Vdoc - Pub Understanding Nanomaterials

understanding
nanomaterials
Malkiat S. Johal
understanding
nanomaterials
understanding
nanomaterials
Malkiat S. Johal
Boca Raton London New York
CRC Press is an imprint of the

Taylor & Francis Group, an informa business
Front cover image was provided by Dr. Gerry Ronan, Farfield Group, Manchester, United Kingdom.
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2011 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works

Version Date: 20110720
International Standard Book Number-13: 978-1-4398-9112-4 (eBook - PDF)
This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors and
publishers have attempted to trace the copyright holders of all material reproduced in this publication
and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced,
transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying, microfilming, and recording, or in any information stor-
age or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copy-
right.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222
Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that pro-
vides licenses and registration for a variety of users. For organizations that have been granted a pho-
tocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
This book is dedicated to my
beautiful daughter Simran
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
About the Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
Chapter 1 A Brief Introduction to Nanoscience . . . . . . . . . . . 1

1.1 The Need for Nanoscience Education . . . . . 1
1.2 The Nanoscale Dimension and the
Scope of Nanoscience . . . . . . . . . . . . . . . . . . 2
1.3 Self-Assembly . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Supramolecular Science . . . . . . . . . . . . . . . . 6
1.5 Sources of Information on
Nanoscience . . . . . . . . . . . . . . . . . . . . . . . . . 7
Chapter 2 Intermolecular Interactions and

Self- Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Chapter Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Intermolecular Forces and
Self- Assembly . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Ion-Ion Interactions . . . . . . . . . . . . . 11
2.1.2 Ion-Dipole Interactions . . . . . . . . . . 13
2.1.3 Dipole-Dipole Interactions . . . . . . . 16
2.1.4 Interactions Involving Induced
Dipoles . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.5 Dispersion Forces . . . . . . . . . . . . . . 20
2.1.6 Overlap Repulsion . . . . . . . . . . . . . 21
2.1.7 Total Intermolecular Potentials . . . 24
2.1.8 Hydrogen Bonds . . . . . . . . . . . . . . . 25
2.1.9 The Hydrophobic Effect . . . . . . . . . 27
2.2 Electrostatic Forces Between Surfaces:
The Electrical Double Layer . . . . . . . . . . . . 28
2.2.1 The Electrical Double Layer . . . . . . 28
2.2.2 The Debye Length . . . . . . . . . . . . . . 31
2.2.3 Interactions Between Charged
Surfaces in a Liquid . . . . . . . . . . . . 32
2.3 Intermolecular Forces and Aggregation . . 34
2.4 Simple Models Describing Electronic
Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
vii
viii Contents
2.4.1 The Particle in a Box Model . . . . . . 37

2.4.2 Conjugation in Organic
Molecules . . . . . . . . . . . . . . . . . . . . 43
2.4.3 Aggregation and Electronic
Structure . . . . . . . . . . . . . . . . . . . . . 47
2.4.4 π-π Stacking Interactions . . . . . . . . 49
References and Recommended Reading . . . . . . . 49
End of Chapter Questions . . . . . . . . . . . . . . . . . . 51
Chapter 3 Rudiments of Surface Nanoscience . . . . . . . . . . . 55

Chapter Overview . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.1 Fundamentals of Surface Science . . . . . . . 55
3.1.1 The Surface Energy of Solids
and Liquids . . . . . . . . . . . . . . . . . . . 55
3.1.2 Surface Free Energy of
Adsorbed Monolayers . . . . . . . . . . . 56
3.1.3 Contact Angles and Wetting
Phenomena . . . . . . . . . . . . . . . . . . . 61
3.1.4 Nanomaterials and
Superhydrophobic Surfaces . . . . . . 62
3.2 Adsorption Phenomena: Self-
Assembled Monolayers . . . . . . . . . . . . . . . . 68
3.2.1 Simple Adsorption Isotherms . . . . 73
3.2.2 Other Useful Adsorption
Isotherms . . . . . . . . . . . . . . . . . . . . . 78
3.3 Surfactant Chemistry . . . . . . . . . . . . . . . . . 79
3.3.1 Micelle and Microemulsion
Formation . . . . . . . . . . . . . . . . . . . . 82
3.3.2 The Determination of Surface
Excess: The CMC and the Cross-
Sectional Area per Molecule . . . . . 84
References and Recommended Reading . . . . . . . 88
End of Chapter Questions . . . . . . . . . . . . . . . . . . . 89
Chapter 4 Characterization at the Nanoscale . . . . . . . . . . . 97

Chapter Overview . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.1 Surface Tensiometry: The Surface
Tensiometer . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2 Quartz Crystal Microbalance . . . . . . . . . . 101
4.2.1 The Piezoelectric Effect . . . . . . . . 102
4.2.2 QCM Principles . . . . . . . . . . . . . . . 103
4.2.3 QCM and Dissipation (D) . . . . . . . 105
4.2.4 Modern QCM-D Setup . . . . . . . . . 107
Contents ix
4.3 Ellipsometry . . . . . . . . . . . . . . . . . . . . . . . 108

4.3.1 Basic Principles of
Electromagnetic Theory and
Polarized Light . . . . . . . . . . . . . . . 109
4.3.2 Basic Principles of Ellipsometry . 113
4.3.3 Obtaining the Thickness of
Films: Optical Parameters Del
(Δ) and Psi (ψ) . . . . . . . . . . . . . . . . 115
4.3.4 The Ellipsometer . . . . . . . . . . . . . . 117
4.4 Surface Plasmon Resonance . . . . . . . . . . . 119
4.4.1 Principles of SPR . . . . . . . . . . . . . 120
4.4.2 SPR Instrument Setup . . . . . . . . . 122
4.5 Dual Polarization Interferometry . . . . . . 124
4.5.1 Waveguide Basics . . . . . . . . . . . . . 124
4.5.2 Waveguide Interferometry and
the Effective Refractive Index . . . 125
4.5.3 Principles of Dual Polarization
Interferometry . . . . . . . . . . . . . . . . 128
4.5.4 Parameters of a DPI Instrument
and Common Applications . . . . . . 130
4.6 Spectroscopic Methods . . . . . . . . . . . . . . 130
4.6.1 Interactions Between Light and
Matter . . . . . . . . . . . . . . . . . . . . . . 132
4.6.2 UV-Visible Spectroscopy . . . . . . . 135
4.6.2.1 Principles of UV-Visible
Spectroscopy . . . . . . . . . . 135
4.6.2.2 Setup of a UV-Visible
Spectrophotometer . . . . . 136
4.6.3 The Absorption of Visible Light
by a Nanofilm . . . . . . . . . . . . . . . . 138
4.6.4 Molecular Fluorescence
Spectroscopy . . . . . . . . . . . . . . . . . 141
4.6.4.1 Principles of
Fluorescence and
Fluorescence Quantum
Yield . . . . . . . . . . . . . . . . . 141
4.6.4.2 Setup of a Fluorometer
for Bulk Phase and
Thin Film Fluorescence
Measurements . . . . . . . . . 143
4.6.5 Vibrational Spectroscopy
Methods . . . . . . . . . . . . . . . . . . . . . 144
4.6.5.1 Introduction to
Vibrational Modes . . . . . . 144
x Contents
4.6.5.2 Attenuated Total

Reflection IR
Spectroscopy . . . . . . . . . . 148
4.6.5.3 Reflection Absorption
IR Spectroscopy . . . . . . . 150
4.6.6 Raman Spectroscopy . . . . . . . . . . 150
4.6.6.1 Rayleigh and Raman
Light Scattering . . . . . . . . 150
4.6.6.2 Surface Enhanced
Raman Spectroscopy . . . 154
4.7 Nonlinear Spectroscopic Methods . . . . . 155
4.7.1 An Introduction to Nonlinear
Optics . . . . . . . . . . . . . . . . . . . . . . 155
4.7.2 Second-Harmonic Generation . . . 159
4.7.3 Sum-Frequency Generation
Spectroscopy . . . . . . . . . . . . . . . . . 162
4.8 X-Ray Spectroscopy . . . . . . . . . . . . . . . . . 167
4.8.1 Absorption . . . . . . . . . . . . . . . . . . . 167
4.8.2 Fluorescence . . . . . . . . . . . . . . . . . 167
4.8.3 Diffraction . . . . . . . . . . . . . . . . . . . 168
4.9 Imaging Nanostructures . . . . . . . . . . . . . . 169
4.9.1 Imaging Ellipsometry . . . . . . . . . . 169
4.9.1.1 Imaging Using
Conventional
Ellipsometry . . . . . . . . . . 169
4.9.1.2 Principles of Modern
Imaging Ellipsometry . . . 169
4.9.1.3 Methods for Extracting
Ellipsometric Data in
Imaging Ellipsometry . . . 170
4.9.1.4 Image Focusing . . . . . . . . 171
4.9.1.5 Resolution of an
Imaging Ellipsometer . . . 171
4.9.2 Scanning Probe Methods . . . . . . . 172
4.9.2.1 Scanning Tunneling
Microscopy . . . . . . . . . . . 172
4.9.2.2 Atomic Force
Microscopy . . . . . . . . . . . 176
4.9.3 Transmission Electron
Microscopy . . . . . . . . . . . . . . . . . . 179
4.9.3.1 Principles of TEM . . . . . . 179
4.9.3.2 TEM Instrumentation . . . 183
4.9.4 Near-Field Scanning Optical
Microscopy . . . . . . . . . . . . . . . . . . 185
Contents xi
4.9.4.1 History and Principles

of NSOM . . . . . . . . . . . . . 186
4.9.4.2 Modern NSOM
Instrumentation
and Different NSOM
Operating Modes . . . . . . . 187
4.10 Light Scattering Methods . . . . . . . . . . . . . 191
4.10.1 The Measurement of
Scattered Light: Determining
the Aggregation Number
of Micelles . . . . . . . . . . . . . . . . . . 192
4.10.2 Dynamic Light Scattering . . . . . . 195
References and Recommended Reading . . . . . . 200
End of Chapter Questions . . . . . . . . . . . . . . . . . 200
Chapter 5 Types and Uses of Some Nanomaterials . . . . . . 207

Chapter Overview . . . . . . . . . . . . . . . . . . . . . . . . 207
5.1 Supramolecular Machines . . . . . . . . . . . . 207
5.1.1 Model Dye System . . . . . . . . . . . . 208
5.1.2 Photorelaxation . . . . . . . . . . . . . . . 211
5.1.3 Formation and Properties of the
Exciton . . . . . . . . . . . . . . . . . . . . . . 212
5.2 Nanowires . . . . . . . . . . . . . . . . . . . . . . . . . 215
5.2.1 Basic Quantum Mechanics of
Nanowires . . . . . . . . . . . . . . . . . . . 216
5.2.2 Conductivity . . . . . . . . . . . . . . . . . 217
5.2.3 Nanowire Synthesis . . . . . . . . . . . 218
5.2.4 Summary . . . . . . . . . . . . . . . . . . . . 219
5.3 Carbon Nanotubes . . . . . . . . . . . . . . . . . . 220
5.3.1 Carbon Nanotube Structure . . . . . 220
5.3.2 Some Properties of Nanotubes . . . 222
5.3.3 Methods for Growing
Nanotubes . . . . . . . . . . . . . . . . . . . 222
5.3.3.1 Arc Discharge . . . . . . . . . 223
5.3.3.2 Laser Ablation . . . . . . . . . 223
5.3.3.3 Chemical Vapor
Deposition . . . . . . . . . . . . 224
5.3.4 Catalyst-Induced Growth
Mechanism . . . . . . . . . . . . . . . . . . 224
5.4 Quantum Dots . . . . . . . . . . . . . . . . . . . . . . 225
5.4.1 Optical Properties . . . . . . . . . . . . . 226
5.4.2 Synthesis of Quantum Dots . . . . . 227
5.4.2.1 Precipitative Methods . . . . 227
xii Contents
5.4.2.2 Reactive Methods in

High-Boiling-Point
Solvents . . . . . . . . . . . . . . 228
5.4.2.3 Gas-Phase Synthesis
of Semiconductor
Nanoparticles . . . . . . . . . 229
5.4.2.4 Synthesis in a
Structured Medium . . . . 229
5.4.3 In Vivo Molecular and Cell
Imaging . . . . . . . . . . . . . . . . . . . . . 229
5.5 Langmuir-Blodgett Films . . . . . . . . . . . . . 230
5.5.1 Langmuir Films . . . . . . . . . . . . . . 232
5.5.2 Langmuir-Blodgett Films . . . . . . . 234
5.6 Polyelectrolytes . . . . . . . . . . . . . . . . . . . . . 237
5.6.1 Electrostatic Self-Assembly . . . . . 238
5.6.2 Charge Reversal and
Interpenetration . . . . . . . . . . . . . . 241
5.6.3 Multilayer Formation . . . . . . . . . . 243
5.7 Model Phospholipid Bilayer Formation
and Characterization . . . . . . . . . . . . . . . . 244
5.7.1 Black Lipid Membranes . . . . . . . . 245
5.7.2 Solid Supported Lipid Bilayers . . 246
5.7.3 Polymer Cushioned
Phospholipid Bilayers . . . . . . . . . . 250
5.7.4 Fluorescence Recovery after
Photobleaching . . . . . . . . . . . . . . . 250
5.7.5 Fluorescence Resonant Energy
Transfer . . . . . . . . . . . . . . . . . . . . . 252
5.7.6 Fluorescence Interference
Contrast Microscopy . . . . . . . . . . . 253
5.8 Self-Assembled Monolayers . . . . . . . . . . . 255
5.8.1 Thiols on Gold . . . . . . . . . . . . . . . . 256
5.8.2 Silanes on Glass . . . . . . . . . . . . . . 257
5.9 Patterning . . . . . . . . . . . . . . . . . . . . . . . . . 258
5.9.1 Optical Lithography . . . . . . . . . . . 259
5.9.2 Soft Lithography . . . . . . . . . . . . . . 262
5.9.3 Nanosphere Lithography . . . . . . . 262
5.9.4 Patterning Using AFM . . . . . . . . . 263
5.9.5 Summary . . . . . . . . . . . . . . . . . . . . 266
5.10 DNA and Lipid Microarrays . . . . . . . . . . . 266
5.10.1 Using a DNA Microarray . . . . . . . 266
5.10.2 Array Fabrication . . . . . . . . . . . . . 267
5.10.3 Optimization . . . . . . . . . . . . . . . . . 270
5.10.4 Applications . . . . . . . . . . . . . . . . . 271
Contents xiii
5.10.5 Arrays of Supported Bilayers

