorption, and DNA
surface displacement. NEXAFS has been used to examine the surface molecular
orientation and ordering for a variety of materials, including SAMs on gold [28] and
polymers [2932]. In previous DNA studies, NEXAFS has been used to character-
ize DNA nucleobases and nucleotides [3337], as well as the orientation of poly-dA
[37] and poly-dT [37, 38] DNA homo-oligomers on gold.
Figure 11.4 shows the nitrogen K-edge NEXAFS spectra from pure ssDNA and
mixed ssDNA/MCU monolayers on gold at normal (90) and glancing (20) inci-
dent X-ray angles [24]. For the pure ssDNA monolayer, the intensity of the * peaks
was slightly higher when the X-ray beam was at a glancing angle of incidence
compared to that at normal incidence. Overlap of the E vector of the polarized
298 C.-Y. Lee et al.

X-ray source with the aromatic nitrogen bonds that cause the 1s * transition
at glancing incidence indicates that the DNA bases in a pure DNA film were, on
average, oriented more parallel to the surface. Similar trends have previously
been observed with surface-bound double-stranded DNA oligomers on gold [39].
This indicates that on average the ssDNA chains had a slightly perpendicular orien-
tation to the substrate. It is believed that the electrostatic repulsive forces between
the ssDNA chains may cause the chains to stand, on average, slightly upright on the
surface although it is likely that overall they remain rather disordered. As the diluent
molecules first incorporate into preadsorbed DNA monolayers (<1 h backfill time),
DNA oligomers adopt a more upright orientation, due to MCU or OEG displace-
ment of DNA nucleotide base amine groups from the gold interface [23, 24].
The ssDNA surface orientation (followed by monitoring the N 1s * transition)
reaches a maximum at ~0.51 h of MCU or OEG exposure, beyond which the
ssDNA surface orientation decreases slightly due to DNA displacement. With less
DNA on the surface, electrostatic repulsive interactions between ssDNA chains are
less effective at holding the DNA molecules perpendicular to the surface, allowing
more disorder among the immobilized DNA chains. Despite the reduction in the
N 1s nitrogen signal at longer diluent backfill times, NEXAFS polarization depen-
dence at 18 h was still greater than that of the pure ssDNA surface, leading to the
conclusion that the change in polarization dependence was due to relative orientation
changes in the DNA layer [23, 24].
DNA hybridization on these mixed ssDNA/diluent thiol monolayers in purified
target DNA samples as well as blood serum was measured by SPR [23, 25]. SPR is
a surface-sensitive optical technique frequently used to detect binding of biological
molecules such as DNA [26, 4045] and proteins [4651] onto chemically and bio-
logically modified gold surfaces without the need for target labeling and complex
sample preparation. For the pure ssDNA surface, SPR response to a noncomple-
mentary ssDNA in buffer solution indicates a small amount of nonspecific ssDNA
binding to the ssDNA surface (Fig. 11.5a). With complementary DNA target, a
much higher SPR wavelength shift was observed, indicating the hybridization of
surface-bound ssDNA probes with complementary ssDNA target. The SPR sensor-
gram also showed that target hybridization occurs slowly on the pure ssDNA sur-
face, as the SPR curve for complementary ssDNA does not reach saturation even
after 60 min of target incubation. In contrast, when MCU or OEG diluent was incor-
porated into the ssDNA monolayer before hybridization, the nonspecific binding of
noncomplementary ssDNA to any of the mixed monolayer probe surfaces was not
detected (Fig. 11.5b). This suggests that the alkylthiol diluents effectively block
unoccupied surface sites to prevent nonspecific ssDNA surface binding, thereby
increasing the hybridization selectivity to target. Furthermore, target hybridization
to the mixed DNA/thiol diluent probe surfaces also reached saturation much more
rapidly than to the pure DNA probe surface. The observed increase in hybridization
kinetics as well as hybridization efficiency on surfaces with short-term exposure
(30 min) to MCU or OEG (for which 32P-labeling results indicate DNA surface
probe densities are comparable to the nonbackfilled pure DNA adlayer [25]) indi-
cates that more than just the ssDNA surface density may be controlling the
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 299

Fig. 11.5 Real-time SPR measurement of target DNA hybridization on (a) pure ssDNA and (b)
mixed ssDNA/MCU adlayers from buffer (1 M NaCl-TE, pH 7.0) demonstrating the effects of
MCU diluent backfill on DNA target hybridization. Kinetic sensorgram traces were characterized
by the following common stepwise features: A measurement baseline was first established by run-
ning buffer over the probe surface. Then noncomplementary DNA (1 M) in buffer was injected to
test nonspecific target binding onto the probe surface (1). As the noncomplementary DNA adsorp-
tion approached saturation, the noncomplementary DNA solution was replaced with buffer to rees-
tablish the baseline (2). Complementary DNA target (1 M) was then injected to determine the
amounts of hybridization (3). As the DNA hybridization approached saturation, the complemen-
tary DNA solution was replaced with buffer to rinse away loosely bound target DNA molecules
from the probe surface (4). Data indicate that the target hybridization reaches a maximum at 1 h
MCU backfill (surface density of 3.55 1013 molecules/cm2) (Adapted with permission from Ref.
[25]. Copyright 2006 American Chemical Society)

hybridization efficiency of surface-bound ssDNA. Orientation studies of ssDNA

chains in identical adlayers using NEXAFS as described previously [23, 24] dem-
onstrate that the initial MCU and OEG addition into the pure DNA adlayer displaces
nonspecifically adsorbed DNA nucleobase amines from gold surface sites and reori-
ents the probe DNA chains to a more upright configuration. With nucleobase amines
300 C.-Y. Lee et al.

Fig. 11.6 Amounts of DNA hybridization from various serum concentrations determined by SPR
on ssDNA/MCU and ssDNA/OEG adlayers. For the ssDNA/MCU adlayer, the hybridization effi-
ciency was significantly reduced (by roughly 50%) from 15% serum, and in 50% serum and higher
concentrations, no hybridization was detected on the probe surface because of rapid, overwhelm-
ing amounts of nonspecific protein adsorption. OEG incorporation into the ssDNA adlayer signifi-
cantly improved the surface resistance to both nonspecific protein and DNA adsorption, allowing
the detection of small DNA target sequences from concentrated, complex biological mixtures

detached from the gold surface, probe DNA molecules are more likely single-point
end-tethered and therefore more configurationally mobile than those in a pure DNA
monolayer. This produces mixed DNA adlayers that are more accommodating for
target DNA molecules to approach and to hybridize. SPR results indicate that the
target hybridization signal reaches a maximum on surfaces following 1 h diluent
backfill, after which the target hybridization signal decreases due to significant dilu-
ent thiol displacement of DNA probes from the adlayer surface [25].
In order for microarray technology to be widely adapted as a diagnostic tool and
to perform reliably in field-based or direct sample-to-answer analyses, surface-
capture nucleic acid biosensors will need to achieve improved target-detection lim-
its and be able to capture target from complex milieu such as whole blood or other
complex biological mixtures rather than depend on initially purified analytes free
from assay-complicating proteins [52, 53]. Figure 11.6 demonstrates the effects of
capturing target analyte from a model complex biological mixture. Although similar
amounts of target capture were detected by SPR on both ssDNA/MCU and ssDNA/
OEG adlayers from purified target DNA samples in buffer solution, the target
hybridization efficiencies of these two probe surfaces differed significantly when
performed in complex biological mixtures (e.g., serum) (Fig. 11.6). Although MCU
addition into the ssDNA adlayer improved the surface hybridization by both orient-
ing immobilized probe DNA (e.g., preventing ssDNA nonspecific interactions with
the gold surface) and providing effective resistance to adsorption of noncomple-
mentary DNA, these ssDNA/MCU surfaces were not sufficiently protein-resistant
to perform measurements in complex milieu. Results of SPR-based DNA hybridiza-
tion from various serum dilutions showed that both the DNA hybridization kinetics
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 301

and the capture efficiency were adversely affected by nonspecific protein adsorp-
tion, even at a minimum serum concentration of 1 v/v% when compared to target
capture from pure buffer [25]. No target hybridization was detected on the
ssDNA/MCU surfaces in SPR from serum concentrations above 30 %, indicating
the substantial interference of nonspecific protein adsorption with specific DNA
capture and hybridization. However, the target hybridization on ssDNA/OEG sur-
faces decreases only slightly (by roughly 20 %) as the serum concentration is
increased to 50 %. In undiluted serum (100 %), the target hybridization on the probe
surface was reduced by approximately 80 %, but still well above baseline noise, to
1.3 1012 molecules/cm2. Comparing these results to data for DNA hybridization on
ssDNA/MCU-probe surfaces, Lee et al. found that OEG backfill significantly
improves the selectivity of the DNA probe surface, allowing target detection directly
from undiluted serum [23]. One conclusion is that an OEG diluent background is
more effective at preventing nonspecific protein binding than the hydroxyl-
terminated MCU alkylthiol diluent background, making it possible to detect a given
target DNA strand within a solution containing a substantial protein background.
Nevertheless, the differences between assay results from buffer versus those in com-
plex media demonstrate the performance challenge for nucleic acid capture in the
presence of substantial competing, nonspecific adsorption noise.