and Microfluidic Platforms . . . . . 271
5.10.6 Summary . . . . . . . . . . . . . . . . . . . . 274
Cited References . . . . . . . . . . . . . . . . . . . . . . . . . 274
References and Recommended Reading . . . . . . 276
End of Chapter Questions . . . . . . . . . . . . . . . . . . 278
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Preface
TO THE STUDENT
Nanoscience is a rapidly changing field where new innova-
tions and discoveries are being made every day. To write a
book that captures even a fraction of what the scientific liter-
ature has produced over the last few years would be a monu-
mental task. With this in mind, the topics presented in this
book are carefully selected to provide a basic understanding
of the field. Many important topics, such as computational
chemistry and solid state physics, have been given limited
coverage, largely because I want this book to be accessible
to any student who has taken introductory college-level sci-
ence courses.
This book is written for a full- or half-semester course in
nanoscience with an emphasis on understanding nanoma-
terials. The stress on “understanding” is the key behind the
objective of the text: to provide fundamental insight into the
molecular driving force underlying self-assembly processes,
as well as to explain how to characterize the resulting nano-
materials. Knowledge of self-assembly and characterization is
essential for an understanding of these interesting systems.
It should be noted that this book does not draw heavily from
scientific papers; rather, it should be used in conjunction with
the primary literature.
TO THE INSTRUCTOR
I have drawn relevant material from many scientific disci-
plines, assuming only a basic level of competency in physics,
chemistry, and biology. Mathematical rigor has been limited
to presenting key results and simple proofs. Instructors should
use their discretion in placing emphasis on or “filling holes”
in areas that may seem somewhat inadequate or limited in
scope. The half-course model is suggested for teaching mate-
rial directly from the book and solving the end-of-chapter
problems. For a full-semester course, the book should be used
in a course that requires students to refer to the primary lit-
erature. The latter may be more suitable for intermediate to
xv
xvi Preface
advanced-level classes, although I strongly suggest training

students to read papers early in their careers.
The approach taken in this book is to focus on preparing
the student to read papers in this field, and so I have limited
specific examples to landmark papers. This book should pro-
vide the necessary background to enable the student to com-
prehend articles from scientific journals. I teach the material
from this book in conjunction with a student seminar series in
which the students select interesting papers for presentation
and class discussion. This approach combines discussion-
based and problem-solving skills, and provides exposure to a
highly interdisciplinary field of study.
Acknowledgments
Writing this book would not have been possible without the
support of friends and family. My research students and mem-
bers of my Soft Nanomaterials class (2008 and 2009) have been
crucial in helping me to develop a textbook that meets the
needs of students interested in this field. I would like to express
sincere gratitude to Robert Rawle, Theodore Zwang, Michael
Haber, Michael Gormally, and Thomas Lane for making valu-
able contributions, including editing and extensive proofread-
ing of the manuscript. In addition, for their detailed review of
the manuscript and constructive comments, I’m very grateful
to Professors Lisa Klein at Rutgers University, Joseph Tracy at
North Carolina State University, and Marcus Lay at University
of Georgia. Last but not least I would like to thank my editor,
Luna Han, for her invaluable suggestions and her unwavering
commitment. This book would not have been written without
her dedication.
Malkiat S. Johal
Claremont, California
[email protected]
xvii
About the Author
Malkiat S. Johal is Asso

ciate Professor of Physical
Chemistry at Pomona College,
Claremont, California. He
obtained a first-class honors
degree in chemistry from
the University of Warwick,
UK. After earning his Ph.D.
in physical chemistry from
the University of Cambridge,
UK, Dr. Johal joined Los
Alamos National Laboratory,
New Mexico, as a post-doctoral
research associate, where he
worked on the nonlinear opti-
cal properties of nanoassem-
blies. His research laboratory
at Pomona College focuses
on using self-assembly and
ionic adsorption processes
to fabricate nanomaterials for optical and biochemical appli-
cations. Professor Johal’s laboratory also explores fundamental
phenomena such as ion-pair complexation, adsorption, sur-
face wettability, and intermolecular non-covalent interactions
that lead to highly ordered nanostructures. He teaches courses
in physical chemistry and soft nanomaterials. Understanding
Nanomaterials is his first book.
xix
One
A Brief Introduction
to Nanoscience
1.1 THE NEED FOR NANOSCIENCE EDUCATION

Significant growth in areas of scientific research inevita-
bly impacts education, first at the graduate level, and then
at the undergraduate level. In the last decade a healthy flow
of publications has addressed the needs of graduate students
and trained professionals in the field of nanoscience. Rapid
research advances in areas such as soft matter, supramolecu-
lar science, and biophysical chemistry have fueled the recent
surge in the number of professional journals in nanotechnol-
ogy. Over 50 journals are publishing research in nanotechnol-
ogy, and over 30 are devoted solely to nanomaterials.
Examples of new, devoted publications include the Royal
Society of Chemistry’s Soft Matter and Nanoscale, which cover
the interdisciplinary science underpinning the properties
and applications of soft matter at the nanoscale, as well as the
American Chemical Society’s ACS Nano and Nano Letters. In
addition, there has been steady growth in funding for nanosci-
ence from both private sources and government agencies such
as the National Science Foundation, Department of Energy, and
National Institute of Standards and Technology. Together with
the growth in research activity and sharp increase in profes-
sional publications, this rise in funding commitment provides
a compelling reason to begin serious training of our future
workforce in this area. Thus, the present textbook has been
developed with the needs of the undergraduate curriculum
in mind. This introductory textbook encompasses the funda-
mental principles governing the fabrication of nanomaterials,
processes such as self-assembly, patterning, and nanolithogra-
phy. Significant coverage is also given to the characterization
and applications of such materials. The book captures the
interdisciplinary nature of this field, and attempts to provide
a well-balanced approach to teaching nanoscience (although
1
2 Understanding Nanomaterials
much of the material emphasizes liquids and surface science).

The book also showcases nanomaterials research originating
from the chemistry, biology, physics, medicine, and engineer-
ing communities. In particular the book emphasizes, wherever
possible, topics of current global interest (energy, the environ-
ment, and medicine). The target audience is undergraduates
with at least one year of introductory chemistry, physics, and
(preferably) biology.
1.2 THE NANOSCALE DIMENSION AND

THE SCOPE OF NANOSCIENCE
The word “nano” is derived from the Latin word nanus, mean-
ing “dwarf,” and is often used in the context of miniaturiza-
tion. It is given the abbreviation n. In the international systems
of units, nano is the prefix used when multiplying a unit, such
as a given length, by 10 –9. Thus, one can speak of a nanometer
(1 nm = 1 × 10 –9 m), a nanosecond (1 ns = 1 × 10 –9 s), and even
a nanogram (1 ng = 1 × 10 –9 g). The nano term is typically used
to refer to objects with length scale approaching the order of
10 –9 m. Thus, one can speak of nanotubes, nanofossils, nano-
wires, and nanofilms as materials in which at least one dimen-
sion is on the order of 10 –9 m. The term nanofabrication refers
to the procedure used to construct materials with nanoscale
dimension.
To put the nanoscale in perspective, consider the size of a
hydrogen atom. You may have learned from introductory phys-
ics or chemistry that the Bohr radius (the distance from the 1s
electron to the central proton in hydrogen) is about 52.0 pm, or
roughly 0.05 nm. This distance arguably represents the lowest
limit with respect to atomic distances. In fact, atoms and their
ions vary in size between this number and about 0.3 nm. This
range represents the atomic scale. The hydrogen molecule, H2,
has a proton-proton distance (bond length) of about 0.07 nm.
The much larger I2 molecule has a bond length of about 0.3
nm, and the diameter of the benzene ring is about 0.5 nm. The
size of molecules increases rapidly with structural complex-
ity. The polyatomic molecule dodecanol, CH3(CH2)10CH2OH,
has a length approaching 2 nm. If 48 such molecules were
stacked together as shown in Figure 1.1, then it is conceivable
that an aggregate of length ~10 nm and height ~10 nm could
be formed. Nanostructures may be comprised of thousands of
such molecules resulting in aggregates on the scale of hun-
dreds of nanometers. Furthermore, macromolecular systems
A Brief Introduction to Nanoscience 3
CH3(CH2)11 group OH group
~10 nm
~~10 nm
10 nm
Figure 1.1 An assembly of 48 dodecanol molecules forming a hypothetical aggregate

of length and width 10 nm. The filled circles represent carbon atoms, and the open circle
represents the oxygen of the terminal hydroxyl group of the molecule. Hydrogen atoms are not
shown for convenience. This two-dimensional assembly is held together by a combination of
hydrogen bonds and hydrophobic interactions (discussed further in Chapters 2 and 3).
such as polymers and proteins have sizes approaching 10 nm;

the aggregation of such molecules may result in structures on
the scale of micrometers.
There is no clear boundary between what one considers
molecular scale and the nanoscale that characterizes aggre-
gated systems. It is simply a matter of the relative size of the
molecule making up the aggregate and the size of the aggre-
gate itself. It is sufficient to say that the field of nanotechnol-
ogy or nanoscience deals with the manipulation and control
of structures of a length scale 1000 nm or smaller. The sci-
ence is fascinating because physical and chemical phenomena
at these scales are markedly different from those observed in
bulk (macroscopic) matter. Sometimes the difference is just a
result of the much larger surface-area-to-volume ratio as par-
ticles shrink in size. Therefore, surface science plays a central
role in understanding nanomaterials.
Nanoscience is full of possibilities and presents us with
the potential for significant technological breakthroughs
in the near future. Nobel laureate Richard P. Feynman real-
ized the importance of this field almost six decades ago. In
his legendary speech, “There’s Plenty of Room at the Bottom,”

he stated: “This field is not quite the same as others in that
it will not tell us much of fundamental physics in the sense
of, ‘What are the strange particles?’ But it is more like solid
state physics in the sense that it might tell us much of great
interest about strange phenomena that occur in complex situ-
ations. Furthermore, a point that is most important is that it
would have an enormous number of technical applications”
(see http://www.zyvex.com/nanotech/feynman.html).
Nanotechnology is defined as the engineering or manipula-
tion of functional systems at the molecular scale. A functional
system is used to describe a material that has a well-defined
responsibility and performs that responsibility with little
or no “side-effects.” Although the term nanotechnology was
popularized in the 1980s, scientists have been studying nano-
structures for well over a century. As early as the mid-1800s,
Michael Faraday investigated the properties of colloidal gold
and attempted to explain the ruby color of such solutions.
These solutions were made by treating an aqueous solution
of NaAuCl4 with a reducing agent. Faraday concluded that
the resulting ruby-colored fluid was very finely divided gold
dispersed in the aqueous solution. A century later, electron
microscopy showed that these particles were indeed gold par-
ticles with average diameters of around 6 nm.
Over the last few decades, nanotechnology has focused
largely on the use of colloidal systems and simple nanopar-
ticles in coatings and materials. For example, silver nanopar-
ticles have found use in hundreds of products because of their
antimicrobial properties. More recently, nanotechnology has
been used to explore biologically active materials as novel
biosensors and targeted drug delivery vehicles for the treat-
ment of diseases. The field is also impacting microelectronics,
or more appropriately nanoelectronics, with the development
of new transistors, amplifiers, and adaptive structures. The
next few decades will inevitably move nanotechnology to the
point where we will be able to create complex nanosystems
and molecular devices by design. Richard Feynman envi-
sioned the capability of nanotechnology over 50 years ago
stating, “I want to build a billion tiny factories, models of each
other, which are manufacturing simultaneously. … The prin-
ciples of physics, as far as I can see, do not speak against the
possibility of maneuvering things atom by atom. It is not an
attempt to violate any laws; it is something, in principle, that
can be done; but in practice, it has not been done because we
are too big.”
1.3 SELF-ASSEMBLY
In general there are two strategies for designing molecular sys-
tems. A “top-down” approach (or step-wise design) describes
the breaking down, or decomposing, of a system to construct
the material of interest or to gain insight into its compositional
subunits. Nanoscale materials are carved into shape by physi-
cal methods such as lithography. In contrast, a “bottom-up”
approach describes the piecing together, or synthesis, of
f ragments to form the larger molecular system. In the most
fundamental sense, nanotechnology refers to the ability to con-
struct molecular assemblies from the bottom up, using current
methods and tools to make useful products. The bottom-up
approach may take advantage of specific chemical reactions or
may involve intermolecular interactions between molecular
fragments. Self-assembly, which is one of the important terms
used in nanoscience, describes a process in which a collec-
tion of disordered building blocks (molecules or nano-objects)
come together to form an organized structure. One common
example is the crystallization of small ions into a definite lat-
tice structure. Techniques such as x-ray diffraction show that
such highly organized structures are comprised of a repetition
of smaller “unit cells” of nanoscale dimension. Self-assembled
structures may also result from just a few molecules, but the
cooperative interplay between these molecules within the
structure may impart a specific function or property.
It is also worthwhile to note that self-assembly may lead to
discrete or extended entities. A discrete entity is well defined
in terms of the number of molecules it contains. A simple
example is the dimer that is formed when two acetic acid mole-
cules interact through hydrogen bonding (Figure 1.2). Another
good example, which is discussed in detail later, is a surfac-
tant “micelle”—a stable, often spherically shaped aggregate
comprised of a few hundred molecules. An extended entity
is undefined in at least one dimension. An example is a thin
film with a well-defined thickness, say, one molecule thick,
but with undefined length and width. A polymer molecule,
in which the exact number of monomeric building blocks is
O H O
C CH3
CH3 C
O H O
Figure 1.2 The acetic acid dimer. The dashed lines represent hydrogen bond interac-
tions (discussed in Chapter 2) between the carbonyl oxygen and the hydroxyl hydrogen. The
dimer is an example of a discrete entity.
usually unknown, is another example of an extended entity.