11.3 Surface Chemical State Image Analysis of DNA

Microarrays Using XPS and ToF-SIMS

In addition to the immobilization chemistry used for attaching ssDNA probes to the
solid supports, another important aspect of microarray fabrication is the physical
deposition method used to create the DNA microspot patterns on the surface.
Methods for fabricating micron-scale DNA patterns fall into two broad categories:
(1) in situ synthesis, a base-by-base attachment to build up different DNA strands at
different sites in the array; or (2) ex situ synthesis and the subsequent attachment of
different complete strands at individual array sites [53]. For the former approach,
Affymetrix has developed a photolithographic process to create DNA arrays using
a series of exposures through masks to control the addition of bases to specific sites
on the surface [54]. This method offers the advantage of having the oligonucleotide
synthesized on the same support used for assay hybridization, obviating the need to
remove the DNA strand from its synthetic support and reattach it to the microarray
surface. However, in situ methods of DNA strand synthesis on the surface do not
allow for independent confirmation of the fidelity of the probe synthesis, provide
variation in probe density, or allow purification of the oligonucleotide after synthe-
sis to eliminate incomplete strands or strands with excess bases, thereby reducing
the overall array reliability. In addition, DNA probe lengths are limited practically
to shorter oligomers (~2550-mers).
In the second method, completely synthesized and prepurified DNA strands are
attached to the array surface. Several factors must be considered in the production
302 C.-Y. Lee et al.

of these microarrays using presynthesized oligonucleotides, including derivatiza-

tion of the substrate with functional groups, derivatization of nucleic acid probes
with complementary functional groups, delivery of minute volumes (e.g., nanoliters)
of DNA solution to the surface using contact or noncontact printing techniques, and
deactivation of remaining unreacted surface-bound functional groups after
immobilization.
Micro-printing techniques are widely used for DNA microarray fabrication on
commercial array slides containing hundreds to thousands of spotted micron-sized
features. The printing process generally involves dispensing nanoliter drops of DNA
print solutions onto solid reactive array surfaces using a robotic spotter interfaced
with DNA probe source plates. The nanoliter drops of DNA solution evaporate
within a few seconds of surface residence without reaching equilibrium in terms of
DNAsurface interactions, covalent reactions on surfaces, or constant ionic strength.
This rapid evaporative process produces distinct differences in immobilized DNA
structure, density, and chemistry compared to bulk solution-coupling reactions
between DNA and surfaces [55, 56]. The formation of dry DNA microspots with
greater DNA density at the edges than in the middle (e.g., donut morphologies) is
commonly observed, as drying or wetting print solution flows to spot edges upon
rapid evaporation on micro-printed array surfaces [57]. When contact-based pin-
printing methods are used, surface damage may also occur during the array printing
process [1]. The resulting immobilized DNA density and distribution within indi-
vidual microarray spots, completely distinct from that produced by bulk surface
immobilization, have profound influences on the subsequent target-capture perfor-
mance (see Sect. 11.2) [19, 26]. Spot-to-spot and intraspot variations in DNA surface
density and distribution can therefore lead to inconsistent target capture, substantial
spotspot and assayassay variability, inaccurate data quantification, and mislead-
ing results. Thus, the accurate quantitative analysis of printed DNA microarray
diagnostics is possible only if controlled and reliable spot uniformity (i.e., spot density,
intraspot DNA density control, spot size, and shape repeatability) is achieved.
XPS and ToF-SIMS each exhibit intrinsic strengths and weaknesses with respect
to generating surface chemical state information at a high spatial resolution but used
together provide a powerful complementary set of techniques. The quantitative
nature of XPS combined with its 210-nm sampling depth makes it ideal for deter-
mining surface concentrations of biomolecules, including DNA [23, 24, 55, 5861].
Innovations in X-ray focusing and lens/analyzer technology now permit XPS imaging
at spatial resolutions lower than 10 m [6265]. Although this spatial resolution
remains orders of magnitude above that obtained with other microscopy techniques,
imaging XPS has significant advantages in quantifying the sample surface composi-
tion. In contrast to imaging XPS, imaging ToF-SIMS is a more surface-sensitive
technique (12-nm sampling depth), providing a significantly higher spatial resolu-
tion that allows a more detailed analysis of the compositional variability within a
biomolecular pattern on solid substrates [6668]. Previous studies have shown that
static ToF-SIMS, in combination with multivariate analysis statistical methods such
as principal component analysis (PCA), can provide the distribution of chemical
species across a patterned surface at a submicrometer resolution [69, 70]. An example
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 303

Fig. 11.7 XPS elemental images (800 m 800 m) of aminated DNA probes printed onto
CodeLink microarray slides at 40-M DNA concentration. While phosphorus is unique to DNA,
silicon is unique to the substrate. In combination, these elemental images enable unambiguous
identification of the spatial distribution of DNA for the XPS region-of-interest (ROI) composi-
tional analyses of the printed DNA microarray spots. The scale bar represents 200 m (Reproduced
with permission from Ref. [71]. Copyright 2007 American Chemical Society)

of the value of the dual use of imaging XPS and imaging ToF-SIMS to determine
the chemical composition, spatial distribution, and hybridization efficiency of
amine-terminated ssDNA bound to commercial polyacrylamide-based microarray
slides was published in 2007 [71].
Figure 11.7 shows XPS elemental images (P, N, Na, and Si) of noncontact printed
DNA microarray spots (100150-m diameter) on CodeLink (GE Amersham
microarray substrate) slides [55, 71]. Background-corrected P 2p images (Fig. 11.7)
show a higher signal intensity in the immobilized DNA microspot regions. Although
N, Na, and Si are present throughout the substrate surface, higher N 1s and Na 1s
and lower Si 2p signal intensities were observed in the DNA microspots. Higher N
and Na signal intensities correlate with nitrogen-containing DNA bases and sodium
counterions associated with the DNA polyphosphate backbone. DNA coverage in
printed regions attenuates the signal intensity from elements present in the underly-
ing glass substrate (i.e., Si). Also, XPS imaging can distinguish between hybridized
and unhybridized microspots on a commercial microarray slide without the use of
radioactive or fluorescent labels. Figure 11.8 shows XPS overlay images displaying
P, N, Na, and Si signal intensities from a hybridized (Fig. 11.8a) and an unhybrid-
ized (Fig. 11.8b) microarray slide. Prior to exposing the microarray slide to comple-
mentary target capture, similar XPS P 2p, N 1s, and Na 1s signal intensities were
detected from all microspots, regardless of whether the microspots were printed
utilizing complementary or noncomplementary probe sequences. After target
hybridization, microspots with complementary probe sequences had higher XPS
P 2p, N 1s, and Na 1s signal intensities compared to microspots with noncomple-
mentary probe sequences, as expected [71].
304 C.-Y. Lee et al.

Fig. 11.8 XPS overlay of phosphorous (P2p), nitrogen (N1s), and sodium (Na1s) with the
substrate silicon (Si2p) signal intensity images (800 m 800 m) from printed DNA probes on
CodeLink polymer microarray slides before (a) and after (b) DNA target hybridization.
Consistent with the target capture signal, the XPS P2p, N1s, and Na1s signal intensities from the
hybridized regions are significantly higher than from the unhybridized regions. Amounts of
DNA on CodeLink surfaces are proportional to the P2p atomic percent (at. %) (c). Microarrays
were printed at three DNA probe concentrations (10, 20, and 40 M). Target hybridization effi-
ciencies for the microarray shown in parentheses above each concentration were derived as a
percentage of probe molecules hybridized {[(P2p at. % of hybridized spot/P2p at. % of probe
spot) 1] 100 %} (Reproduced with permission from Ref. [71]. Copyright 2007 American
Chemical Society)

Small-area region-of-interest (ROI) analyses were performed to obtain quantita-

tive elemental composition information about individual DNA spots both before
and after hybridization. The results from these ROI analyses were then used to
determine target hybridization efficiencies for microarrays printed at various printed
probe concentrations (e.g., 10, 20, and 40 M). Microspots with hybridized comple-
mentary sequences show higher percentages of phosphorus (2.2 atomic percent, or
at.%), nitrogen (13.1 at.%), and sodium (5.6 at.%) and lower substrate oxygen (23.0
at.%) and silicon (3.3 at.%) signals compared to microspots with unhybridized non-
complementary probe spots (P = 1.3, N = 11.8, Na = 3.1, O = 25.7, and Si = 5.4 at.%).
Due to the substantial amounts of nitrogen (~8 at.%) in the unprinted CodeLink
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 305

acrylamide-based polymer layer [55], the phosphorus signal from the DNA back-
bone is the only unique characteristic element useful for quantifying relative
amounts of surface-immobilized and hybridized DNA oligomers. The DNA surface
coverage is proportional to the phosphorus atomic concentration, as has been shown
previously for immobilized DNA on gold [25, 61] and polymer-modified silicon
substrates [55, 58]. Figure 11.8c shows relative amounts of surface-immobilized
probe and hybridized DNA obtained for each probe printing concentration from
small-area XPS analysis [71]. Figure 11.8c shows the P 2p at.% from the DNA
microspots increases with increasing spotting-probe solution concentration. The
target hybridization efficiencies shown in parentheses above each concentration in
Fig. 11.8c were calculated from the XPS P 2p signal to determine the percentage of
probe molecules hybridized {[(P 2p at.% of hybridized spot/P 2p at.% of probe
spot) 1] 100%}. A hybridization efficiency of 86% was obtained for microspots
printed at a probe concentration of 10 M. At higher probe printing concentrations
(20 and 40 M), slightly lower hybridization efficiencies (80% and 73%) were
obtained [71]. A reduction in the hybridization efficiency at a higher probe coverage
has been reported previously [25, 26, 55] and can be explained by steric and elec-
trostatic crowding effects in closely packed DNA probes that hinder DNA target
duplex formation on the microarray surface [72].
In addition to determining DNA microspot composition, knowing the distribu-
tion of immobilized DNA molecules within individual DNA microspots is desir-
able, as described above. Imaging ToF-SIMS is able to provide distinct analytical
data on the lateral distribution of DNA within single-array microspots before and
after target hybridization via detection of the characteristic DNA ion fragments (i.e.,
DNA bases and phosphate backbone) [59, 60, 73, 74]. Negative-ion ToF-SIMS ROI
spectra from the DNA microspotted regions (Fig. 11.9b) show phosphate fragments
from the DNA backbone (PO, PO2, PO3 , and H2PO4) at m/z 47, 63, 79, and 97
[59, 60, 73, 74], as well as unique fragments from DNA bases such as adenine
(AdeH, m/z = 134), thymine (ThyH, m/z = 125), guanine (GuaH, m/z = 150), and
cytosine (CytH, m/z = 110) [59, 60]. These DNA-related peaks were absent in the
negative-ion spectra from the unspotted substrate region (Fig. 11.9a), confirming
their origin from surface-immobilized DNA molecules. Individual negative-ion
ToF-SIMS images for selected masses show the distribution of printed DNA within
an unhybridized microspot containing noncomplementary DNA probes (Fig. 11.10a)
and a hybridized microspot containing complementary probes (Fig. 11.10b) from
the microarray region printed with 40-M DNA solution [71]. A microarray spot
diameter of approximately 150 m was determined from the ToF-SIMS POx
images, which is comparable to that observed using fluorescence microscopy after
printing (Fig. 11.10c). ToF-SIMS images of the Si peak on the unhybridized
microspots reveal a donut feature around the probe spot consistent with those
seen in fluorescent images (Fig. 11.10c) (PCA analysis of these halos is presented
later in the chapter). Also, consistent with XPS data, images for characteristic DNA
fragments (m/z 47, 63, 79, 97, 110, 125, 134, and 150) show higher signal intensities
for the hybridized DNA microspot (Fig. 11.10b) [71].
306 C.-Y. Lee et al.