Nanotechnology exploits both discrete and extended entities
for specific functions.
Although scientists use artificial methods of self-assembly
to construct functional materials, there are many examples of
natural self-assembly processes. The folding of proteins and
other biomacromolecules and the formation of lipid bilayer
cell membranes are examples of natural self-assembly in biol-
ogy. Artificial methods may involve covalent building blocks
based on chemically robust starting materials or may exploit
the shapes, sizes, and intermolecular interactions of these
building blocks to direct the formation of nanostructures.
Self-assembly processes can be static or dynamic. Static
self-assembly describes the irreversible formation of a stable
structure. Examples of static self-assembly include the forma-
tion of the double helix of DNA and the folding of polypep-
tide chains into a protein molecule. Dynamic self-assembly
describes a reversible process such as adsorption of molecules
onto a surface and oscillating chemical reactions.
1.4 SUPRAMOLECULAR SCIENCE

Self-assembly in nanomaterials falls into the broader field
of supramolecular science. Supramolecular science refers to
the branch of science that focuses on systems composed of
a discrete number of molecular subunits (typically of nano-
scale dimension). These subunits are sometimes referred to
as molecular building blocks. Generally speaking, the spatial
organization of the building blocks is influenced by revers-
ible weak interactions such as hydrogen bonds, van der Waals
interactions, and electrostatic forces (Chapter 2). Although
irreversible interactions such as covalent bonds may also play
a vital role, supramolecular chemistry is concerned mainly
with non-covalent interactions.
Supramolecular science is important in a host of processes
such as protein folding, molecular recognition, self-assembly,
and host-guest chemistry. The acetic acid dimer mentioned
earlier is a discrete supramolecular entity. The hybridization of
single-stranded DNA in solution to the double-stranded form is
driven by hydrogen bonds being formed between base pairs—
this process results in an extended supramolecular entity.
Supramolecular science and the study of non-covalent inter-
actions touches every scientific discipline from biology (e.g.,
biological cell structure, protein-protein interactions, drug
delivery using nano-vehicles), chemistry (e.g., colloid stability,
micellar nano-reactor synthesis), and physics (e.g., organic

photovoltaic systems, holography, optical coatings, data trans-
mission and storage), to engineering (e.g., tertiary oil recovery,
large-scale synthesis of nano-wires) and environmental sci-
ence (e.g., remediation on nano-pores, detection of hazardous
materials on nano-films). For example, an extended entity such
as a thin film containing an antigen on the surface may be used
to detect the presence of a specific antibody. The complexation
of the antigen and the antibody, driven by non-covalent inter-
actions, may result in a supramolecular “bilayer.”
In this book, the terms supramolecular chemistry and self-
assembly are always used in the context of nanomaterials.
1.5 SOURCES OF INFORMATION ON NANOSCIENCE

In this book each chapter ends with a short list of relevant
further reading. These books, review articles, and papers have
been selected in part due to their clarity, depth, and manage-
able mathematical rigor. They should be accessible reading
material once the student has completed relevant chapters in
this text. However, some advanced books, which are typically
used by graduate students, technicians, or academic and indus-
trial professionals, are also listed for the ambitious student.
Readers are also encouraged to browse scientific journals for
up-to-date information on nanomaterials. The following jour-
nals are highly recommended: Nano Letters, Langmuir, Journal
of Physical Chemistry, Biomacromolecules, Advanced Materials,
ACS Nano, Applied Materials and Interfaces, and Chemistry of
Materials (all published by the American Chemical Society).
Other useful journals include Thin Solid Films, Nano Today,
Nanomedicine (Elsevier), Soft Matter, Nanoscale (Royal Society
of Chemistry), Nature Nanotechnology, and Nature Materials
(Nature Publishing Group).
The following books are additional recommended starting
points for students beginning their education in nanoscience.
• Deffeyes, K. S. and Deffeyes, S. E. Nanoscale: Visualizing

an Invisible World, 2009, Massachusetts Institute of
Technology, Cambridge, MA. This is a beautifully
illustrated book containing many examples of nano-
materials. The book describes the local structure of
materials at the nanoscale. It is an excellent read for
those beginning their studies in nanoscience.
• Ratner, M. A. Nanotechnology: A Gentle Introduction to
the Next Big Idea, 2002, Prentice Hall, New Jersey. This
book focuses on the technical and business aspects of

the field. It provides a wide perspective on the subject,
from science and economics to ethics.
• Jones, R. A. Soft Machines: Nanotechnology and Life,
2008, Oxford University Press, New York, NY. Although
not mathematically rigorous, this book does an excel-
lent job of presenting fundamental physical laws gov-
erning nanoscience.
• Understanding Nanotechnology from the editors of
Scientific American, 2002, Warner Books, New York,
NY. This is a popular science book that does a good job
of describing the technological implications of nano-
science. The mathematical and scientific background
is limited in this book, so it is an accessible overview
of the field and nanoscience terminology.
Two
Intermolecular Interactions
and Self-Assembly
CHAPTER OVERVIEW
Nanostructures assemble, often spontaneously, from simple
molecular building blocks. It is therefore important to begin
this chapter with a discussion of the forces between such mol-
ecules. The types of intermolecular interactions (for example
ion-ion, ion-dipole, dipole-dipole, dipole-induced dipole,
London forces, hydrogen bonds, and electrostatic forces) will
ultimately determine the degree and type of intermolecular
aggregation, as well as the structure of the resulting aggregate.
Such interactions are examined in both bulk media and on sur-
faces. This chapter concludes with some coverage on electronic
structure, and how simple quantum mechanical models can
be applied to predict some optical properties of nanomaterials.
In particular, conjugation in simple organic molecules is used
to make important connections between electronic structure,
intermolecular interactions, and molecular self-assembly.
2.1 INTERMOLECULAR FORCES AND
SELF-ASSEMBLY
This section introduces selected fundamental physical ideas
relating to the assembly and properties of nanomaterials in
order to provide a sufficient background for understanding
subsequent chapters. Intermolecular interactions play a central
role in surface chemistry and the process of self-assembly, both
of which affect the structure and properties of nanomaterials.
Such interactions also determine the properties of surfactants,
influence adsorption phenomena, and even affect the interaction
between molecules and electromagnetic radiation.
Self-assembly is the process during which molecular frag-
ments spontaneously and often reversibly organize themselves
into nanomaterials. The organization of these molecular
9
building blocks is driven by a combination of thermodynamic

factors, kinetic factors, and intermolecular interactions. The
interactions may be covalent in nature, leading to strong
bonds between the molecules and resulting in an irreversibly
self-assembled nanostructure. Covalent interactions are men-
tioned throughout this chapter with specific examples given in
Chapter 5. It is important to appreciate that self-assembly can
be spontaneous and directed. We begin this chapter by review-
ing some important non-covalent intermolecular interactions
that govern the formation of self-assembled nanomaterials.
Various forces are responsible for intermolecular interac-
tions. Most of the forces are electrostatic in origin, and we dis-
cuss them from a classical perspective, although it should be
noted that a quantum mechanical approach to understanding
intermolecular forces is perhaps more correct.
Any interaction between two molecules can be thought of
as a sum of a variety of different forces. We will discuss many
of these forces, including ion-ion forces, ion-dipole forces,
dipole-dipole forces, induced dipole forces, dispersion forces,
and hydrogen bonds. Depending on the types of molecules
interacting, one force or another may predominate.
Scientists often express intermolecular interactions not as
forces, but as intermolecular potentials (or the potential energy
of interaction). The potential energy (V) and force (F) between
two interacting molecules are related by
dV ( r )
F (r ) = − (2.1)
dr
where r is the distance between the two molecules and may

be defined differently for different types of interacting mol-
ecules. The negative sign on the derivative means that as the
potential energy of the interacting molecules increases with
increasing r, the force will tend to move it toward smaller r to
decrease the potential energy.
The van der Waals interaction is the collective term used
to describe attractive or repulsive forces, or non-covalent
interactions, between molecules. Named after Dutch scientist
Johannes Diderick van der Waals, this type of intermolecu-
lar interaction generally refers to molecules involving ion-ion,
ion-dipole, dipole-dipole forces, and interactions involving
induced dipoles (including London dispersion forces). Van
der Waals forces play a key role in biology, polymer science,
surface science, nanotechnology, and material science. They
Intermolecular Interactions and Self-Assembly 11
govern self-assembly processes, protein-protein interactions,

and crystallization processes. These interactions are also
found in nature. For example, the ability of geckos to climb
smooth surfaces (such as glass) is attributed to van der Waals
interactions and likely involves a nanofilm of water trapped
between the glass surface and the foot. In fact, research is cur-
rently being done in many nanoscience laboratories to mimic
this behavior and allow people to scale walls, or to create
“gecko tape” that exploits this ability.
Geckos possess the ability to cling to nearly any surface,
sometimes even by a single toe, because they have millions
of branching hairs called setae on their toe-pads that present
enough surface area for van der Waals interactions to have an
influence at the macroscopic scale. Recent advances in nan-
otechnology have yielded reusable “adhesives” that are four
times more “adhesive” than a gecko’s foot. These adhesives
are comprised of flexible polymers connected by silicon bases
to carbon nanotubes, which are cylindrical graphene columns
held together by van der Waals interactions. Structure, proper-
ties, and uses of carbon nanotubes will be further discussed
in Chapter 5.
2.1.1 Ion-Ion Interactions

Ion-ion forces are perhaps the most well-known intermolecu-
lar forces and are among the strongest intermolecular forces
of those we’ll be discussing. Ion-ion forces arise between two
ionic (charged) species, such as the force between Na+ and Cl–
that holds together crystals of common table salt. The poten-
tial energy of interaction V(r) between two charges q1 and q2 is
often called the Coulombic energy and is given as
q1q2
V (r ) = (2.2)
4π ε 0 r12
where ε0 is the permittivity of free space (8.854 × 10 –12 m–3 kg–1

s4 A2) and r 12 is the distance between the two ionic species.
For atomic or molecular ions, q is often calculated as q = ze
where z is the formal charge on the ion and e is the charge on
an electron, 1.60217 × 10 –19 C.
We can express the Coulombic force using the relationship
between potential energy and force (Equation 2.1) as
dV ( r ) q1q2
F (r ) = − = (2.3)
dr 4π ε 0 r122
Equation 2.3 is obtained by differentiating Equation 2.2, and

realizing that d/dr(1/r) is –1/r2. From Equation 2.3, we see that
the Coulombic force between two ions changes as 1/r2. We also
see that the force is negative when the two ions are attracted
to each other (when q1 and q2 have opposite signs) and positive
when they repel each other.
Example 2.1 The Coulombic Energy Between Ions
The ionic radius of Na+ is determined to be 95 pm and the ionic

radius of Cl– is 181 pm. Calculate the Coulombic energy between
two isolated ions of Na+ and Cl– if they are in contact as shown in
Figure 2.1.
Solution If the two ions are “in contact,” then the distance between
their centers is the sum of their two ionic radii. So,
rNa–Cl = 95 pm + 181 pm = 2.76 × 10 –10 m
Each ion has a formal charge of +1 or –1, so
qNa+ = ze = (+1)(1.60217 × 10 –19 C) = 1.60217 × 10 –19 C
qCl– = ze = (–1)(1.60217 × 10 –19 C) = –1.60217 × 10 –19 C
Then, using Equation 1.2,

qNa+ qCl− (1.602 × 10−19 C)(−1.602 × 10−19 C)
V (r )Na−Cl = =
4π ε 0rNa−Cl 4π(8.854 × 10−12 m−3 kg −1 s4 A2 )(2.76 × 10−10 m)
= − 8.36 × 10−19 J
where 1 J = 1 kg m2 s –2 and 1 C = 1 A s.
One example of ion-ion interactions can be found in the for-

mation of a charged polymer layer, or nanofilm, on a silicon
surface. Silicon usually has an oxide layer (SiO2) of about 1 nm
Na+ Cl–
Figure 2.1 A schematic depiction of Na+ and Cl– “in contact” with each other.
in thickness on its surface. In a network of surface SiO2 groups,

each silicon atom has a tetrahedral molecular geometry, result-
ing in a layer of oxygen atoms each covalently bonded to a single
silicon atom at the surface. At very low pH, these oxygen atoms
are protonated, but at neutral to high pH, the oxygen atoms are
deprotonated and thus create a layer of negative ionic charge
(Si-O –) along the surface. Polyethylenimine (PEI), a polycation
with several amine functional groups, can then be exposed to
this negatively charged surface to create a positively charged
PEI layer of relatively uniform thickness. A negatively charged
ion or polymer can then be exposed to this surface-bound PEI
layer to create a secondary layer, resulting in a new negatively
charged surface that allows the process to be repeated. This
process is called electrostatic self-assembly and will be used
to develop several techniques in future chapters.
2.1.2 Ion-Dipole Interactions

Many molecules possess permanent dipoles and are classi-
fied as polar molecules. Polar molecules do not have a perma-
nent charge, but because of the differing electronegativities of
the atoms bound in the molecule, certain regions of the mol-
ecule may have a partial positive or a partial negative charge.
In certain cases this partial charge can lead to a permanent
dipole. For example, a water molecule has a permanent dipole
due to its bent geometry. The oxygen atom has a much higher
electronegativity than the hydrogen atoms, and so it tends to
draw more electrical charge to itself. As a result, the hydro-
gens have a partial positive charge and the oxygen has a par-
tial negative charge. The net dipole moment passes through
the oxygen atom and bisects the hydrogen atoms as shown in
Figure 2.2.
δ−
δ+ δ+
H H
Figure 2.2 The electronegative oxygen pulls electron density away from the hydrogen
atoms, leaving them with a partial positive charge δ+. Each of the hydrogens has a dipole
moment that points toward the partial negative charge of the oxygen, resulting in a net
dipole moment that passes through the O and bisects the Hs.
Polar molecules such as H2O are often characterized by

their dipole moment µ, which is defined as
μ = qL (2.4)
where L is the distance separating the partial positive and par-

tial negative charges of magnitude q as shown in Figure 2.3.
Dipole moments of some common molecules are shown in
Table 2.1. As shown in the table, dipole moments are often
given in units of Debye (D), where 1D = 3.336 × 10 –30 C m.
When an ion draws near to a molecule with a dipole, there
is an electrostatic interaction between the dipole and the ionic
species. The potential energy of this interaction is given as
qion μ cos θ
V ( r , θ) = (2.5)
4π ε 0 r122
+= qL
q+ q–
Figure 2.3 The dipole moment µ between two charges q+ and q– separated by a dis-
tance L is calculated as µ = qL.
Table 2.1 Dipole Moments of Some Common Molecules

Molecule Dipole Moment (D)
H2O 1.85
CH3OH 1.7
NH3 1.47
CH3Cl 1.9
CHCl3 1.04
CH3COOH 1.7
NaCl 9.00
HCl 1.11
C6H5CN (benzonitrile) 4.18
Source: CRC Handbook of Chemistry and Physics, 88th ed., Web Version.
(2008) pp. 9-47–9-55. With permission.
q+
Ion θ µ
L
r12
q1 = ze
q–
Figure 2.4 A schematic depiction of the variables involved in an ion-dipole interaction.