Fig. 11.9 Negative-ion ToF-SIMS ROI spectra from the (a) substrate and the (b) DNA regions of
the microarray surface. The DNA region shows characteristic nucleic acid peaks at m/z 42 (CNO),
63 (PO2), 79 (PO3), 97 (H2PO4), 110 (C4H4N3O, CytH), 125 (C5H5N2O2, ThyH), 134
(C5H4N5, AdeH), and 150 (C5H4N5O, GuaH) that were absent or present at much lower intensi-
ties in the background substrate region (Reproduced with permission from Ref. [71]. Copyright
2007 American Chemical Society)

ToF-SIMS was also employed to identify the DNA hybridization signal by using a
complementary target sequence in which 50% of the DNA bases were modified with
one Br atom (Fig. 11.11a). Printed probe microarrays exposed to Br-modified DNA
complementary targets produced strong Br signals from hybridized probe spots
(Fig. 11.11c) compared to noncomplementary probes (Fig. 11.11b). ToF-SIMS
intrinsic high sensitivity in detecting brominated molecules, combined with its dem-
onstrated submicron spatial resolution, opens new possibilities to exploit this analyti-
cal method to determine hybridization uniformity across single microarray spots.
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 307

Fig. 11.10 Representative negative-ion ToF-SIMS images showing the distribution of DNA and
substrate fragments within a single (a) unhybridized and (b) hybridized microarray spot. The DNA
fragments are localized to the noncontact printed regions but distributed inhomogeneously within
the microspot. The Si image from the unhybridized probe spot (a) showed a halo feature around
the spot. A brighter pixel intensity corresponds to higher DNA or substrate signals (counts per
pixel). ToF-SIMS images are 200 m 200 m. The fluorescent images of the printed probe (left)
and hybridized target (right) were collected using a fluorescent microscope showing a halo
effect; scale bar represents 50 m (c) (Reproduced with permission from Ref. [71]. Copyright
2007 American Chemical Society)

While specific peaks due to expected fragments can be readily identified in the
SIMS spectra, such as the DNA molecular fragments identified above, the energetic
SIMS process yields hundreds of peaks in the 0200-m/z range, making the inter-
pretation of all peaks detected in a ToF-SIMS experiment difficult. To simplify data
308 C.-Y. Lee et al.

Fig. 11.11 (a) Bromine modification of DNA target (50% brominated DNA bases comprise the
DNA target sequence). Representative negative-ion ToF-SIMS images showing no Br fragments
detected from the noncomplementary (unhybridized) microspots (b). Microarrays exposed to
Br-modified DNA targets produce strong Br signals for complementary (hybridized) microspots
after target hybridization (c). A brighter pixel intensity corresponds to higher DNA or substrate
signals (counts per pixel). Images are 200 m 200 m (Reproduced with permission from Ref.
[71]. Copyright 2007 American Chemical Society)

interpretation and identify image features related to other chemical species

(e.g., salt ions and detergent molecules from the print buffer, organic species from
the polymer layer, contaminants introduced by the printing process, etc.), the multi-
variate analysis technique of PCA was used for more detailed analyses of the
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 309

Fig. 11.12 Image scores and loadings for PC 1 (a), PC 2 (b), and PC 3 (c) for an unhybridized
microspot (negative-ion images). PC 3 loadings (c) from the negative-ion ToF-SIMS image data
matrix indicate that the halo feature detected in spots by ToF-SIMS imaging is characterized by
Si-containing fragments from the polymer-coated glass substrate. Images are 200 m 200 m
(Reproduced with permission from Ref. [71]. Copyright 2007 American Chemical Society)

ToF-SIMS images. PCA was performed on the ToF-SIMS negative-ion image of

the unhybridized DNA microspot shown in Fig. 11.10a to gain a better understand-
ing of the chemical species related to the halo feature. The first three image
scores and loadings from PCA are shown in Figs. 11.12ac. Principal component 1
(PC 1, Fig. 11.12a) clearly distinguishes the image features corresponding to the
DNA microspot (bright regions) and the substrate (dark regions). The PC 1 loadings
plot (Fig. 11.12a) show that most of the major peaks with positive PC 1 loadings
associated with the printed microspots are phosphate- and nitrogen-containing
DNA fragments, while most of the major peaks with negative PC 1 loadings are
hydrocarbon fragments and silicon-containing species from the substrate polymer
310 C.-Y. Lee et al.

layer. PC 2 (Fig. 11.12b) shows image features that correspond to the salt ions,
including Cl, NaOH, CaO, and so forth (bright regions), and SDS fragments
(dark regions). PC 3 (Fig. 11.12c) captures the image feature that corresponds to the
halo around the probe spot (dark regions). The PC 3 loadings plot (Fig. 11.12c)
indicates that most major peaks with negative PC 3 loadings associated with the
halo in the ToF-SIMS image are silicon- and sulfur-containing fragments possibly
from the polymer-coated glass substrate exposed as a result of polymer-layer damage
due to the microarray printing process or from silicon-containing contaminants
wicking to the outside of the spot upon spotting. Studies are currently underway to
examine DNA microspot heterogeneities in further detail.

11.4 Concluding Remarks

The work on immobilized DNA surface analysis presented in this chapter demon-
strates the utility of XPS, ToF-SIMS, NEXAFS, and SPR for obtaining molecular-
level information about the structure, coverage, and hybridization properties of
immobilized ssDNA in model self-assembled monolayer systems as well as on
commercial microarray substrates. Information gained from these studies has facil-
itated an improved understanding of ssDNA behavior on surfaces by providing
quantitative correlations between immobilized DNA surface properties (i.e., den-
sity and orientation) and the resultant analytes target hybridization efficiency from
simple and complex biological samples. Such analyses are needed to understand
the limitations of current DNA microarray assays and to design and engineer new
DNA-based detection systems to meet biomedical and biosensing technological
goals. Despite this substantial progress in quantifying several aspects of DNA on
surfaces, new surface analytical method innovation is required to obtain even more
challenging parameters relevant to DNA real-time assay. Detection limits, higher
throughput, spatial information, and other surface properties could all be improved.
Additionally, sum frequency generation spectroscopy (SFG) and multiplexed
probe microscopy methods (SPM, AFM) that provide in situ information on DNA
surface and DNADNA interactions in wet aqueous conditions are likely to con-
tribute in the near term to this suite of techniques. Optical waveguide methods
interrogating DNA in complex milieu using creative probe designs will also
contribute [75].

Acknowledgments We gratefully acknowledge support from NESAC/BIO (NIH grant no.

EB-002027) and NIH grant no. EB-001473. NEXAFS studies were performed at the NSLS,
Brookhaven National Laboratory, which is supported by the U.S. Department of Energy, Division
of Materials Science and Division of Chemical Sciences. Dr. D. A. Fischer is thanked for his expert
technical assistance with the NEXAFS experiments. Dr. D. Graham is thanked for assistance with
PCA. We also thank Dr. S. Golledge for expert technical assistance with the ToF-SIMS experi-
ments performed at the Center for Advanced Materials Characterization in Oregon.
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 311