L is the distance between the centers of the two partial charges of the dipole. r12 is the
distance between the center of the ion and the midpoint of L. θ is the angle between L and
r12. q1 is the charge on the ion, calculated as ze. µ is the dipole moment.
where µ is the dipole moment, qion is the charge on the ion, r 12

is the distance between the ion and the center of the dipole
moment, and θ is the angle between L and r 12. A schematic of
the geometries involved in a typical ion-dipole interaction is
shown in Figure 2.4.
From Equation 2.5 we see that the potential energy of an
ion-dipole interaction is angle dependent, which makes sense
intuitively. For example, consider the interaction between
a cation (a positively charged ion) and a dipole. The cation
attracts the negative region of the dipole, but repels the posi-
tive region. If the negative region is oriented toward the cat-
ion and the positive region is pointing away from it (θ = π),
we expect the magnitude of the potential energy to be max-
imized. Likewise, if the positive region is pointing toward
the cation and the negative region is pointing away (θ = 0),
the magnitude of the potential energy is maximized (but it
is repulsive rather than attractive). If, however, the dipole is
oriented perpendicularly with respect to the cation (θ = π/2),
the potential energy is zero because the attractive interac-
tion between the cation and the negative region is balanced
out by the repulsive interaction with the positive region.
Therefore, the angle dependence of the ion-dipole interaction
is as expected.
It should be noted that some molecules (such as benzene or
CO2) may have partial charge separations, but do not have a
net dipole moment. For example, in the case of CO2, the dipole
moment from the carbon toward one oxygen cancels out the
dipole moment from the other. However, in these types of
molecules, a multipole may exist and also results in an inter-
molecular force between the molecule and an ionic species.
Different equations than those above must be used to calculate
the interaction between molecules with electrical multipoles

and ionic species.
2.1.3 Dipole-Dipole Interactions

Molecules with permanent dipoles may also interact with each
other through electrostatic means. As might be expected, the
strength of the interaction is also angle dependent. This type
of interaction is analogous to the magnetic attraction between
two bar magnets—the attraction between the two magnets
depends on the angle of rotation of each magnet relative to
the other. A schematic representing the interaction between
two molecular dipoles is depicted in Figure 2.5. The potential
energy for such an interaction between two dipole moments µ1
and µ2 can be calculated as
μ 1μ 2
V ( r , θ1, θ2, φ) = −
4π ε 0 r123
(2 cos θ1 cos θ2 − sin θ1 sin θ2 cos φ (2.6))
where θ1, θ2, and ϕ are defined in Figure 2.5.

It might be tempting to think that the attraction between
two dipoles is always maximized when the two dipoles are
in line “head to tail” with the partial positive charge of one
dipole pointing directly toward the partial negative charge of
the other. However, depending on the lengths L of the inter-
acting dipoles, the most attractive interaction may often be
when the two dipoles are anti-parallel to each other, with the
positive region of one dipole directly adjacent to the negative
q+1
q+2
θ1 φ θ2
µ1 L1 L2 µ2
r12
q–2
q–1
Figure 2.5 A schematic depiction of the variables involved in a dipole-dipole interac-

tion. L is the distance between the centers of the two partial charges of either dipole 1 or 2.
r12 is the distance between the midpoints of L1 and L2. θ1 and θ2 are the angles between L1
and r12 or L2 and r12, respectively. ϕ is the angle of rotation between dipoles 1 and 2. q+ and
q– are the partial positive or negative charges in each dipole. µ is the dipole moment.
region of the other. This anti-parallel orientation allows the

molecules to draw closer together, reducing the value of r 12
and thereby maximizing the attractive interaction energy.
In the purification technique of column chromatography,
solid polar molecules such as silica or alumina are placed in
a vertical glass column and are referred to as the stationary
phase. The liquid mobile phase of solution to be purified is then
flowed through the column. Because dipole-dipole interactions
between polar molecules in the mobile phase and the molecules
of the stationary phase slow the polar molecules’ descent, com-
pounds in the solution flow out the bottom of the column, or
elute, in order of increasing polarity and can thus be separated.
This same technique works for ions in solution as well because
they are slowed by ion-dipole interactions. Polarity of the sol-
vent dictates the rate of movement of compounds through the
column. If a solvent is too polar, the stationary phase attracts
the solvent rather than the solutes and no separation occurs.
However, if a solvent is not polar enough, some more polar sol-
utes may not make it all the way through the column.
2.1.4 Interactions Involving Induced Dipoles

When an ion approaches a non-polar molecule, the electrons
of the non-polar molecule may feel the effect of the electric
field produced by the ion. As a result, the electron cloud sur-
rounding the non-polar molecule may become distorted. For
example, when a cation approaches a non-polar molecule, the
electron cloud of the non-polar molecule is pulled slightly
toward the cation. The result of this distortion of the elec-
tron cloud is an effective charge separation in the non-polar
molecule and is called an induced dipole. A schematic of an
ion-induced dipole interaction is shown in Figure 2.6. Polar
molecules, as well as ions, are also capable of inducing dipoles
in non-polar molecules.
– +
δ– – + δ+
+ – +
– +
–
Ion Induced Dipole
Figure 2.6 As a cation approaches a polarizable atom or molecule, its electric field
produces a distortion of the electron cloud surrounding the polarizable atom or molecule.
The proximity of the cation to the molecule results in an effective charge separation and
produces an induced dipole in the molecule.
The extent to which the electron cloud of a molecule

becomes distorted in the presence of an ion or polar molecule
is termed its polarizability. Polarizability α is mathematically
defined in terms of the strength of the dipole induced in a mol-
ecule due to an electric field of strength E,
μ induced = α E (2.7)
Molecules with high polarizabilities have a larger induced

dipole moment in the presence of an electric field than those
with low polarizabilities. Typical polarizabilities of various
atoms and molecules are shown in Table 2.2. Note that the units
of polarizability given in the table are 10–24 cm3 divided by 4πε0.
The presence of an induced dipole moment in a non-polar
molecule means that a potential energy of interaction exists
between the non-polar molecule and the ion or polar molecule
that is inducing the dipole. Using Coulomb’s law and Equation
2.3, we can calculate the electric field produced by an ion as a
function of the distance r from the center of the ion
F q
E (r ) = = (2.8)
q 4π ε 0 r 2
The interaction energy between an ion and an induced

dipole can then be given as
−q1α
V (r ) = (2.9)
2(4π ε 0 )2 r124
Table 2.2 Polarizabilities of Several Atoms and Molecules

Molecule Polarizability (10–24 cm3/4πε0)
H2O 1.45
D2O 1.26
NH3 2.2
He 0.20
H2 0.8
CH4 2.59
Au 5.8
Si 5.38
C6H6 (benzene) 10.2
Source: CRC Handbook of Chemistry and Physics, 88th ed., Web Version. (2008)
pp. 10-193–10-202. With permission.
Example 2.2 Perturbation of Electron Clouds Due to Ions
The center of a sodium ion (Na+) is located 0.35 nm from the cen-
ter of a gold atom. If the atomic radius of gold is 144 pm, by what
percentage of its atomic radius would the electron cloud of the gold
atom be shifted due to the presence of the sodium ion?
Solution The electric field induced by Na+ at a distance 0.35 nm
from its center is calculated using Equation 2.8.
qNa+ (1.602 × 10−19 C)

E (r ) = =
4π ε 0r 2 4π(8.854 × 10−12 m−3 kg −1 s4 A2 )(0.35 × 10−9 m)2
J
= 1.18 × 1010
Cm
Then, using Table 1.1, the dipole moment induced in the gold
atom is
⎛ J ⎞
μ = α E = 4π ε 0 (5.8 × 10−30 m3 ) ⎜1.18 × 1010 = 7.61 × 10−30 C m
⎝ C m ⎟⎠
Now, from Equation 2.4, we know that μ = qL , and so per unit

charge (e) we can calculate L
μ 7.61´10−30Cm
L= = = 4.75 × 10−11m = 47.5 pm
q 1.602 × 10−19 C
Then the electron cloud of Au is shifted by 47.5 pm/144 pm =

33% of its atomic radius.
In an analogous manner, the electric field strength at a

given point in space produced by a polar molecule with dipole
moment µ is a function of the orientation of the dipole moment
with respect to that point in space and is calculated as
μ (3 cos2 θ + 1)1/2
E ( r , θ) = (2.10)
4π ε 0 r 3
The potential energy of interaction between a polar molecule

with permanent dipole µ1 and an induced dipole is therefore
− μ 12 α (3 cos2 θ + 1)
V ( r , θ) = (2.11)
2(4π ε 0 )2 r126
where θ is the angle between the dipole moment of the polar

molecule and the line connecting the midpoint of the polar
molecule with the center of the induced dipole.
Finally, we note that the interaction between an ion or a
polar molecule and an induced dipole is always attractive. It
is inherently so because the electric field produced by the ion
or polar molecule always induces a dipole in the polarizable
molecule that is oriented such that it is attracted toward the
species inducing the dipole.
2.1.5 Dispersion Forces

Aside from the forces that are essentially electrostatic in
nature described in the previous sections, there also exists a
force between all molecules, regardless of charge or polarity,
that results from the quantum mechanical correlation between
electrons in neighboring molecules. This force is called the
dispersion or London force. Although dispersion forces are
inherently quantum mechanical in nature and a rigorous
description of their origin is beyond the scope of this book, we
can gain an intuitive grasp of dispersion forces by considering
in a somewhat classical manner their contribution to the inter-
action between two neutral, non-polar molecules.
Even though a neutral, non-polar molecule has no perma-
nent dipole moment, at any given moment the distribution
of its electrons may be asymmetrical, resulting in an instan-
taneous or momentary dipole moment. This instantaneous
dipole moment creates an electric field that perturbs the elec-
trons of a neighboring molecule, producing in it an induced
dipole moment and resulting in an attractive force between
the two molecules.
In order to calculate the dispersion force between two mol-
ecules, a quantum mechanical calculation must be performed,
the accuracy of which generally depends on the level of the-
ory used. One of the earliest calculations was performed by
London in the 1930s using quantum mechanical perturbation
theory. His calculation produces reasonably accurate results,
and although more precise calculations have been performed
in more recent years, London’s equation is less complex and
therefore more suitable for our purposes.
According to London’s results, the approximate potential
energy of interaction due to dispersion between two molecules
can be calculated in terms of their electronic polarizabilities α
and their ionization potentials I. For two identical molecules
(or atoms), the result is
−3 α 2I −Cdispersion
V (r ) = = (2.12)
4(4π ε 0 ) r126
2
r126
and for two different molecules
−3 α1α 2 I 1I 2
V (r ) = (2.13)
2(4π ε 0 )2 r126 (I 1 + I 2 )
As with the interaction energy for dipole-induced dipole

interactions in Section 2.1.4 (Equation 2.11), we see that the
potential energy of interaction for a dispersion interaction
according to London’s equations goes as 1/r 6 and is always
attractive between any two molecules.
The dispersion interaction plays an important role in the
liquid and solid phases of many materials and is the main con-
tributor to cohesion. However, it turns out that the strength of
the dispersion interaction does not vary much between differ-
ent types of molecules (i.e., the interaction between any two
given molecules is of similar strength). Therefore, the electro-
static interactions described in the earlier sections, and not
dispersion interactions, are generally responsible for such
behaviors as phase separation and self-assembly in condensed
phases, behaviors that are of utmost importance in the devel-
opment and study of nanomaterials.
2.1.6 Overlap Repulsion