References

1. Pirrung MC. How to make a DNA chip. Angew Chem Int Edit. 2002;41:127789.
2. Wang J. From DNA biosensors to gene chips. Nucleic Acids Res. 2000;28:30116.
3. Sanchez-Carbayo MS, Bornmann W, Cordon-Cardo C. DNA microchips: technical and practi-
cal considerations. Curr Org Chem. 2000;4:94571.
4. Debouck C, Goodfellow PN. DNA microarrays in drug discovery and development. Nat
Genet. 1999;21:4850.
5. Churchill GA. Fundamentals of experimental design for cDNA microarrays. Nat Genet.
2002;32:4905.
6. Yu DH, Blankert B, Vire JC, Kauffmann JM. Biosensors in drug discovery and drug analysis.
Anal Lett. 2005;38:1687701.
7. Zhai JH, Cui H, Yang RF. DNA based biosensors. Biotechnol Adv. 1997;15:4358.
8. Rivas GA, Pedano ML, Ferreyra NF. Electrochemical biosensors for sequence-specific DNA
detection. Anal Lett. 2005;38:2653703.
9. Bloomfield VA, Crothers DM, Tinoco I. Nucleic acids structures, properties, and functions.
Sausalito: University Science Books; 2000.
10. Huang E, Zhou FM, Deng L. Studies of surface coverage and orientation of DNA molecules
immobilized onto preformed alkanethiol self-assembled monolayers. Langmuir. 2000;16:
327280.
11. Wolf LK, Fullenkamp DE, Georgiadis RM. Quantitative angle-resolved SPR imaging of
DNA-DNA and DNA-drug kinetics. J Am Chem Soc. 2005;127:174539.
12. Piliarik M, Vaisocherova H, Homola J. A new surface plasmon resonance sensor for high-
throughput screening applications. Biosens Bioelectron. 2005;20:210410.
13. Shumaker-Parry JS, Aebersold R, Campbell CT. Parallel, quantitative measurement of protein
binding to a 120-element double-stranded DNA array in real time using surface plasmon reso-
nance microscopy. Anal Chem. 2004;76(7):207182.
14. Wark AW, Lee HJ, Corn RM. Long-range surface plasmon resonance imaging for bioaffinity
sensors. Anal Chem. 2005;77:39047.
15. Lazerges M, Perrot H, Zeghib N, Antoine E, Compere C. In situ QCM DNA-biosensor probe
modification. Sens Actuator B-Chem. 2006;120:32937.
16. Su XD, Wu YJ, Knoll W. Comparison of surface plasmon resonance spectroscopy and quartz
crystal microbalance techniques for studying DNA assembly and hybridization. Biosens
Bioelectron. 2005;21:71926.
17. He PA, Xu Y, Fang YZ. A review: electrochemical DNA biosensors for sequence recognition.
Anal Lett. 2005;38:2597623.
18. Love JC, Estroff LA, Kriebel JK, Nuzzo RG, Whitesides GM. Self-assembled monolayers of
thiolates on metals as a form of nanotechnology. Chem Rev. 2005;105:110369.
19. Herne TM, Tarlov MJ. Characterization of DNA probes immobilized on gold surfaces. J Am
Chem Soc. 1997;119:891620.
20. Moses S, Brewer SH, Lowe LB, Lappi SE, Gilvey LBG, Sauthier M, et al. Characterization of
single- and double-stranded DNA on gold surfaces. Langmuir. 2004;20:1113440.
21. Levicky R, Herne TM, Tarlov MJ, Satija SK. Using self-assembly to control the structure of
DNA monolayers on gold: a neutron reflectivity study. J Am Chem Soc. 1998;120:978792.
22. Peterlinz KA, Georgiadis RM, Herne TM, Tarlov MJ. Observation of hybridization and dehy-
bridization of thiol-tethered DNA using two-color surface plasmon resonance spectroscopy.
J Am Chem Soc. 1997;119:34012.
23. Lee C-Y, Gamble LJ, Grainger DW, Castner DG. Mixed DNA/Oligo(ethylene glycol) func-
tionalized gold surfaces improve DNA hybridization in complex media. Biointerphases.
2006;1:8292.
24. Lee C-Y, Gong P, Harbers GM, Grainger DW, Castner DG, Gamble LJ. Surface coverage and
structure of mixed DNA/alkylthiol monolayers on gold: characterization by XPS, NEXAFS,
and fluorescence intensity measurements. Anal Chem. 2006;78:331625.
312 C.-Y. Lee et al.

25. Gong P, Lee C-Y, Gamble LJ, Castner DG, Grainger DW. Hybridization behavior of mixed
DNA/alkylthiol monolayers on gold: characterization by SPR and 32P radiometric assay. Anal
Chem. 2006;78:332634.
26. Peterson AW, Heaton RJ, Georgiadis RM. The effect of surface probe density on DNA hybrid-
ization. Nucleic Acids Res. 2001;29:51638.
27. Stohr J. NEXAFS spectroscopy, vol. 25. New York: Springer; 1992.
28. Hahner G, Kinzler M, Thummler C, Woll C, Grunze M. Structure of self-organizing organic
films a near edge X-ray absorption fine-structure investigation of thiol layers adsorbed on
gold. J Vac Sci Technol A-Vac Surf Films. 1992;10:275863.
29. Gamble LJ, Ravel B, Fischer DA, Castner DG. Surface structure and orientation of PTFE films
determined by experimental and FEFF8-calculated NEXAFS spectra. Langmuir. 2002;18:
21839.
30. Nagayama K, Sei M, Mitsumoto R, Ito E, Araki T, Ishii H, et al. Polarized NEXAFS studies
on the mechanical rubbing effect of poly(tetrafluoroethylene) oligomer and its model com-
pound. J Electron Spectrosc Relat Phenom. 1996;78:3758.
31. Ziegler C, Schedelniedrig T, Beamson G, Clark DT, Salaneck WR, Sotobayashi H, et al.
X-ray-absorption study of highly oriented poly(Tetrafluoroethylene) thin-films. Langmuir.
1994;10:4399402.
32. Castner DG, Lewis KB, Fischer DA, Ratner BD, Gland JL. Determination of surface-structure
and orientation of polymerized tetrafluoroethylene films by near-edge X-ray absorption fine-
structure, X-ray photoelectron-spectroscopy, and static secondary ion mass-spectrometry.
Langmuir. 1993;9:53742.
33. Fujii K, Akamatsu K, Muramatsu Y, Yokoya A. X-ray absorption near edge structure of DNA
bases around oxygen and nitrogen K-edge. Nucl Instrum Methods Phys Res Sect B-Beam
Interact Mater Atoms. 2003;199:24954.
34. Fujii K, Akamatsu K, Yokoya A. Near-edge X-ray absorption fine structure of DNA nucleo-
bases thin film in the nitrogen and oxygen K-edge region. J Phys Chem B. 2004;108:80315.
35. Furukawa M, Fujisawa H, Katano S, Ogasawara H, Kim Y, Komeda T, et al. Geometrical char-
acterization of pyrimidine base molecules adsorbed on Cu(110) surfaces: XPS and NEXAFS
studies. Surf Sci. 2003;532:2616.
36. Kirtley SM, Mullins OC, Chen J, Vanelp J, George SJ, Chen CT, et al. Nitrogen chemical-
structure in DNA and related molecules by X-ray absorption-spectroscopy. Biochim Et
Biophys Acta. 1992;1132:24954.
37. Samuel NT, Lee C-YL, Gamble LJ, Fisher DA, Castner DG. NEXAFS characterization of
DNA components and molecular-orientation of surface-bound DNA oligomers. J Electron
Spectrosc Relat Phenom. 2006;152:13442.
38. Petrovykh DY, Perez-Dieste V, Opdahl A, Kimura-Suda H, Sullivan JM, Tarlov MJ, et al.
Nucleobase orientation and ordering in films of single-stranded DNA on gold. J Am Chem
Soc. 2006;128:23.
39. Crain JN, Kirakosian A, Lin JL, Gu YD, Shah RR, Abbott NL, et al. Functionalization of sili-
con step arrays II: molecular orientation of alkanes and DNA. J Appl Phys. 2001;90:32915.
40. Brockman JM, Frutos AG, Corn RM. A multistep chemical modification procedure to create
DNA arrays on gold surfaces for the study of protein-DNA interactions with surface plasmon
resonance imaging. J Am Chem Soc. 1999;121:804451.
41. Nelson BP, Grimsrud TE, Liles MR, Goodman RM, Corn RM. Surface plasmon resonance
imaging measurements of DNA and RNA hybridization adsorption onto DNA microarrays.
Anal Chem. 2001;73:17.
42. Smith EA, Wanat MJ, Cheng YF, Barreira SVP, Frutos AG, Corn RM. Formation, spectro-
scopic characterization, and application of sulfhydryl-terminated alkanethiol monolayers for
the chemical attachment of DNA onto gold surfaces. Langmuir. 2001;17:25027.
43. Shumaker-Parry JS, Zareie MH, Aebersold R, Campbell CT. Microspotting streptavidin and
double-stranded DNA Arrays on gold for high-throughput studies of protein-DNA interactions
by surface plasmon resonance microscopy. Anal Chem. 2004;76(4):91829.
11 Multi-technique Characterization of DNA-Modified Surfaces for Biosensing 313