In our discussion of the different types of intermolecular
potentials in the previous sections, we ignored the fact that
atoms and molecules occupy some finite space. For example, if
we examine the equation for Coulombic force (Equation 2.3) by
itself, we would be led to conclude that two oppositely charged
ions are drawn toward each other with increasing force until
they occupy the same point in space. Obviously, this does not
occur with atoms and molecules in nature. To account for the
finite size of atoms and molecules, we then include another
contributor to the interaction potential energy between two
atoms or molecules called overlap repulsion. Overlap repul-
sion is the interaction that accounts for the fact that two atoms
or molecules cannot occupy the same point in space.
What, then, is the size of an atom or a molecule? This is
not a trivial question. From the results of quantum mechanics,
we realize that the electron “clouds” of atoms and molecules
do not have definite boundaries. Determining where an atom
“ends” is therefore somewhat tricky. As a result, radii of atoms

are often experimentally defined and, depending on the type of
measurement made (and consequently the property measured),
a different result might be obtained. For example, one way to
measure the radii of atoms would be to assume that in solids
they act as tiny, hard spheres that have packed closely together
(see Figure 2.7). Using x-ray or neutron diffraction methods,
one could then observe how closely the atoms pack together
in a crystal and thereby deduce the atomic radius. The results
of this method yield a type of atomic radius called the hard
sphere radius or van der Waals packing radius. Other methods
also exist, such as measuring the distance between two atoms
in a covalent bond (rather than the distance in a crystal). This
method yields the covalent bond radius. The atomic radius
calculated depends on the method used. In certain cases, the
results obtained from these different methods might vary by
as much as 30%. The type of measurement one chooses to use
generally depends on the type of system being studied.
After using the most suitable method to determine an atomic
radius, one can then calculate the overlap repulsion between
two atoms. A variety of models with increasing sophistication
and complexity are used to calculate the repulsive potential
energy between two atoms due to overlap. Perhaps the sim-
plest model is to characterize the atom as a “hard sphere”
with a definite boundary (i.e., the repulsive force between
two atoms would be infinite at any distance smaller than the
atomic radius). This hard sphere model between two atoms at
Figure 2.7 Atoms in crystal lattices can often be modeled as tiny, hard spheres in order
to calculate their atomic radius. X-ray or neutron diffraction methods can then be used to
experimentally determine the atomic radius.
a distance r from each other can be represented mathemati-

cally as
∞
⎛ σ⎞
V (r ) = ⎜ ⎟ (2.14)
⎝r⎠
where σ is the atomic or molecular diameter (i.e., two times the

atomic radius). As expected, when r > σ, then V(r) is essentially
zero and when r < σ, V(r) is infinitely large. A graph of V(r) ver-
sus r for the hard sphere model is shown in Figure 2.8(a).
A more realistic model is the soft sphere model, which
assumes that atoms are “compressible” to some degree and do
not have completely rigid boundaries. One mathematical rep-
resentation for the soft sphere model can be given as a power
law as
n
⎛ σ⎞
V (r ) = ⎜ ⎟ (2.15)
⎝r⎠
where n is usually an integer between 9 and 16 and σ is defined

as before. In this model, V(r) quickly becomes quite small
when r is much bigger than σ. Conversely, the overlap repul-
sion becomes large rather quickly when the distance between
the two atomic centers is less than σ. Figure 2.8(b) depicts a
graph of V(r) versus r for the soft sphere power law model.
Thus far in our discussion of overlap repulsion, we have
operated under the assumption that the atomic or molecular
Hard sphere model Soft sphere model
V(r) 0 r V(r) 0 r
σ σ
(a) (b)
Figure 2.8 (a) The hard sphere model of overlap repulsion. r is the intermolecular dis-
tance and σ is the molecular diameter. (b) The soft sphere model of overlap repulsion. Note
that r can assume some values slightly smaller than σ without V(r) becoming infinitely large,
as is the case with the hard sphere model.
geometry is essentially spherical. While this assumption is

relatively sound for atoms and some small molecules (CH4, for
instance, can be modeled as nearly spherical), most molecules
possess other geometries. The concepts of overlap repulsion
we have developed still apply to these species, but different
calculations of their interaction energy are required to account
for their differing geometries. However, such methods of cal-
culation are beyond the scope of this text.
2.1.7 Total Intermolecular Potentials

The previous sections provided a fundamental description
of the main forces involved in van der Waals intermolecu-
lar interactions. Ultimately, the total interaction potential
energy between two molecules is the sum of all the different
interactions that we have discussed (as well as a few more
complex interactions).
In a very basic treatment of the interactions between two
atoms or molecules, the total intermolecular potential is often
modeled by the Lennard-Jones potential, which is the sum of
a soft sphere repulsion term and an attractive term that goes
as 1/r 6 (analogous to the London dispersion attractive interac-
tion). The Lennard-Jones potential is given as
⎡⎛ σ ⎞ 12 ⎛ σ ⎞ 6 ⎤
V ( r ) = 4ε ⎢ ⎜ ⎟ − ⎜ ⎟ ⎥ (2.16)
⎢⎣⎝ r ⎠ ⎝r⎠ ⎥
⎦
where –ε is the minimum energy and σ is a constant parame-

ter (not the molecular diameter). A graph of the Lennard-Jones
potential is shown in Figure 2.9.
σ
V(r) 0 r
–ε
Figure 2.9 The Lennard-Jones total intermolecular potential curve. –ε is the minimum
energy. r is the intermolecular distance.
Although the Lennard-Jones potential is a relatively primi-

tive model of the total intermolecular potential between two
molecules, it provides us with a qualitatively useful picture of
a common interaction between two molecules. Starting from
the far right-hand side of Figure 2.9, we see that the poten-
tial energy decreases as the distance between the molecules
grows smaller, until the energy reaches a minimum value of
–ε. If r decreases beyond this minimum energy value, then the
potential energy quickly increases (i.e., the force between the
two molecules becomes strongly repulsive) due to the overlap
repulsion term.
A more complete total intermolecular potential would be
the sum of all interaction potential energies. Using only the
interactions we have discussed so far, the complete total inter-
molecular potential would look like that shown in Equation
2.17. However, it must be realized that simple systems will not
exhibit all of these kinds of interactions, so that some terms
in Equation 2.17 will be zero.
V ( r )total = V ( r )ion-ion + V ( r )dipole-diople + V ( r )ion-dipole
+V ( r )ion-induced dipole + V ( r )dipole-induced dipole

(2.17)
+V ( r )dispersion + V ( r )overlap
If the sum of attractive interaction terms is greater than the

repulsive interaction terms, then the two molecules are drawn
together until the repulsive interactions eventually overwhelm
the attractive interactions (remember that the overlap repul-
sion quickly becomes prohibitively large at distances smaller
than the atomic or molecular radii).
In conclusion and as a cautionary addendum, note that
many of the models of intermolecular forces discussed in this
and previous sections are for the most part rudimentary and
hence not fully useful if one wishes to make exhaustive calcu-
lations of the potential energies between atoms or molecules.
However, the interactions we have discussed are qualitatively
very useful for the purposes of this text and provide the con-
ceptual tools to understand the intermolecular forces at play
in the realm of nanomaterials.
2.1.8 Hydrogen Bonds

The hydrogen bond is a special type of intermolecular
interaction of great importance in colloidal systems and
n anomaterials in general. The hydrogen bond is essentially
electrostatic in origin and so is a subset of the dipole inter-

actions already discussed. However, it is of particular
importance and strength, and so has acquired a special
classification.
Hydrogen bonds occur between molecules that have hydro-
gen covalently bonded to a strongly electronegative atom such
as N, O, or F. In such cases, the electron density surrounding
the hydrogen atom is mostly drawn toward the more electro-
negative atom, leaving the hydrogen atom “exposed” with a
strong partial positive charge through which it may form a
strong dipole-dipole interaction with the electronegative ele-
ment on an adjacent molecule. Perhaps more importantly, the
reduced electron density around the hydrogen atom means
that the neighboring molecule with which it is hydrogen bond-
ing can draw much closer than it could otherwise (within ~1.5
to 2.0 Å). In other words, the overlap repulsion is minimized
because the electron “cloud” surrounding the hydrogen atom
has been reduced in size. Because the neighboring molecule
can draw much closer, the magnitude of the attractive interac-
tion energy is much greater than it would normally be (remem-
ber that in our equations for electrostatic interaction energies,
r 12 was in the denominator and usually raised to some power;
so if r 12 is smaller, then the magnitude of the interaction poten-
tial energy is much greater). Figure 2.10 depicts a hydrogen
bond between two water molecules.
rhydrogen bond
H H
Figure 2.10 The hydrogen bond between two water molecules. The electronegative oxy-
gen atom pulls much of the electron density surrounding the hydrogen atom to itself, giving
the oxygen a large partial-negative charge and leaving the hydrogen atom with a partial-
positive charge and very little electron density. The oxygen atom of a neighboring water
molecule can therefore approach much closer to the hydrogen atom than would normally be
possible for a dipole-dipole interaction.
2.1.9 The Hydrophobic Effect

The van der Waals interactions discussed so far are respon-
sible for many of the physical properties (e.g., solubility) of
organic molecules. For example, in methanol (CH3OH), the
hydrocarbon portion of the molecule is relatively small and
the polar hydroxyl group is largely responsible for the weak
intermolecular van der Waals interactions. However, as the
length of the hydrocarbon moiety increases [e.g., as in decanol,
CH3(CH2)9OH] the non-polar hydrocarbon portion of the mol-
ecule dominates the interactions and defines the solubility.
Hydrocarbon chains are essentially oil and have little or no
tendency to interact with water. If the hydrocarbon chain is
long enough the molecules may drop out of solution (precipi-
tate) and interact with themselves instead of the water mole-
cules. This brings us to a discussion of the hydrophobic effect,
which describes the tendency of non-polar molecules to form
aggregates of like molecules in water.
Experience shows us that at the macroscopic level oil and
water do not mix, but rather form separate layers. We also
know that water forms beads, or distinct droplets, on oil-like
surfaces such as the surface of a leaf. Often when organic mol-
ecules containing large non-polar hydrocarbon moieties are
placed in water they spontaneously self-organize themselves
into larger aggregates because this arrangement minimizes
overall contact between the hydrocarbon part of the molecule
and the water molecules of the solvent. The origin of this
phase separation is primarily to maximize the water-water
intermolecular interactions, because the dipole-dipole inter-
actions are stronger than the interactions between non-polar
molecules. In other words, the enthalpic force is in the water,
not the non-polar phase, meaning that this type of phase sepa-
ration is enthalpically driven.
In order to complete our understanding of the hydrophobic
effect, we need to say something about entropy. In addition to
the enthalpic contribution discussed above, entropy may play
an important role, especially in more complex systems. From
general chemistry, entropy is described as a thermodynamic
property related to the number of energy levels among which
the energy of a system is spread. Basically this is saying that a
collection of water molecules confined and “ordered” within
a small region of space will have smaller entropy compared
to the same molecules “disordered” and occupying a larger
volume. Entropy changes are positive for a spontaneous pro-
cess, which means that it is favorable for a water molecule to
be “released” from a small space and enter a larger volume,

especially if such a process leads to a loss of order.
Consider an organic molecule with a large hydrocarbon
chain [e.g., CH3(CH2)9OH] placed in water. It turns out that water
molecules, in avoiding contact with the hydrocarbon chain,
form a cage around the chain. This cage contains ordered
water molecules essentially immobilized around the organic
molecule. When two such “hydrated” molecules approach each
other and make contact, this cage is disrupted and the con-
fined water molecules are liberated into the bulk solution.
This process is accompanied by an increase in entropy of
the water molecules. In a sense the van der Waals attractions
between the two hydrocarbon chains is driven in part by the
large increase in entropy due to the disruption of the cage.
If this process occurs among many organic molecules it will
lead to aggregation, typically resulting in nanoscale entities
dispersed within the aqueous solution.
The hydrophobic effect is paramount in many self-assembly
processes, including some biological processes such as the for-
mation of the cell membrane. Reference to the effect is made
throughout the text and, in particular, Section 3.3 contains a
discussion of the hydrophobic effect in the context of surfac-
tant chemistry.
2.2 ELECTROSTATIC FORCES BETWEEN SURFACES:

THE ELECTRICAL DOUBLE LAYER
Surface chemistry plays a vital role in the self-assembly of
nanomaterials. The forces discussed so far (van der Waals inter-
actions, hydrogen bonds, hydrophobic interactions, etc.) may
exist between a planar surface and a molecule some distance
away. The strength and nature of these interactions will deter-
mine the extent to which molecules adsorb to the surface and
perhaps initiate the growth of a nanomaterial. Furthermore,
surface forces play an important role in catalysis, where a
surface-bound molecule may be immobilized on the surface
in an optimal geometry for a reaction to ensue. This section
focuses on electrostatic interactions at surfaces. A more thor-
ough treatment of the subject can be found in Israelachvili’s
classic book Intermolecular and Surface Forces.
2.2.1 The Electrical Double Layer

Our discussion of electrostatic forces would not be complete
without a brief overview of the electrostatic double layer and
its role in the interactions between surfaces in liquids at the
nanoscale. The electrostatic double layer is the term given to

the diffuse layer of counterions in a solution that are associ-
ated with a charged surface. As shown below, the electrostatic
double layer plays an important role in determining the forces
that operate between charged surfaces in liquids.
When a surface comes into contact with a liquid, it may
become charged by adsorbing ions from solution or releas-
ing ions into solution. Many surfaces, for instance, are pH
labile, and at either high or low pH might become positively
or negatively charged. For example, a surface containing pri-
mary amine groups becomes positively charged at pH < 10 as
the amine groups acquire an extra proton. Another common
example of a surface-charging mechanism is the binding of
Ca2+ ions by the zwitterionic head groups of many phospho-
lipid bilayers, resulting in a positively charged surface.
In solution, we would expect the charges on a surface to
be balanced by the appropriate counterions that have been
released from the surface itself or drawn in from the surround-
ing solution. Indeed, this is the case and the result is the for-
mation of two regions of counterion charge to neutralize the
surface charge. The first region is a compact layer of counteri-
ons that is closely bound to the charged surface. This compact
region of bound counterions is called the Stern or Helmholtz
layer. It should be noted that the counterions in the Stern layer
are not necessarily irreversibly bound to the surface and can
often be exchanged with those in the surrounding solution.
The second region is a more diffuse and extensive layer of
counterions that is in rapid equilibrium with the surround-
ing solution. This region is referred to as the electrical double
layer or the diffuse electrical double layer and is the focus of
our discussion. These regions are shown schematically in
Figure 2.11.
A common example of the electrical double layer can be
found in the colloid milk. As a mixture of primarily non-polar
butterfat droplets in water, milk particles would seem to be
expected to aggregate and coagulate into butter due to hydro-
phobic interactions. However, trace amounts of the highly
polar phosphoprotein casein at the water-milk interface result
in an electrical double layer forming around each milk par-
ticle. This double layer creates enough repulsion to overcome
the hydrophobic particle’s tendency to aggregate. Inks, paints,
and blood provide further examples of heterogeneous liquid
mixtures that are stabilized by electrical double layers.
The presence of the electrical double layer is a direct con-
sequence of the tug-of-war between the energy and entropy
–
–- x
–
Diffuse electrical
double layer –
– –
– –
– – –
– – – –
Stern/Helmholtz – –
–
layer – – – – – – – – – –
+ + + + + + + + + + + + + + + + + + + + + 0
Positively charged surface
Figure 2.11 The Stern/Helmholtz layer and the diffuse electrical double layer. Ions
within the Stern/Helmholtz layer are bound to the surface, although generally not rigidly.
of the charged-surface/bulk solution system. The electrostatic

energy of the system is minimized when the charge separa-
tion is at a minimum—i.e., when the counterions in solu-
tion become closely associated with the charged surface to
the point of neutralization. The entropy, on the other hand,
is maximized when the counterions are able to move freely
through the entire volume of the bulk solution. Based on
energy considerations alone, we would not expect an elec-
trical double layer to exist—the surface charges would be
completely neutralized by counterions closely bound to the
surface. Entropic considerations, however, demand some sort
of give and take. The resulting compromise between energy
minimization and entropy maximization of the system pro-
duces the diffuse double layer with an equilibrium concen-
tration of counterions that gradually decreases with distance
away from the charged surface until it reaches a value equiva-
lent to that of the bulk solution.
The actual distribution of the counterions at equilibrium
can be calculated by the Poisson-Boltzmann equation
d 2Ψ ⎛ e ⎞

dx 2
= −⎜
⎝ ε ε 0 ⎟⎠ ∑z ρ
i
i io (
exp − zieΨ( x )/kT ) (2.18)
where zi is the valency of the i-th electrolyte (i.e., +1 for Na1+),
e is the standard unit of charge, k is Boltzmann’s constant,
and T is temperature. Ψ(x) is the electrostatic potential at a
distance x away from the surface. The zero of the potential can
be arbitrarily assigned and ρio is the number density of the ith

electrolyte at that same distance x (often chosen to be the limit
as x approaches ∞, which is equivalent to the bulk phase).
2.2.2 The Debye Length