44. Peterson AW, Heaton RJ, Georgiadis R. Kinetic control of hybridization in surface immobi-
lized DNA monolayer films. J Am Chem Soc. 2000;122:78378.
45. Peterson AW, Wolf LK, Georgiadis RM. Hybridization of mismatched or partially matched
DNA at surfaces. J Am Chem Soc. 2002;124:146017.
46. Xia N, Shumaker-Parry JS, Zareie MH, Campbell CT, Castner DG. A streptavidin linker layer
that functions after drying. Langmuir. 2004;20:37106.
47. Green RJ, Frazier RA, Shakesheff KM, Davies MC, Roberts CJ, Tendler SJB. Surface plasmon
resonance analysis of dynamic biological interactions with biomaterials. Biomaterials.
2000;21:182335.
48. Pavey KD, Olliff CJ. SPR analysis of the total reduction of protein adsorption to surfaces
coated with mixtures of long- and short-chain polyethylene oxide block copolymers.
Biomaterials. 1999;20:88590.
49. Green RJ, Davies MC, Roberts CJ, Tendler SJB. Competitive protein adsorption as observed
by surface plasmon resonance. Biomaterials. 1999;20:38591.
50. Morgan H, Taylor DM. A surface-plasmon resonance immunosensor based on the streptavidin
biotin complex. Biosens Bioelectron. 1992;7:40510.
51. Daniels PB, Deacon JK, Eddowes MJ, Pedley DG. Surface-plasmon resonance applied to
immunosensing. Sensor Actuator. 1988;15:118.
52. Gong P, Grainger DW. Microarrays: methods and protocols. 2nd ed. Totowa: Humana Press;
2007.
53. Grainger DW, Greef CA, Gong P, Lochhead MJ. Microarrays: methods and protocols. 2nd ed.
Totowa: Humana Press; 2007.
54. Fodor SPA, Read JL, Pirrung MC, Stryer L, Lu AT, Solas D. Light-directed, spatially address-
able parallel cemical synthesis. Science. 1991;251:76773.
55. Gong P, Harbers GM, Grainger DW. Multi-technique comparison of immobilized and hybrid-
ized oligonucleotide surface density on commercial amine-reactive microarray slides. Anal
Chem. 2006;78:234251.
56. Dufva M. Fabrication of high quality microarrays. Biomol Eng. 2005;22:17384.
57. Ramakrishnan R, Dorris D, Lublinsky A, Nguyen A, Domanus M, Prokhorova A, et al. An
assessment of Motorola CodeLink (TM) microarray performance for gene expression profiling
applications. Nucleic Acids Res. 2002;30:e30.
58. Shen G, Anand MFG, Levicky R. X-ray photoelectron spectroscopy and infrared spectroscopy
study of maleimide-activated supports for immobilization of oligodeoxyribonucleotides.
Nucleic Acids Res. 2004;32:597380.
59. Lee CY, Canavan HE, Gamble LJ, Castner DG. Evidence of impurities in thiolated single-
stranded DNA oligomers and their effect on DNA self-assembly on gold. Langmuir.
2005;21:513441.
60. May CJ, Canavan HE, Castner DG. Quantitative X-ray photoelectron spectroscopy and time-
of-flight secondary ion mass spectrometry characterization of the components in DNA. Anal
Chem. 2004;76:111422.
61. Petrovykh DY, Kimura-Suda H, Whitman LJ, Tarlov MJ. Quantitative analysis and character-
ization of DNA immobilized on gold. J Am Chem Soc. 2003;125:521926.
62. Walton J, Fairley N. Characterisation of the Kratos Axis Ultra with spherical mirror analyser
for XPS imaging. Surf Interface Anal. 2006;38:12305.
63. Walton J, Fairley N. Transmission-function correction for XPS spectrum imaging. Surf
Interface Anal. 2006;38:38891.
64. Vohrer U, Blomfield C, Page S, Roberts A. Quantitative XPS imaging new possibilities with
the delay-line detector. Appl Surf Sci. 2005;252:615.
65. Walton J, Fairley N. Quantitative surface chemical-state microscopy by x-ray photoelectron
spectroscopy. Surf Interface Anal. 2004;36:8991.
66. Belu AM, Yang ZP, Aslami R, Chilkoti A. Enhanced TOF-SIMS imaging of a micropatterned
protein by stable isotope protein labeling. Anal Chem. 2001;73:14350.
314 C.-Y. Lee et al.

67. Yang ZP, Belu AM, Liebmann-Vinson A, Sugg H, Chilkoti A. Molecular imaging of a
micropatterned biological ligand on an activated polymer surface. Langmuir.
2000;16:748292.
68. Belu AM, Graham DJ, Castner DG. Time-of-flight secondary ion mass spectrometry: tech-
niques and applications for the characterization of biomaterial surfaces. Biomaterials.
2003;24:363553.
69. Wagner MS, Graharn DJ, Castner DG. Simplifying the interpretation of ToF-SIMS spectra and
images using careful application of multivariate analysis. Appl Surf Sci. 2006;252(19):
657581.
70. Tyler BJ. Multivariate statistical image processing for molecular specific imaging in organic
and bio-systems. Appl Surf Sci. 2006;252(19):687582.
71. Lee CY, Harbers GM, Grainger DW, Gamble LJ, Castner DG. Fluorescence, XPS, and TOF-
SIMS surface chemical state image analysis of DNA microarrays. J Am Chem Soc.
2007;129:942938.
72. Heaton RJ, Peterson AW, Georgiadis RM. Electrostatic surface plasmon resonance: direct
electric field-induced hybridization and denaturation in monolayer nucleic acid films and
label-free discrimination of base mismatches. Proc Natl Acad Sci U S A. 2001;98:37014.
73. Hellweg S, Jacob A, Hoheisel JD, Grehl T, Arlinghaus HF. Mass spectrometric characteriza-
tion of DNA microarrays as a function of primary ion species. Appl Surf Sci. 2006;252:
67425.
74. Hashimoto H, Nakamura K, Takase H, Okamoto T, Yamamoto N. Quantitative TOF-SIMS
imaging of DNA microarrays produced by bubble jet printing technique and the role of TOF-
SIMS in life science industry. Appl Surf Sci. 2004;2312:38591.
75. Moiseev L, Unlu MS, Swan AK, Goldberg BB, Cantor CR. DNA conformation on surfaces
measured by flourescence self-interference. Proc Natl Acad Sci U S A. 2006;103:26238.
Index

A orientation and ordering, 207208

ADME. See Adsorption, distribution -sheets yield chiral signals, 214216
metabolism and excretion (ADME) Amide I vibrational mode, 197198
Adsorption, distribution metabolism and Angle-dependent XPS (ADXPS)
excretion (ADME), 127128 and depth profiling, 1112, 17, 31
ADXPS. See Angle-dependent XPS (ADXPS) instruments, 12
AES. See Auger electron spectroscopy (AES) Aperture, NSOM
AFM. See Atomic force microscopy (AFM) aluminum coating, 231
Ahimou, F., 27 diffraction limit, 227
Alpha()-helix fiber optics, 228
melittin light emerging, 228
antimicrobial activity, 210 metal-coating, 238
fibrinogen coiled-coils, 212213 Appelhans, A.D., 72
Gaussian distribution, 210211 Application modes, AFM
G-protein, 214, 215 amplitude and phase changes,
maximum-entropy function, 211212 oscillation, 262
orientation distribution, 211 cantilever oscillation, 261
orientation, 200201 characteristic image artifacts, 261
Amer, N.M., 259 dynamic mode, 261
Amide I protein signals electronic feedback, 261
advantages, 206207 material properties, nanometer scale, 262
detection, 207 surface roughness/external vibrations,
-helical melittin 260261
antimicrobial activity, 210 Arlinghaus, H.F., 53
fibrinogen coiled-coils, 212213 Artyushkova, K., 30
Gaussian distribution, 210211 Atomic force microscopy (AFM)
G-protein, 214, 215 application modes, 260262
maximum-entropy function, 211212 and biological assays, 30
orientation distribution, 211 biological systems (see Biological
measurements systems, AFM)
crystallization, 209 biophysical properties, 285
FWM and CARS, 208, 209 cellular structures and dynamics, 262
hyperpolarizability, 210 commercial cantilevers, 265
polarization mapping, 210 constant current imaging, 257
protein, -helices and -sheets, depth calibration, 85, 86
208209 designed laser optic, 260

V.S. Smentkowski (ed.), Surface Analysis and Techniques in Biology, 315

DOI 10.1007/978-3-319-01360-2, Springer International Publishing Switzerland 2014
316 Index

Atomic force microscopy (AFM) (cont.) SEs, 181184

and DPPC, 239 SIMS, 187
electronic interaction, 258 probe formation, 177178
EM image, cantilever, 259260 statistical methods, 178
and FIB, 247 Benninghoven, A., 38
force-clamp measurements, 285 Berger, M.J., 179
force-measuring tool, 285 Berman, E.S., 57, 60
frequency spectrum, cantilever, 263 Beta()-sheet
high-resolution imaging, 262 chiral signals, 214216
hydrodynamic effects, 265 orientation, 201202
IBM patent application, 258 peptide tachyplesin, 207
imaging tool, 285 Betzig, E., 238
LE/LC phases, 234, 240 Biesinger, M.C., 5556
local/scanning probe Binnig, G., 256, 258
microscopes, 256 Biointerface engineering, 2930
mass components, 264 Biological structures, NSOM
mechanical stress, 257 bilayers and multilayers
mechanical vibrations, 257 acceptor emission, 241, 242
Ni/Cr layers, 7980 AFM image, 239241
optical detection scheme, 259 cellular function, 240
oxide-sharpening technique, 260 donor fluorophore, 241, 243
resonance curve, 263 DPPC, 239, 240
silicon nitride, 241, 247 electric fields, 238, 240
standard detection scheme, 259 fiber-optic probe, 242
STM imaging, 256 FRET, 240242
time-dependent external forces, 263 high-resolution fluorescence, 240
unbombarded sugar trehalose, 8081 LangmuirBlodgett (LB)
Attenuated total reflectance Fourier transform technique, 239
IR spectroscopy (ATR-FTIR) microdomain structures, 240
amide I band, 206 PMMA, 241
-helical melittin (see Alpha-helix) rhodamine, 241
PSLB, 216 single-molecule fluorescence image,
Auger electron spectroscopy (AES), 2 238, 240
biological films and membranes, 233
cells
B confocal images, 245
Backscattered helium ions, 184, 185 DCs, 245
Backscatter electrons, 172 F-actin fibers, 244
Barber, M., 107 fluorescence, 243, 244
Beamspecimen interaction HASM, 243, 244
beam penetration HLA protein, 244
electron-greedy, 180 human fibroblast, 244
energy transfer, 178 lectins, 245
Monte Carlo modeling, 180181 MHC, 244
stopping power, 178180 microdomains, 244
Monte Carlo computer simulations, 178 plasma membrane, 244
particles and detectors polarization, 245
backscattered helium, 184185 topography, 243, 244
cathodoluminescence (CL) effect, 186 fluorescence, 233
detected photons, 185, 186 monolayers (see Lipid monolayers)
imaging modes, 188 tipsample interactions (see Tipsample
mass spectrometry techniques, 187 interactions)
MCP, 185, 186 topography, 233
myelin sheets, 187188 Biological systems
secondary ions and neutrals, 187 biotinavidin, 265
Index 317