For most situations, the solutions to the Poisson-Boltzmann
equation are rather complicated and should be obtained
numerically by a computer. However, in the limit of a very
small electrostatic potential such that zeψ(x)/kT << 1, then the
Poisson-Boltzmann equation reduces to
d 2Ψ ⎛ e ⎞
=
dx 2 ⎜⎝ ε ε 0 ⎟⎠ ∑ z ρ ( z eΨ(x )/kT )
i
i io i (2.19)
d 2Ψ
= κ 2 Ψ( x ) (2.20)
dx 2
where
1/2

⎛
⎜ ∑ρ
i
ze ⎞
2 2
i0 i
⎟ (2.21)
κ=⎜ ⎟
⎜ ε ε 0 kT ⎟
⎝ ⎠
and has units of m–1. In this case, ρi0 is defined as the number
density of the ith electrolyte in the bulk solution.
The second-order differential equation in 2.20 is called the
Debye-Hückel equation and has a well-known solution of
Ψ( x ) = Ψ0e − κx (2.22)
where Ψ0 is the potential at the charged surface.

From Equation 2.22, we see that the characteristic decay
length of the electrostatic potential for the Debye-Hückel
model is 1/κ. This length is often called the Debye length or
the Debye screening length and can be used as a rough approx-
imation for the “thickness” of the electrical double layer. If a
charge is within the Debye length, it “feels” the effect of the
charged surface, and if it is too far outside the Debye length, it
effectively is screened from the charged surface by the inter-
vening cloud of counterions.
From Equation 2.21 we also see that the Debye length is

independent of the properties of the surface itself—that is
to say for a given liquid at a certain temperature it depends
only on the concentrations and valencies of ions in the bulk
solution, not on the surface charge or surface potential. For
example, for an aqueous 100-mM NaCl solution at 25°C, the
Debye length is 0.96 nm, independent of the charge density or
potential of the surface itself.
2.2.3 Interactions Between Charged

Surfaces in a Liquid
In general, we can solve for the pressure due to the presence of
ions at a position x between two charged surfaces as
P( x ) = kT ∑ ρ (x )
i
i (2.23)
where ρi(x) is the number density of the ith electrolyte at x (mea-

sured in molecules per cubic meter). Since the distribution of
ions at any given point must be calculated using the Poisson-
Boltzmann equation, then solving the Poisson-Boltzmann
equation for the system must precede a calculation of the pres-
sure between two surfaces. However, as mentioned previously,
the Poisson-Boltzmann equation is rather complicated to solve
for most systems of practical interest. It is beyond the scope
of this text to discuss the solutions to this equation, but stu-
dents should be aware of this general approach for calculating
the pressure between two surfaces. We limit our discussion to
qualitative descriptions of the forces operating between two
interacting surfaces, descriptions resulting from the applica-
tion of the approach described previously.
In the most elementary example, suppose we have a flat,
neutral surface approaching a flat, charged surface in a par-
allel orientation, such as that shown in Figure 2.12. Before
the approach of the neutral surface, the charged surface has
associated with it a diffuse electrical double layer extending
out into solution. As the neutral surface approaches, however,
the counterions in the double layer must become confined to a
smaller and smaller volume, resulting in a decrease in entropy
of the system. The approach of the neutral surface likely
causes some of the counterions to bind to the charged surface,
resulting in a slight decrease in energy. However, this favor-
able decrease in energy is offset by the much larger decrease
in the entropy of the system. For this reason, the interaction
Neutral surface
–
–
–
– –
– –
– –
–
– – –
–
– –
–
– – – – – – – – – –
+ + + + + + + + + + + + + + + + + + + + +
Figure 2.12 The interaction between a neutral and charged surface is repulsive.

+ + + + + + + + + + + + + + + + + + + + +
– – – – – – – – – –
– –
– – – –
– –
– –
– –
–
– –
–
– – –
– – – –
– –
– – –
– – – – – – – –
+ + + + + + + + + + + + + + + + + + + + +
Figure 2.13 The interaction between two surfaces of like charge is repulsive.
between a neutral and charged surface in a liquid must always

be repulsive.
Now let’s consider the interaction between two charged
surfaces of like charge, as shown in Figure 2.13. As the sur-
faces approach each other, their electrical double layers begin
to overlap, resulting in an effective decrease in the entropy

of the system and making the interaction unfavorable. Hence,
the interaction between two surfaces of like charge is always
repulsive. By comparison, this repulsive interaction begins to
occur when the two surfaces are at a greater distance than if
one of the surfaces was neutral.
Finally, let’s consider the case of the interaction between
two surfaces of opposite charge. As we would expect, electro-
static attraction between the surfaces dominates at long range.
One surface can, in effect, operate as the “counterion” for the
apposing surface. As the surfaces approach, counterions are
released into solution and expelled from the gap between the
two surfaces, resulting in an increase in entropy of the sys-
tem. If the charge densities of each surface are equivalent, the
electrostatic attraction continues up until contact between the
surfaces. If the charge densities are not equal, some counte-
rions must remain in the gap. Gradually they become more
concentrated as the surfaces approach each other, and at some
point the repulsive force from these counterions balances the
electrostatic attraction between the surfaces.
2.3 INTERMOLECULAR FORCES AND
AGGREGATION
Supramolecular chemistry is dominated by the host of non-
covalent interactions present in molecular subunits. A simple
illustration can help us understand the interplay between the
various interactions discussed so far, and how this interplay
leads to molecular self-assembly into nanomaterials with a spe-
cific structure. Consider the set of generic molecules shown in
Figure 2.14. For simplicity, only characteristics emphasizing
interactions are shown such as ionic moieties, hydrophobic
regions, dipoles, and hydrogen bonding groups. The mole-
cules represent molecular building blocks and the aggregation
of these individual blocks will be affected by intermolecular
interactions.
The organization of the building blocks into more com-
plex structures will largely be driven by a minimization of
energy and thermodynamic constraints. The latter factor, for
example, can be an entropic gain due to the hydrophobic effect.
Minimization of energy will be achieved by minimizing unfa-
vorable interactions such as bringing two like-charged moi-
eties to the same vicinity. Possible ways the molecular building
blocks may assemble are shown in Figure 2.14(b). Self-assembly
will lead to a three-dimensional aggregate, but if the assembly
(i) (ii) (iii) (iv)

Cationic group + +
Hydrocarbon chain
Dipole
Hydroxyl group –
(a)
(i) (ii) (iii) (iv)

+ + +
– + –
+ + + +
– –
(b)
(c)
Figure 2.14 (a) Representation of four molecular building blocks containing various
interacting functionalities. (b) Possible aggregation patterns driven by (i) like-charge repul-
sion, (ii) dipole-dipole and H-bonding interactions, (iii) opposite-charge attraction, and (iv)
strong hydrophobic interactions. Hydrophobic interactions probably play a role in all of these
aggregates. (c) Strong substrate-molecule interactions cause the molecules to tilt in order to
minimize like-charge interactions.
is occurring on a surface, then a two-dimensional aggregate

will form. A planar support will also impose some restrictions
on the exact structure. For example, in Figure 2.14(c), a sup-
ported monolayer is forced into a tilted and/or staggered state
because of strong substrate-molecule interactions and inter-
molecular head group repulsions.
A thermodynamically stable structure will be formed based
on the conditions of the assembly process (temperature, the
presence of a surface, pH, concentration, etc.). If the structure
is formed in a solvent such as water, it is important to appre-
ciate that the molecules comprising the aggregate may be in
dynamic equilibrium with “free monomers” in solution. One
consequence of this is that the aggregate size and shape may
change with monomer concentration and other conditions

such as pH, temperature, and salt concentration.
Up until now, we have ignored direct electron coupling
between neighboring molecules in a self-assembled aggregate.
Although induced dipole effects are electronic in origin, the
molecules may be relatively far apart so that electrons are still
localized on each molecule. When one is confronted with a
dense aggregate comprised of molecules that are essentially
touching one another, the electronic characteristics of the mate-
rial may change. For example, molecular building blocks may
result in a nanostructure in which the electrons are delocal-
ized over the entire aggregate. This will change the material’s
optical and electronic properties, which will be determined
by the shape and size of the aggregate. The following sections
provide some background in basic electronic structure per-
taining to electron delocalization and the effect of size.
2.4 SIMPLE MODELS DESCRIBING

ELECTRONIC STRUCTURE
Electrons interact with radiation, and this interaction is
responsible for the absorption and emission of radiation.
Phenomena such as fluorescence, phosphorescence, and pho-
to-electricity depend on how light interacts with molecules.
This interaction can be exploited to gain information about
molecular structure (the basis of spectroscopy). Spectroscopy
and the nature of light-matter interactions are covered in
Chapter 4. Here some pertinent elements of electronic struc-
ture are covered. It is assumed that the student has a general
chemistry level grasp of Lewis structures, molecular orbital
(MO) theory of simple molecules, and a quantum mechani-
cal interpretation of light (photons) and electronic structure
(energy levels). One important equation worth recalling is the
Planck’s equation, which relates energy between two energy
levels (∆E) to the wavelength (λ) of light absorbed or emitted as
a result of an electronic transition between these two energy
levels (Equation 2.24).
hc
ΔE = hν = (2.24)
λ
The wavelength and frequency (ν) of light is related by λ = c/ν,

where c is the speed of light in a vacuum (2.998 × 108 ms–1) and
h in Equation 2.24 is the Planck’s constant (6.626 × 10 –34 Js).
2.4.1 The Particle in a Box Model

Chemical reactivity and physical phenomena such as the
absorption of light is largely determined by the electronic
structure in molecules. Since electronic energy levels are quan-
tized, Equation 2.24 provides the wavelength (or color) of light
absorbed or emitted due to transitions between such levels. For
example, electron configurations of atoms and molecules pro-
vide an excellent explanation of such observations. However,
in many cases an electron (or electrons) is free to move within
a certain region of space. This region may actually be an aggre-
gate that has nanoscale dimensions, such as a metal nanopar-
ticle. Metals and conjugated molecules are other examples
where electrons are not restricted to the individual nuclei, but
rather are delocalized over a larger region of space.
The particle in a box model is a quantum model that
describes the energy levels available to an electron confined to
a well-defined region of space. For simplicity we will consider
this region to be a line. The line represents a one-dimensional
region with infinite walls (known as potential barriers) that
keep the electron confined on the line (Figure 2.15). The elec-
tron does not have the energy to jump the potential barrier,
and so the probability that it is present beyond the regions
0 and a is zero. The quantum-mechanical solution of such a
model tells us that the electron can only have certain values of
energy, and the exact energy depends on the quantum number
n, which can take values 1, 2, 3, 4, and so on. These values of
energy are given by Equation 2.25, where a is the length of the
line, and m is the mass of the electron.
h2n2
E= (2.25)
8ma2
e –-
0 a
Figure 2.15 An electron confined on a line between 0 and a. At points 0 and a, the
infinite potential barriers prevent the electron from crossing over.
Example 2.3 Calculating the Energy of an Electron

in a One-Dimensional Nanoscale Region
Consider an electron that is free to move along a line of length

200 nm. What is the ground state energy of the electron? What is
the energy of the electron in the n = 3 state? What is the effect on
the spacing between neighboring energy levels if the length of the
line increased from 200 nm to 300 nm?
Solution The ground state energy represents the lowest energy
the electron can have. This is the case when n = 1. The mass of
the electron is 9.109 × 10 –31 kg and h = 6.626 × 10 –34 Js. Using
Equation 2.25,
h 2n 2 (6.626 × 10−34 Js)2 (1)2

E = =
8ma 2 8(9.109 × 10−31kg)(200 × 10−9 m)2
4.390 × 10− 67 J2 s2
= = 1.506 × 10−24 J
2.915 × 10− 43 kgm2
(note: 1 J = 1 kgm2s –2)

When n = 3, the energy is
h 2n 2 (6.626 × 10−34 Js)2 (3)2

E = =
8ma 2 8(9.109 × 10−31kg)(200 × 10−9 m)2
3.951 × 10−66 J2 s2
= = 1.355 × 10−23 J
2.915 × 10−43 kgm2
The difference between neighboring energy levels (∆E) is given by
h2
ΔE = E 2 − E1 = (n22 − n12 )
8ma 2
The subscripts 1 and 2 denote lower and upper energy levels, respec-
tively. If a increases from 200 nm to 300 nm, ∆E will decrease by
22/32 or about 45%.
In the particle in a box model the electron does not exist

as a discrete particle moving along the line. Rather it resem-
bles a standing wave whose exact form depends on the value
of n. The standing waves can be viewed as a smear of elec-
tron density, with regions of high and low electron probabil-
ity. Regions of zero electron probability are known as nodes.
Figure 2.16 shows an energy level diagram for the first few
energy levels based on Equation 2.25 and representations of the
Probability Density
Energy Level
n=5 E = 25h2/8ma2
n=4 E = 16h2/8ma2
n=3 E = 9h2/8ma2
n=2 E = 4h2/8ma2
n=1 E = h2/8ma2
0 a
(a) (b)
Figure 2.16 (a) The allowed energy levels of an electron confined on a line between 0
and a. The energy values are given as a function of the quantum number n. As n increases,
the spacing between the energy levels increases. (b) The corresponding probability density
as a function of the quantum number n. The regions of zero probability represent nodes. The
number of nodes is equal to n – 1 (ignoring the zero probability at 0 and a).
corresponding electron density for each level. If the potential

barriers were not infinite, the quantum mechanical solution
tells us that the electron can actually reside beyond the line
region, a phenomenon known as quantum mechanical tunnel-
ing. However, the probability density in the regions beyond
the line decays very rapidly.
The particle in a box model is arguably the simplest quan-
tum mechanical model describing the energy and probabil-
ity density of an electron. Despite its simplicity it has been
extremely valuable in describing the absorption properties of
simple molecules in which the electron is freely moving along
a line. Such molecules can be linear carbon chains contain-
ing alternating single and double bonds (conjugation) in which
electrons are delocalized along the entire chain.
It must be pointed out that linear conjugated molecules
have more than one free electron moving along the chain.
C C C
Unhybridized
p-orbital
C C C
containing
one electron
C-C σ-framework
CH2=CH-CH=CH-CH=CH2
(a)
n=4 LUMO n=4 LUMO
n=3 HOMO n=3 HOMO
n=2 n=2
Excitation
n=1 n=1
(b)
Figure 2.17 (a) Overlap of unhybridized p-orbitals on each carbon atom produces a
π molecular orbital resulting in the delocalization of electron density along the hexatriene
chain. (b) The particle in a box model as applied to the hexatriene molecule results in an
energy-level diagram showing electron pairs in three levels. Excitation of an electron to the
n = 4 occurs by the absorption of energy.
For example, hexatriene has 6 free electrons as shown in

Figure 2.17. Each carbon atom is sp2 hybridized, and overlap
of these orbitals forms the σ backbone of the carbon chain,
including all the C-H bonds. The electrons due to σ-bonding
are localized between the two atoms forming the bond. The
particle in a box model does not apply to these electrons.
However, each carbon atom has an unhybridized p-orbital con-
taining one electron. The p-orbitals on all the carbon atoms are
able to overlap forming the π-framework [shown by the shaded
regions in Figure 2.17(a)]. The 6 π-electrons are delocalized
along the chain and can be described using the particle in a
box model.
In order to deal with the 6 π-electrons in hexatriene, an energy-
level diagram is constructed like that shown in Figure 2.17(b).
The energy levels are filled up with the appropriate number
of electrons such that each level contains a maximum of two

electrons. The Pauli Exclusion Principle introduced in general
chemistry tells us that only two electrons can enter a given
energy level, and this pairing occurs with the electrons having
opposite spins to each other. In Figure 2.17, the 6 free electrons
in hexatriene enter the first three energy levels. Higher energy
levels (n > 3) are empty when the molecule is in the ground
state. The n = 3 level is known as the HOMO (highest occupied
molecular orbital) level and the n = 4 level is known as the
LUMO (lowest unoccupied molecular orbital) level. An elec-
tron in the n = 3 state can absorb a photon of energy and as a
result enter the n = 4 level. This is known as the HOMO-LUMO
electronic transition. Of course, the photon causing the excita-
tion must have an energy that is equal to the HOMO-LUMO
energy gap.
The particle in a box model can be expanded to describe
the electron moving in both a two-dimensional rectangular
region (Equation 2.26) and a three-dimensional cubical region
(Equation 2.27).
2
h2 ⎛ nx2 ny ⎞
E 2D = ⎜ + ⎟ (2.26)
8m ⎝ a 2 b 2 ⎠
2
h2 ⎛ nx2 ny nz2 ⎞
E 3D = ⎜ + + ⎟ (2.27)
8m ⎝ a 2 b 2 c 2 ⎠
In these equations a, b, and c represent the length, width, and

height of a cube (or just a and b for a rectangle), and the x, y,
and z subscripts denote quantum numbers in the three differ-
ent directions of a Cartesian coordinate system. These three
quantum numbers independently assume values of 1, 2, 3, 4,
and so on.
An interesting solution arises when dealing with an elec-
tron moving around a ring. The energy levels are given by
Equation 2.28.
h2
E= m2 (2.28)
8π 2I
I = me r 2 (2.29)
I is the moment of inertia of the electron going around the

ring (Equation 2.29), r is the radius of the ring, and me is the
mass of the electron. The quantum number m can take on
values of 0, ±1, ±2, ±3, and so on. This means that when m
is 1, there are two energy levels with the same energy. We
say that the electron in the m = 1 state is twofold degenerate.
In general, +m and –m represent a doubly degenerate energy
state.
This model can be used to describe the electronic structure
of benzene. There are six π electrons in benzene, one in each
unhybridized p-orbital on each carbon atom [Figure 2.18(a)].
These six electrons are delocalized around the ring and are
regarded as free electrons. Figure 2.18(b) shows an energy
level diagram based on Equation 2.28. We can place two of
them into the m = 0 level and four into the m = ±1 level.
The first electronic transition would be the 1 → 2 transition,
and the energy associated with this transition is given by
Equation 2.30.
h2 2 2
ΔE = (2 − 1 ) (2.30)
8 π 2I
C6H6
Unhybridized C C
p-orbital C C
containing
C C
one electron
C-C σ-framework
(a)
E = 4h2 8π2I m = ±2
E = h2 m = ±1
8π2I
E=0 m=0
(b)
Figure 2.18 (a) Overlap of unhybridized p-orbitals on each carbon atom produces a π
molecular orbital resulting in the delocalization of electron density along the benzene ring.
(b) The particle in a box model as applied to the benzene molecule results in an energy level
diagram showing electron pairs in three levels. The two energy levels corresponding to m =
±1 are doubly degenerate.
Example 2.4 Estimating the Size of the Benzene Ring
Benzene absorbs light of wavelength ~250 nm. Estimate the radius

of the benzene ring.
Solution We first change the absorption wavelength into the corre-
sponding energy using Equation 2.24.
hc (6.626 × 10−34 Js)(2.998 × 108 ms −1 )

ΔE = = = 7.946 × 10−19 J
λ (250 × 10−9 m)
Assuming that the 1 → 2 transition is associated with the absorp-

tion wavelength of 250 nm, we rearrange Equation 2.30 to deter-
mine the moment of inertia.
h 2 (22 − 12 ) (6.626 × 10−34 Js)2 (3)

I= = = 2.102 × 10−50 Js2
Δ E (8π 2 ) (7.946 × 10−19 J)(8π 2 )

= 2.102 × 10−50 kg m2
(note: 1 Js2 = 1 kgm2)

Finally, the radius can be estimated using Equation 2.29.
I 2.102 × 10−50 kgm2

r = = = 1.519 × 10−10 m = 0.152 nm
me 9.109 × 10−31kg
This is close to the experimentally measured value of around

0.25 nm.
The particle in a box model can be used to describe the

energy of an electron confined within a three-dimensional
region of nano-scale dimension. This is known as quantum
confinement, and examples involving quantum confinement
in solid nanoparticles are given in Chapter 5.
2.4.2 Conjugation in Organic Molecules

The free electron model can be applied to conjugated organic
molecules. In fact, the model can help us explain why the
absorption wavelength decreases as the length of the molecule
increases. We will use the term conjugation length to describe
the length of the alternating double- and single-bonded hydro-
carbon chain in which the electrons are delocalized over. First
let’s review some background on the structure of conjugated
organic molecules.
As briefly mentioned in the last section, a conjugated sys-
tem has three sp2 hybridized orbitals on every carbon atom,
which form covalent bonds with nearby atoms. This accounts

for the σ-bonding in the molecule. The leftover unhybridized
pz orbital combines with other pz orbitals to form a delocal-
ized π MO that spans the length of the molecule. For example,
ethylene has two atomic pz orbitals, ϕ1 and ϕ2. From these two
atomic orbitals, two MOs, Ψ1 and Ψ2*, are formed by taking
linear combinations. The bonding MO, Ψ1, results from the in-
phase combination of the wave functions of the two p orbitals,
whereas the antibonding orbital, Ψ2*, results from the out-of-
phase combination. The overlap results in two new MOs—one
bonding orbital with an energy lower than either of the origi-
nal p orbitals and one antibonding orbital with an elevated
energy. The relative energies of these MOs are illustrated
in Figure 2.19. It must be pointed out that only the π MOs
are shown in the figure. These MOs result from the overlap
Antibonding orbital (Ψ2*)
C C
C π π* C
Atomic orbital (φ1) Atomic orbital (φ2)
C C
Bonding orbital (Ψ1)
Figure 2.19 A partial MO energy-level diagram for ethylene emphasizing the π bonding
(Ψ1) and antibonding (Ψ2) MOs. MOs are constructed by the linear combination of the AOs
of ethylene, in this illustration the unhybridized p-orbitals (ϕ1 and ϕ2) on each carbon atom.
The two electrons from each p-orbital are placed in the bonding MO (Ψ1). The MOs due to
σ-bonding are not shown.
between unhybridized p-orbitals on the carbons. The MOs

resulting from σ-bonds are not shown, as this is not relevant
to our discussion of conjugation.
In terms of shape, the bonding MO has electron density
above and below the line connecting the two carbon atoms.
There is no node between these atoms. In contrast, the anti-
bonding MO has a node between the two carbons. The anti-
bonding MO has electron density concentrated near each of
the two carbon atoms, but zero electron density between these
atoms. The presence of a node means no bond, thus the term
a ntibonding MO.
The simple two p-orbital overlap scheme can be expanded
to conjugated systems. We start by considering 1,3-butadiene,
the simplest conjugated system. Because 1,3-butadiene has four
atomic orbitals, four molecular orbitals must result. Figure 2.20
shows the atomic and molecular orbitals of 1,3- butadiene and

C C C C
C C C C
C C C C π π* C C C C
Atomic orbitals (φ1) Atomic orbitals (φ2)
C C C C
C C C C

Figure 2.20 A partial MO energy-level diagram for butadiene emphasizing the π bonding
(Ψ1 and Ψ2) and antibonding (Ψ3* and Ψ4*) MOs. MOs are constructed by the linear combina-
tion of the AOs of butadiene, in this illustration the unhybridized p-orbitals on each carbon
atom grouped together and described as ϕ1 and ϕ2. The four electrons from each p-orbital are
placed in the bonding MOs (Ψ1 and Ψ2). The MOs due to σ-bonding are not shown.
the corresponding energy levels. Like ethylene, only the π MO

are shown, which result from the overlap between unhybrid-
ized p-orbitals on the carbons. The MOs resulting from σ-bonds
are not shown.
The important aspect of Figure 2.20 is the π → π* transi-
tions. Notice that the Ψ2 → Ψ3* transition in 1,3-butadiene is
much smaller than the Ψ1 → Ψ2* transition in ethylene. As
the number of p orbitals increases in a conjugated system, the
energy gap, ∆E, between the HOMO and the LUMO becomes
progressively smaller. Figure 2.21 shows the MO energy levels
of ethylene and the first three conjugated systems: 1,3-buta-
diene; 1,3,5-hexatriene; and 1,3,5,7-octatetraene. The vertical
arrows emphasize the HOMO-LUMO transitions. Further, the
energy difference between the HOMO and LUMO determines
the position of the absorption maximum (λmax). Since wave-
length is inversely proportional to ∆E, λmax increases with the
length of the chain of the conjugated carbon atoms.
In the extreme case of a conjugated polymer, the π bond-
ing orbitals and the π* antibonding orbitals form broad bands
separated by a small ∆E. A long conjugated polymer is able
Ψ8*
Ψ6*
Ψ4*
Ψ7*
Ψ2 * Ψ5*
Ψ6*
Ψ3*
Ψ4*
Ψ5*
Energy
Ψ3 Ψ4
Ψ2
Ψ1 Ψ3
Ψ2
Ethylene Ψ1 Ψ2
Ψ1
Butadiene
Ψ1
Hexatriene
Octatetraene
2 4 6 8
Conjugation Length (number of C atoms)
Figure 2.21 The MO energy-level diagram for a series of linear conjugated hydrocar-
bons. The figure shows how ∆E for the π → π* transition changes as the conjugation length
increases.
to conduct electricity because electrons in the valence band

(π) can be excited into the conduction band (π*). There, they
are free to move and carry charge. Because electrons in conju-
gated polymers must be excited in order to conduct electricity,
conjugated polymers make excellent semiconductors.
Alan Heeger, Alan MacDiarmid, and Hideki Shirakawa
shared the Noble Prize in Chemistry in 2000 for their discov-
ery and development of conductive (conjugated) polymers in
the late 1960s and early 1970s. They did much of their work
on polyacetylene, the simplest conjugated polymer, with
alternating single and double bonds in a linear carbon chain.
Polyacetylene, however, is very susceptible to photo-oxidation,
and more recent studies have focused on other, more stable
conjugated polymers. Derivatives of polythiophene, polyani-
line, polyfluorene, and poly(phenylene vinylene) or PPV have
similar conductivities to polyacetylene but are much more
stable against oxidative degradation. Polymers and other mac-
romolecules will be discussed in Chapter 5.
2.4.3 Aggregation and Electronic Structure