electrophysiology, 278284 Cross, S.E., 274

intermolecular forces, 266267 Curie, J., 256257
intramolecular forces, 267273 Curie, P., 256257
MD computer simulations, 265
molecular structures, 265
and optical microscopy, 273278 D
Biomedical devices, 3 Davies, J., 3
Biomolecular analysis, ToF-SIMS. DCA. See Dynamic contact angle (DCA)
See Time-of-flight secondary 3D chemical imaging. See Three-
ion mass spectrometry dimensional (3D) chemical
(ToF-SIMS) imaging
Biosensing, DNA-modified surfaces. DCs. See Dendritic cells (DCs)
See DNA-modified surfaces de Bakker, B.I., 245
Boughton, A.P., 195218 Delcorte, A., 59, 60, 119
Breitenstein, D., 3763 Dendritic cells (DCs), 245
Briggs, D., 142, 157 Deoxyribonucleic acid (DNA)
Brunelle, A., 56 elemental compositions, 24, 25
lipids, mucins and enzymes, 2425
modified surfaces, biosensing
C (see DNA-modified surfaces)
Cantilever Depth profiling
constant-height mode, 261 and ADXPS, 1112
mass components, 264 chemical distribution, 17
mechanical properties, 262 coronene, C 1s spectra, 18
optical detection scheme, 259 electron impact (EI) ion source, 54
split photodiode, 259 measurements, 79
tapping mode, 262 molecular, 8184, 87, 91
Caprioli, R.M., 126 sputtering, 48
Castaing, R., 105 ToF-SIMS, 38, 58
Castner, D.G., 21, 289310 XPS, 19
Caudodorsal cell hormone Dickenson, N.E., 225250
(CDCH), 122 3 Dimensional analysis
Chemical composition amino acids and phospholipids, 59
MIC, 29 glutardialdehyde-fixated NRK cells,
SIMS imaging, 73 4950, 52, 59
XPS, 27 organic samples, 62
Cheng, J., 58 Dipalmitoylphosphatidycholine (DPPC)
Chen, Z., 195218 bilayer, AFM image, 239240
Chichester, R.J., 238 high-resolution NSOM fluorescence,
Cluster ion source 234, 235
beam characteristics, 46 LE and LC phases, 234
chromatic aberration, 46 monolayer, fluorescence, 233234
EI ion source, 44 RDS, 236
G-SIMS, 167 DNA. See Deoxyribonucleic acid (DNA)
ion-optical system, 45 DNA-modified surfaces
LMIS, 4344 adeninethymine (AT) and cytosine
separation and pulsing, 4445 guanine (CG), 291
Cluster SIMS. See Secondary ion mass hybridization, 293
spectrometry (SIMS) microarrays, 291292
Collision cascades, 3940 structure and hybridization properties
Conboy, J.C., 216 hybridization, 300301
Constant force mode, 261 immobilization, 294, 295
Constant height mode nitrogen K-edge NEXAFS
faster imaging, 257 spectra, 297
imaging cells, 273 SAMs, 294
318 Index

DNA-modified surfaces (cont.) micromechanical measurements, 283

SPR, 298, 299 sound-pressure level, 282283
ssDNA, 293, 298 transduction current measurements, 283
XPS, 294, 296 Electrospray-deposition (ESD) system, 121
X-ray absorption, 297 Elemental composition, 24, 25
surface analytical techniques, 293 Erickson, E.S., 225250
surface chemical state image analysis Evans-Nguyen, K.M., 204
categories, 301
fluorescence microscopy, 305, 307
hybridization efficiency, 305 F
image scores and loadings, 309310 Fairley, N., 30
micro-printing techniques, 302 Fartmann, M., 57
microspots, 304 FIB. See Focused ion beam (FIB)
negative-ion ToF-SIMS ROI spectra, Field ion microscope (FIM), 176
305, 306 Fluorescence
XPS and ToF-SIMS, 302303, 306, 308 DNA hybridization, 293
DPPC. See Dipalmitoylphosphatidycholine microscopy, 305, 307
(DPPC) NSOM
Drug metabolism confocal, 234, 236, 237
IMS protocols and instruments, 100 DCs, 245
MRMbased imaging, 129 F-actin fibers, 244
profile and ADME, 127128 high spatial resolution, 234
signal intensities, 128 images, 234, 240
Duke, C.B., 2 lipid monolayer, 230
Dunn, R.C., 225250 model membranes, 233
Dynamic contact angle (DCA), 3 phalloidin, 244
tetramethylrhodamine, 244
radioactive, 303
E Fluorescence resonance energy transfer
Easton, C.D., 931 (FRET), 240242
Eckert, R., 267 Focused ion beam (FIB), 247
Edidin, M., 233, 244 Formalin-fixed and paraffin-embedded (FFPE)
Electron impact (EI) ion source, 44 tissue, 113
Electron spectroscopy for chemical analysis Fourier transform ion cyclotron resonance
(ESCA), See also X-ray (FT-ICR), 109, 111112
photoelectron spectroscopy Fragmentation pathway
(XPS), 10, 294 G-SIMS (see G-SIMS fragmentation
Electrophysiology pathway mapping
single cells and excised membrane (G-SIMS-FPM))
patches SMILES (see Simplified molecular input
electromechanical motors, 279280 line entry specification (SMILES))
optical transparent plastic plates, 279 FRET. See Fluorescence resonance energy
single-ion-channel recording, 278 transfer (FRET)
voltage-sensitive ion channels, 279 Friction mode, 261
tissue sections FT-ICR. See Fourier transform ion cyclotron
cochlear hair cells, inner ear, 280 resonance (FT-ICR)
DIC, 280 Full-width half-maximum (FWHM), 230
differential interference contrast FWHM. See Full-width half-maximum
(DIC), 280 (FWHM)
force interaction, 281
force transmission, 281
mechanically independent G
stereocilia, 282 Gamble, L.J., 289310
mechanical stimulus, 282 Garcia-Parajo, M.F., 245
mechanosensitive ion channels, 284 Gardella, J.A., 60
Index 319

Gas-phase basicity (GB), 105 mass scale and instrument optimization, 156
Gasteiger, J., 160 positive-and negative-ion data, 158
Gentle secondary ion mass spectrometry SIMS emission process, 157
(G-SIMS) surface plasma temperature, 157, 158
bovine serum albumin, 154 G-SIMS imaging, 166167
cholesterol, caffeine and poly-l-lysine, Guerquin-Kern, J.-L., 38
154, 155
drug-delivery systems and biomedical
implants, 152 H
energy distribution, excited atoms, 145 Hagenhoff, B., 3763
fingerprint approach, 148 Harbers, G.M., 289310
folic acid, molecular structure, 154, 155 HASM. See Human arterial smooth muscle
fragmentation process, 143 (HASM)
G-SIMS-FPM (see G-SIMS fragmentation Heeren, R.M.A., 99133
pathway mapping (G-SIMS-FPM)) Helium ion microscope (HIM)
hydrocarbon molecule, series CnH2n, 146 Acidovorax, 172, 173
ion beam parameters, 150151 beam energies, 173
Irganox 1010, 152, 153 beamspecimen interaction, 178188
low surface plasma temperature, 147 detectable particles and detector, 174
mass-scale calibration, 156 imaging applications, biology, 192
mass spectra, 142 inherent surface specificity, 172
molecular fragments, 146 ion source and ion column, 174
multivariate methods, complementarity, 143 probe formation, 177178
optimization, instrument parameters, 150 sample charging and preparation
partition function relation, 146 characteristics, 189
80:20 PLGA co-polymer, 153, 154 imaging biological specimens, 188
polycarbonate (PC) and Irganox 1010, intercalated and principal cells,
166167 kidney, 191
polycyclic aromatic hydrocarbon ions, 152 rat kidney preparation, 191
polymer industry, 142 ultrastructure, mouse tooth enamel,
polystyrene (PS) spectral intensities, 189, 190
147149 signal-to-noise ratio, 174, 175
secondary ion formation, 153 source
spectra acquisition, 151152 electron and gallium ion beams., 176
surface-energy distribution, 145 nanometer-level precision, 177
ToF systems, 154 ultrahigh-vacuum (UHV) vessel,
Giessibl, F.J., 258 175, 176
Gillen, G., 82 Hillenkamp, F., 103
Gilmore, I.S., 141168 Hookes law, 258
Goetze, B., 171192 Hrber, J.K.H., 255285
Gong, P., 289310 Human arterial smooth muscle (HASM),
G proteins, 201, 214, 215 243, 244
Graham, D.J., 60
Grainger, D.W., 289310
Green, F.M., 141168 I
G-SIMS. See Gentle secondary ion mass Identification and interpretation
spectrometry (G-SIMS) G-SIMS fragmentation pathways, 165166
G-SIMS fragmentation pathway mapping mass spectrum, 168
(G-SIMS-FPM) Imaging
chemical permutations, 156 chemical processes, 172
folic acid, gmax values, 158, 159 high-resolution imaging techniques, 172
and G-SIMS SMILES, 144 high-resolution surface analysis, 172
ion-trap systems, 157 HIM (see Helium ion microscope (HIM))
10-keV argon and cesium ions, 157, 158 STEM, 172
low-energy collision, 156 surface specificity, 171
320 Index

Imaging (cont.) -helices, spectrin, 271, 272

ToF-SIMS isothermal restoring force, 270271
applications, 58 modular structure, proteins, 267
epithelial cells, EthDI, 51, 54 protein-folding and-unfolding pathways,
LB films and SAM, 5556 267, 271
semiquantitative interpretation, 54 recombinant DNA techniques, 269
single cells, 57 spectrin, mechanical behavior, 273
tissue sections, 5657 thermodynamic studies, 271
Imaging mass spectrometry (IMS) Ion-mobility mass spectrometry (IMMS), 112
applications, 109 Ion scattering spectroscopy (ISS), 2
biological surface analysis, 100101 Isotope labeling, 205206
compositions and structures, surfaces, 100 ISS. See Ion scattering spectroscopy (ISS)
description, 132133
detection efficiencies, 109
FT-ICR, 111112 J
IMMS, 112 Joens, M.S., 192
instruments, 111113
ion-optical microscope elements, 112
MALDI (see Matrix-assisted laser K
desorption/ionization (MALDI)) Karas, M., 103
mass analyzer characteristics, 109 Kozole, J., 7197
microprobe and microscope modes, 102, 103 Kulp, K.S., 60
molecular and spatial resolution, 99
molecular imaging technique, 99
proteins, lipids and carbohydrates, 120 L
publications, 100 Langer, M.G., 284
silicon wafer masks, microstructures, 113 LangmuirBlodgett (LB) films
SIMS (see Secondary ion mass hydrogen atoms, 55
spectrometry (SIMS)) lipid mixtures, 55
ToF, 110111 Survanta monolayer, 236, 237
Immobilized temperature, 56
DNA molecules, 292, 305 Lee, C.-Y., 25, 289310
ssDNA probe, 292 LE/LC phases. See Liquid-expanded/
IMMS. See Ion-mobility mass spectrometry condensed (LE/LC) phases
(IMMS) Leone, L., 27
IMS. See Imaging mass spectrometry (IMS) Leufgen, K.M., 55
Interfaces, SFG Levi-Setti, R., 57
amide I protein signals (see Amide I Lewis, K.B., 12
protein signals) Lipid imaging
CH and NH spectral signals single cells, 9192
amino acids and peptides, 205 tissue imaging (see Tissue imaging)
BSA, 204 Lipid monolayers
hydrophobic groups, 204 AFM, 233235
pH values, 205 buckling, 237
proteinprotein interactions, 205 cellular membrane fractions, 235
isotope labeling, 205206 confocal fluorescence, 233234, 236, 237
nucleic acids, 217 DPPC lipid monolayer, 233235
PSLB, 216217 electric fields, 238
Intramolecular forces far-field images, 237
double-cantilever AFM, 268269 fluorescence image, 233, 236
energy scheme, pathway, 272 GM1 rafts, 235236
forceextension relationship, spectrin, LB films, 233, 234, 236, 237
269270 LE/LC, 233235
Gaussian distributions, 271 lipid rafts, 234, 235
Index 321

microdomains, 234, 235 specific mass spectrometry imaging,

PMMA, 238 129, 130
RDS, 236 system control, 104
single-molecule fluorescence image, 238 thermodynamics, reaction, 105
spatial resolution, 234 Matrix-enhanced SIMS (ME-SIMS)
sphingomyelin-cholesterol-DOPC, 235, 236 dried droplet deposition, 116
topography, 234, 235, 237 glycerol matrices, 107
transition dipole moment, 239 ionization enhancement, 118
Lipidomics, 101, 113 lipids and peptides, 108109
Lipids Lymnaea stagnalis nervous tissue, 121, 123
enzymes, and DNA, 2425 vs. MALDI, 121, 122
and mucins, 10 matrices and applications, 115, 117
Liquid-expanded/condensed (LE/LC) phases, surface analysis, 121
234, 240 May, C.J., 53, 60
Liquid metal ion gun (LMIG), 72, 121 McArthur, S.L., 931
Liquid metal ion source (LMIS) McLafferty, F.W., 40
advantages, 43 MCP detector. See Microchannel plate (MCP)
gold and bismuth sources, 4344 detector
needle, 43 MD simulations. See Molecular dynamics
parameters, 43 (MD) simulations
Liu, J., 216 Medium energy ion spectroscopy, 180
Liu, Z.C., 25 Melittin
LMIG. See Liquid metal ion gun (LMIG) antimicrobial activity, 210
LMIS. See Liquid metal ion source (LMIS) fibrinogen coiled-coils, 212213
Lou, L., 29 Gaussian distribution, 210211
G-protein, 214, 215
maximum-entropy function, 211212
M orientation distribution, 211
Mahoney, C.M., 58 Metal-assisted SIMS (MetA-SIMS)
Major histocompatibility complex (MHC), 244 assistance, chemical modifications, 125
MALDI. See Matrix-assisted laser desorption/ cellular localization, 121, 124
ionization (MALDI) enhanced secondary ion signal, 107
Matrix-assisted laser desorption/ionization gold islands, migration, 121122
(MALDI) SSIMS, 108
carbohydrates and metabolites, plants, surface analysis, 121
129130 Meyer, E., 259
cluster SIMS, 87 MHC. See Major histocompatibility complex
drug distribution and quantification, (MHC)
127129 MIC. See Microbiological corrosion (MIC)
element/metal detection, 131 Michel, R., 53
HCCA classical vs. HCCA/ANI ionic Microarrays
matrix, 105, 107 and biosensor surfaces, 293
image reconstructions, 105 DNA hybridization, 293
imaging, lipids, 130, 131 micro-printing techniques, 302
ion-formation mechanisms, 104105 Microbiological corrosion (MIC), 29
ionic matrices, 105 Microchannel plate (MCP) detector, 185, 186
laser irradiation, 104 Milillo, T.M., 60
laser performance and stability, 104 Mishra, G., 931
mass spectral analysis, 105 Molecular depth profiling
matrices and structures, 105, 106 SIMS
microprobe mode, 103104 advantages, cluster projectiles, 82
microscope, 111112 chemical damage, 8182
peptide and protein detection, 126 C60+ primary ion fluence, PLA, 81, 82
proteomics and clinical proteomics, 126127 description, 8384
322 Index

Molecular depth profiling (cont.) shear-force feedback, 229

parameters, 83 spatial resolution, 226, 227
polyatomic projectiles, 82 SPM and TEM, 226
SF5+ primary ion fluence, 82, 83 tissue and cellular structure, 225
variables, 84 Nonspecific binding, 298, 301
ToF-SIMS Notte, J., 171192
EI ion source, 44 Nucleic acids
layered polymer system, 49, 51, 58 biosensors, 300
LB films, 58 hybridization, 292
primary ions, 48 probes, 302
sputter erosion, 48 Nygren, H., 56
Molecular dynamics (MD) simulations
Ag solid crystal, 15-keV Ga, 73, 74
Au atoms, 76 O
C atom, 15-keV C60, 7374 Obst, M., 172
5-keV Au3 and 5-keV C60, 7576 Odom, R.W., 107108
physical damage, 79 Ogaki, R., 152
Molecular structure Operational modes, ToF-SIMS
G-SIMS (see Gentle secondary ion mass data collection, 49
spectrometry (G-SIMS)) depth profiling, 58
SMILES (see Simplified molecular input 3 dimensional analysis, 4950, 52, 5859
line entry specification (SMILES)) imaging
Molecular weight applications, 58
biochemical properties, 120 epithelial cells, EthDI, 51, 54
secondary-ion images, polymer, 110 LB films and SAM, 5556
Monroe, E.B., 5657 semiquantitative interpretation, 54
Mooren, O.L., 225250 single cells, 57
Multiple reaction monitoring (MRM)based tissue sections, 5657
imaging, 129 intensity, sputtering, 49, 51
Multi-technique, DNA-modified surfaces. mass-spectrometric techniques, 49, 50
See DNA-modified surfaces spectrometry
Multivariate statistical analysis, 302 DNA oligomers, 53
DPPC, 50, 53
enzyme activity, 54
N proteins, 53
Near-field scanning optical microscopy Optical microscopy See also Near-field
(NSOM) scanning optical microscopy
and AFM, 226 (NSOM)
Airy disk pattern, 227 biological process, 276277
aperture, 228 cell membrane, exocytosis, 276
biological applications (see Biological cell physiology-measuring techniques, 274
structures, NSOM) cell suspension, 276
chemical etching, 228, 229 conventional piezotube scanner, 273
and DPPC, 230 exocytosis, progeny virus, 277
fiber optics, 228 live cells, physiological conditions, 274, 275
and FIB instrument, 228, 229 membrane dynamics, 276
fluorescence, 230 modulation measurements, 274
and FWHM, 230 phagocytosis/cell movement, 276
hydrofluoric acid (HF), 228 piezotube scanner, 275
light passing, 227, 228 position-sensitive quadrant
numerical aperture, 227 photodetector, 275
optical measurements, 227 virus reproduction, 276277
probes, sample heating, 230233 Optical waveguide lightmode spectroscopy
and SEM, 228, 229 (OWLS), 9
Index 323

Orientation analysis, SFG Proteomics

CH groups MALDI mass-spectrometric imaging, 127
advantage, 198 microspot deposition, matrix, 126
fitted signal intensity, 199 microspotted matrix depositio, 126
methyl groups, 199200 spraying system, 126
protein secondary structures, 200 PSLB. See Planar-supported lipid bilayer
-helix, 200201 (PSLB)
-sheet, 201202
OWLS. See Optical waveguide lightmode
spectroscopy (OWLS) Q
Quantitative analysis, XPS, 13, 30
Quartz crystal microbalance (QCM), 9, 7677
P
Pacholski, M.L., 3738
Paynter, R.W., 11, 17, 21 R
PCA. See Principal component analysis (PCA) Radial distribution chamber (RDC), 14, 15
Physiological conditions, AFM. See Optical Ratner, B.D., 11, 12, 17, 21
microscopy Respiratory distress syndrome (RDS), 236
Planar-supported lipid bilayer (PSLB) Richter, K., 56
advantage, isotope-labeled lipids, 216 Rohrer, H., 256
asymmetric bilayers, 216217 Rutherford backscatter spectroscopy, 180, 184
LangmuirBlodgett and Langmuir Rutherford, E., 180, 184
Schaefer techniques, 216
PMMA. See Polymethyl-methacrylate
(PMMA) S
Polyatomic ion source, 38, 72 SAMs. See Self-assembled monolayers
Polymethyl-methacrylate (PMMA), 241 (SAMs)
Principal component analysis (PCA) Scanning electron microscopy (SEM)
DNA microspot, 309 charging effects, 172
ToF-SIMS negative-ion image, 309 granules, aluminum coating, 228, 229
Probes high gas pressures, 188
DNA/thiol diluent, 298 high-resolution imaging instrument, 172
hybridization, 301 imaging biological specimens, 188
NSOM, sample heating sample preparation protocols, 191
catastrophic damage, 232, 233 SEs, 181, 182
chemical etching technique, 232 Scanning near-field optical microscopy
electron micrographs, 232, 233 (SNOM), 5
emission spectrum, 231232 Scanning probe microscopy (SPM)
fiber optics, 231 biological structure and function, 226
laser power, 232, 233 constant current imaging, 257
N-allyl-N-methylaniline, 231, 232 spatial resolution, 226
output power, 232 Scanning transmission electron microscope
perylene, 231, 232 (STEM), 172
power technology, 232 Schnieders, A., 3763
thermal expansion, 232, 233 Schultz, J.A., 130
thermochromic polymer, 231, 232 Schwieters, J., 110
nucleic acid, 302 Seah, M.P., 141168
ssDNA, 292 Secondary electrons (SEs)
Proteins and peptides characteristic bright-edge effect, 182, 184
amide I protein signals (see Amide I collagen fibers, 182
protein signals) EverhartThornley detector, 182, 183
high-resolution XPS C 1s spectrum, 21, 23 helium-induced, 182
surface nitrogen, 21, 22 image formation, 181
XPS, 21 type-one secondary electrons (SE1), 182
324 Index

Secondary ion mass spectrometry (SIMS) archetypal molecular entities, 165

Au3+/Bi3+ and C60+ projectiles, 8687 ASCII text string, 161
biological tissue (see Tissue imaging) Daylight Chemical Information Systems, 161
chemical sensitivity, 4 DENDRAL project, 160
cluster ion sources, 72 description, 161
description, 7273 folic acid text, 162
3D image, depth calibration, 8586 fragmentation mass and level, folic acid, 163
2D image, pixel size, 8485 informatics, 161
energetic (keV) atomic projectiles, 72 mass-based tree structure, 163
enhancement, molecular ion yield, 106107 mass spectra prediction, 160
gas-phase ion sources, 72 secondary ion intensities, 164
high-spatial-resolution analytical SIMS library spectra, 164
microscopy, 105 simulated fragmentation pathways, 161162
ion optics, 105 syntax and grammar, 161
MALDI, 87 valine and tyrosine, 164
matrix-enhanced, 107108 SIMS. See Secondary ion mass spectrometry
MD simulations (see Molecular dynamics (SIMS)
(MD) simulations) Simulated spectra
metal-assisted, characteristics, 108109 fragmentation library, 168
molecular depth profiling, 8184 SMILES fragmentation pathway, 164
physical damage Single-cell imaging
ion beam, 79 fusion pores, 91
nickel/chromium (Ni/Cr) layers, 7980 peptides and lipids
sugar trehalose, AFM, 8081 biological applications, 125
surface sensitivity, 79 cellular localization, MetA-SIMS,
postionization, 96 121, 124
sample preparation, 9596 direct molecular imaging, Lymnaea
single cells (see Single-cell imaging) stagnalis nervous tissue, 121, 123
surface analysis and imaging, 106 gold deposition, 124
surface metallization and matrix lateral resolution analyte diffusion,
deposition, 119120 121122
ToF analyzer, 105 MALDI stigmatic imaging, rat brain
ToF-SIMS (see Time-of-fight secondary tissue, 125
ion mass spectrometry (ToF-SIMS)) ME-SIMS vs. MALDI, 121, 122
yield enhancement surface-enhanced SIMS, 121
LangmuirBlodgett techniques, 77 surface metallization, 121
neutrals, 77 Tetrahymena, 9394
organic materials, 77 Xenopus laevis oocyte, 9495
PS-200 and Gramicidin D, 7778 Single molecule detection
QCM, 15-keV C60, 7677 lipid monolayer, 230, 238
Self-assembled monolayers (SAMs), 55, polarization, 245
294, 297 Single-stranded DNA (ssDNA).
SEM. See Scanning electron microscopy see DNA-modified surfaces
(SEM) Sjvall, P., 57
SERS. See Surface-enhanced Raman Slodzian, G., 105
scattering (SERS) Small spot spectroscopy, 1920
SFG vibrational spectroscopy. See Sum Smentkowski, V.S., 15, 60
frequency generation (SFG) SMILES. See Simplified molecular input line
vibrational spectroscopy entry specification (SMILES)
Shubeita, G.T., 241 SNOM. See Scanning near-field optical
Sigmund, P., 39 microscopy (SNOM)
Signor, L., 128 Somorjai, G.A., 204
Simplified molecular input line entry Sostarecz, A.G., 55, 58
specification (SMILES) SPM. See Scanning probe microscopy (SPM)
Index 325

SPR. See Surface plasmon resonance (SPR) Tapping mode, 262

SSIMS. See Static secondary ion mass Tarlov, M.J., 294
spectrometry (SSIMS) TEM. See Transmission electron microscopy
Static secondary ion mass spectrometry (SSIMS) (TEM)
bovine serum albumin samples, 154 Thompson, C.E., 60
degraded fragment products, 156 Thompson, M.W., 39
dominant intensity ions, 148 Three-dimensional (3D) chemical imaging
polyamino acid poly-L-lysine, 154 depth calibration, 8586
polystyrene (PS), 148, 149 Xenopus laevis oocyte cell, 9495
Stauber, J., 99133 Time-of-flight secondary ion mass
Stephens, W.E., 110 spectrometry (ToF-SIMS)
Stokes, G.V., 277 charge compensation, 4748
Stylus, 226, 247 data quality
Sum frequency generation (SFG) vibrational secondary ionization probability,
spectroscopy 5960
amide I protein signals (see Amide I statistical evaluation, 6061
protein signals) depth profiling, 48
blood coagulation, 195 description, 38
description, 217218 development, 3738
experimental procedures DNA hybridization, 306
components, 202203 formation
DR-SFG signal intensity, 203204 cluster primary ions, 4142
instruments, 202 collision cascades, 3940
near total internal reflection, 203 GCIB sources, 62
surface sensitivity, 203 G-SIMS-FPM, 156, 157
interfaces, biological molecules ion detection, 47
(see Interfaces, SFG) life science applications, 63
lipid membranes, 196 operational modes (see Operational modes,
nonlinear optical process ToF-SIMS)
ATR-FTIR probes, 197 primary ions (see Cluster ion source)
CH stretching modes, 198 pumping rate, vacuum system, 61
materials, 196197 quantification, 6162
SERS, 198 separation process, 4647
vibrational modes, FTIR, 197 and XPS, 293, 302
orientation analysis (see Orientation Tipsample interactions
analysis, SFG) AFM, 247249
surface-sensitive in situ probes, 196 biological tissues, 246
Surface analysis instrumentation FIB, 247249
applications, 2 fiber-optic, 246
biological sample analysis, 2, 3 fragile samples, 246
instruments and data analysis protocol, 5 HASM cell, 248, 250
sample preparation methods, 5 hybrid probe, 248, 250
Surface-enhanced Raman scattering (SERS), 198 interferometric signal, 246
Surface metallization (MetA-SIMS), 108109 model membranes, 246
Surface plasmon resonance (SPR) noninvasive imaging, 246, 247
DNA hybridization, 294 refractive index, 248
measurement, 298, 299 silicon nitride probes, 247
and spectroscopy, 293 thin acetate matrix, 248, 249
Surface roughness, 260261 topography, 248
Tissue analysis
components, 25
T hard tissue, 2829
Tachyplesin I, 215 microbial cells, 27, 28
Tandem mass spectrometry (MS), 128, 132 UHV technique, 2728
326 Index

Tissue imaging Vickerman, J.C., 142, 157

cellular membranes, 88 Vroman, L., 206
human atherosclerotic plaque, 91
in vivo analysis, 88
mouse brain W
anatomic structures, 8889 Wagner, M.S., 21, 53
lipid heterogeneity, 89, 90 Walton, J., 30
phosphocholine and cholesterol, 90 Wang, J., 207, 210, 212, 215
raclopride drug, 9091 Weininger D., 161
Tissue surfaces Winograd, N., 3738, 58, 7197
biodistribution and metabolism, 132 Woods, A.S., 130
enhanced-surface IMS techniques, 132, 133 Worm-like chain (WLC) mode, 270
MALDI (see Matrix-assisted laser Wucher, A., 42, 85
desorption/ionization (MALDI)) Wu, K.J., 107108
metallization and matrix-deposition, 119120
SIMS (see Secondary ion mass
spectrometry (SIMS)) X
spatial resolution, 132 XPS. See X-ray photoelectron spectroscopy
surface modifications (XPS)
matrix deposition, 115118 X-ray photoelectron spectroscopy
metal deposition, 118119 (XPS)
organic solvent treatment, 115 and ADXPS, 1112
tissue collection analytical techniques, 293
biological samples, 114115 biointerface engineering, 2930
fast frozen, 114 biomolecules
frozen and FFPE, 113, 114 lipids, mucins, enzymes and DNA,
trypsin digestion, 132 2426
ToF-SIMS. See Time-of-flight secondary ion proteins and peptides, 2124
mass spectrometry (ToF-SIMS) elements, 1011
Topography and ESCA, 10
immunofluorescence, 226 freeze hydration, 12
microdomains, 244 imaging and mapping, 13
semicircular domains, 234 instrumentation
Transmission electron microscopy (TEM), charge neutralization, 1517
172, 226 depth profiling, 1719
Tunneling imaging and mapping, 2021
quantum-mechanical process, 256 organic and biological
vacuum gap, 256 systems, 13
Turecek, F., 40 RDC, 14, 15
Tyler, B.J., 60 small spot spectroscopy, 1920
QCM and AFM, 9
SPR and OWLS, 9
U surface sensitivity, 11
Ultrahigh-vacuum (UHV), 10, 13, 14, 28 and tissue analysis (see Tissue analysis)
Urbassek, H.M., 40 and ToF-SIMS, 302
and UHV, 10

V
Vaidyanathan, S., 57, 59 Z
Van der Waals force, 258, 261 Zeitvogel, F., 172