Molecular aggregates have electronic and spectroscopic prop-
erties that may be considerably different from the mono-
mer. Consider a conjugated molecule that contains a dipole
moment. We can represent this molecular building block as a
rod shown in Figure 2.22, where the arrow indicates the direc-
tion of the dipole. Let’s form a molecular aggregate in which
individual monomers are arranged in a regular fashion. We
can distinguish between two types of aggregation patterns
called the H-type and J-type aggregation.
The J-type aggregate is a one-dimensional molecular assem-
bly in which the dipole moments of the individual monomers
are aligned parallel to the line joining their centers. This is
sometimes referred to as the “end-to-end arrangement.” In
contrast, the H-type arrangement, while still a one-dimen-
sional array, is one in which the dipole moments are aligned
parallel to each other but perpendicular to the line joining
their centers. This is sometimes referred to as the “face-to-face
arrangement.” Figure 2.22 illustrates the difference between a
J-aggregate and an H-aggregate.
One of the most characteristic properties of J-type aggre-
gation is that such materials absorb higher wavelength light
with respect to the monomer absorption. We say that it is
red-shifted in the absorption spectrum with respect to the
monomer absorption (see Chapter 4). The absorption wave-
length of the H-aggregate is a little lower (or blue-shifted) with
J-aggregate H-aggregate
Unaggregated
monomers
Blue shift
Red shift
λmax
Figure 2.22 Non-covalent aggregation patterns in dipolar building blocks. The

J-aggregate is a head-to-tail arrangement and the H-aggregate is a head-to-head arrange-
ment between neighboring molecules.
respect to the monomer absorption wavelength. The energy

shift of the absorption wavelengths of the aggregates has been
explained by exciton theory. This theory will not be covered
in this text, but we can apply our understanding of conjuga-
tion length and the particle in a box models to help under-
stand this observation.
First assume that the electron motion is along the dipole
only. A J-aggregate can be considered a line along which
free electrons can move. Since the dipoles are arranged end-
to-end, the electron motion can be considered as delocal-
ized along the entire length of the aggregate. According to
Equation 2.25, the length a is much greater than the length
of an individual monomer. Therefore, ∆E is small and is
inversely proportional to wavelength (Equation 2.24), and
the aggregate absorbs light energy at a higher wavelength
compared to a monomer. Conversely, an H-aggregate will
have a short conjugation length and a larger ∆E value
corresponding to the absorption of light of a smaller wave-

length. The unaggregated monomer can be considered inter-
mediate between these two extremes, and so the ∆E value
corresponding to the excitation lies between the values for
the H- and J-aggregate.
2.4.4 π-π Stacking Interactions

We end this chapter with a brief discussion of a very weak
electron coupling interaction leading to aggregation known as
the π-π stacking interaction. These interactions occur due to
the presence of p-orbitals in conjugated ring systems such as
benzene. The net effect of such an interaction is face stack-
ing of planar rings as shown in Figure 2.23(a) for naphthalene.
Although the effect is actually insignificant in small systems
such as benzene, the interaction becomes stronger as the num-
ber of π-electrons increases. It should be pointed out, however,
that even in these systems, electrostatic forces usually over-
come π-π stacking interactions. Nonetheless, the interaction
is particularly strong in planar polycyclic aromatic molecules
containing many delocalized rings. Examples include anthra-
cene and triphenylene [Figure 2.23(b)]. For large ring systems
these interactions can be so significant that they may domi-
nate the supramolecular chemistry and determine the overall
structure of the aggregate. For example, the interaction deter-
mines the growth of organic crystals composed of such poly-
cyclic molecules.
π-stacking interactions affect the properties of polymers,
peptides, liquid crystals, and proteins. In biology, π-stacking
occurs between adjacent nucleotides, and this adds to the sta-
bility of double-stranded DNA. A related phenomenon, called
the edge-face interaction, is often observed in proteins where
the hydrogen atom of one aromatic system points perpendicu-
larly to the center of the aromatic plane of the other aromatic
system [Figure 2.23(c)]. This type of interaction is thought of
as related to the partial positive charge on the hydrogen atom
connected to the aromatic ring.
REFERENCES AND RECOMMENDED READING

• Israelachvili, J. Intermolecular and Surface Forces, 3rd
ed., 2011, Academic Press, San Diego, CA, pp. 31–139.
This is a classic textbook on the subject, but is probably
for the more advanced student. Chapters 3 through 8
are particularly relevant to this text.
(a)
(b)
(c)
H H
Figure 2.23 (a) π-π stacking between planar aromatic rings. (b) Structures of anthra-
cene and triphenylene. (c) The edge-face interaction between two planar aromatic ring
systems.
• Lowe, J. P. Quantum Chemistry, 2nd ed., 1993,

Academic Press, San Diego, CA. This book is for the
advanced student. Chapter 2 contains some excellent
applications of the particle in a box model.
• Pavia, D., Lampman, G., and Kriz, G. Introduction
to Spectroscopy, 2nd ed., 1996, Saunders College
Publishing, Orlando. The chapter on electronic spec-
troscopy provides some useful background on MO
diagrams of organic molecules. This book is an excel-
lent read for those interested in the kinds of electronic
excitations discussed in Section 2.4.2.
• For those interested in H- and J-aggregation, the follow-
ing is excellent: Kuhn, H. and Kuhn, C. Chromophore
Coupling Effects. In J-Aggregation, Kobayashi, T., Ed.,
1996, World Scientific: Singapore, 1–140.
END OF CHAPTER QUESTIONS

1. k bT is often used to measure the strength of a given
interaction, where k b is Boltzmann’s constant (1.38065 ×
10 –23 J K–1). Coulombic interactions are generally among
the strongest and the farthest reaching of the interac-
tions that we have discussed. Using the information
given in Example 2.1, determine at what distance Na+
and Cl– would have to be separated in order for V(r)Na–Cl
to be equal in magnitude to k bT at room temperature
(298K).
2. (a) Imagine that a Ca2+ ion interacts with an H2O mol-
ecule in the absence of any other molecules at a dis-
tance of 0.4 nm. Plot the interaction potential energy
as a function of θ. Assume that the only interaction
between the two species is an ion-dipole interaction
and that the distance remains fixed. (b) At what angles
does the magnitude of the potential energy reach a
maximum? Why does this make sense? What do the
positive and negative values of V represent?
3. Obtain an expression for the force between an ion and
its induced dipole in an ion-induced dipole interac-
tion. Is the force attractive or repulsive?
4. (a) Assume that two dipolar molecules (each with
dipole moment µ) are in the same plane. Show that
for all distances r, the interaction potential energy
between the dipoles is smaller (more negative) if the
two dipoles are oriented in a line rather than anti-par-
allel to each other. (b) The answer from 4(a) is slightly
misleading because it might lead us to believe that the
dipoles always prefer to orient themselves in a line
rather than anti-parallel. If the dipolar molecules are
anisotropic in shape (oblong in the direction of the
dipole), explain why the molecules might prefer to ori-
ent themselves in an anti-parallel orientation rather
than in a line.
5. Explain in your own terms why, in the absence of other
interactions, a dipole-induced dipole interaction will
always be attractive. Use diagrams if appropriate.
6. At what value of r (in terms of σ) does the Lennard-Jones
potential reach its minimum value? Mathematically,
what does σ represent?
7. Use the particle in a box model to predict the values
of ∆E and the corresponding values of λmax for the
conjugated linear hydrocarbons shown in Figure 2.21.

Assume the C-C bond length is 154 pm and the C=C
bond length is 134 pm.
8. β-Carotene is an organic molecule responsible for the
red-orange color in plants and fruits. The molecular
structure is shown below. Assuming that the length
of the molecule is 2.94 nm, determine the value of
λmax. Does the result agree with the red-orange color
observed for this molecule?
CH3
CH3 CH3 CH3
H3C
CH3
CH3 CH3 CH3
CH3
9. This question concerns the “quantum” confinement

of an electron within a nano-cube. When sodium dis-
solves in liquid ammonia, some dissociation occurs
according to the following equation:
Na → Na + (solvated ) + e − (solvated )
The solvated electron can be treated as a particle in

a three-dimensional box. Assume that the box is cubic
with an edge length of 1.55 × 10 –7 cm, and suppose
that excitation occurs in all directions simultaneously
from the lowest state (n = 1) to the first excited state
(n = 2). What wavelength, in nm, of radiation would the
electron absorb? Would you expect the solution to be
colored?
10. Discuss the possible types of aggregation patterns
when the following water-soluble molecules self-as-
semble in the aqueous phase. Your discussion should
include intermolecular contributions from van der
Waals interactions, electronic coupling effects, and
predictions about the absorption properties of the
resulting aggregates. How would molecules aggregate
if they were forced at the interface of a hydrophobic
surface and a water phase?
(a) CH3(CH2)12SO4Na
(b) HO-CH=CH-CH=CH-CH2-CO2H,
(c) BraH3N-CH=CH-CH=CH-CH2-CO2Na
(d) O(CH2)10SO3Na
N
N
where X = OH or NO2
(e) CONHC18H37
N
N
N
SO3H
11. The following molecules are identical except for the

structure of dipolar head groups. Place the molecules
in order of increasing head group dipole moment and
discuss the differences in the way these molecules
aggregate into a monolayer at the air-water interface, if
any difference exists.
O
H 3C C N N O (CH2)10 SO3–
O2N N N O (CH2)10 SO3–
NC N N O (CH2)10 SO3–
Three
Rudiments of Surface
Nanoscience
CHAPTER OVERVIEW
Surfaces and interfaces occur everywhere in nature, from bio-
logical cells to the vast expanse of oceans. They play a key
role in nanoscience since they are often used as platforms for
the growth of nanomaterials. Understanding nanomaterials
would therefore not be complete without covering important
elements of surface science. Interfaces represent the two-
dimensional plane between two different bulk phases of mat-
ter, such as oil and water. We will describe a surface as an
interface where one of the bulk phases is a gas, usually air.
Physical and chemical processes occurring at such regions
tend to be very different from corresponding processes in the
bulk phase as such processes are confined to a region in which
one dimension is on the nanoscale order. For instance, a spe-
cial interaction known as the hydrophobic effect influences
chemistry at surfaces, often responsible for the formation of
films of nanoscale thickness. In this chapter, a consideration
of hydrophobicity and the surface energy of solids and liquids
will lead to a discussion of contact angles and wetting phe-
nomena. This naturally leads to a discussion of self-assembled
monolayers and adsorption phenomena. An understanding
of how intermolecular interactions influence the adsorption
and aggregation of molecules into nanostructures (such as
micelles) is provided by considering the amphiphilic nature
of surfactant molecules.
3.1 FUNDAMENTALS OF SURFACE SCIENCE

3.1.1 The Surface Energy of Solids and Liquids
Molecules of the same type tend to experience a net attractive
interaction. In liquids, this cohesive force keeps molecules
55
%
B
$
A
Figure 3.1 Bulk and surface interactions between molecules in a pure phase (e.g., H2O).
Bulk molecule A is surrounded symmetrically by its neighbors. Surface-bound molecule B is
surrounded asymmetrically by its neighbors.
close to each other, and each molecule is symmetrically sur-

rounded by others, resulting in a zero net interaction. The
picture is considerably different on the surface since mol-
ecules here are not surrounded symmetrically by others
(Figure 3.1). At the surface a molecule experiences cohesive
forces from others in the bulk beneath it, but there are no
interactions above it. This asymmetry in force results in a
net inward pull on the surface-bound molecule normal to
the surface. This is the molecular basis of surface tension,
which is defined as the force acting parallel to the surface
and at right angles to a line of unit length on the surface. The
experimental measurements of surface tension are described
in Chapter 4.
The existence of surface tension is the reason why many
liquids, such as water, tend to spontaneously contract and
minimize the surface-area-to-volume ratio. To this end, such
liquids adopt a spherical geometry in the absence of all exter-
nal forces such as gravity. One can determine the work done in
expanding the surface of a liquid against the surface tension
forces. This work actually represents the surface free energy of
the expansion of the interface.
3.1.2 Surface Free Energy of Adsorbed Monolayers

The measurement of surface tension is perhaps the oldest
method used to characterize a nanofilm floating on the surface
of water. Later in this chapter we will encounter a class of
molecules known as amphiphiles, which are capable of self-
assembling into nanostructures of remarkable complexity.
Rudiments of Surface Nanoscience 57
Amphiphilic molecules, such as lipids and surfactants, have

a tendency to accumulate on the surface of water. The driv-
ing force behind this self-assembly process is a combination of
factors including the lack of solubility of the hydrocarbon por-
tion of the molecule in water and the tendency of the charged
or polar head group to point toward the aqueous phase. The
net result is the formation of a film composed of a monolayer
of highly oriented molecules. The closely packed monolayer
is in dynamic equilibrium with individual molecules in the
bulk aqueous phase, and so the packing density of the film
increases as the concentration of molecules in the bulk phase
increases. However, structural and thermodynamic constraints
often limit the packing density, and a saturated monolayer is
usually formed when the bulk concentration is sufficiently
high. The thickness of such a film is approximately the length
of the amphiphilic molecule (nanoscale). Due to this short
dimension, the monomolecular film is often described as a
two-dimensional nanoassembly.
Figure 3.2 shows a monolayer of the surfactant sodium
dodecyl sulfate (SDS) confined to the air-water interface.
Later in the book, we show that nanofilms at the air-water
interface are of vital importance to biological processes, and
their properties can be exploited in a number of important
applications. Thus, it is often crucial to know how well the
film is packed within both the monolayer and the cross-sec-
tional area occupied by each molecule on the liquid surface.
These factors often determine the physicochemical proper-
ties of the film.
Surface tension measurements provide a simple but pow-
erful method for determining a variety of characteristics of
adsorbed monolayers, including the density of the monolayer.
Surface tension measures the stability of a surface. If the sur-
face is relatively unstable, it has a large surface tension value
and is considered a “high energy surface.” For example, liquid
water has a high surface energy because the molecules prefer
not to be on the surface, but rather to be surrounded completely
by other water molecules and interact with them through ener-
getically favorable intermolecular hydrogen bonding.
A more rigorous way to define surface tension is to say that
it is the free energy required to transport a molecule from the
bulk phase to the surface and hence expand the area of the
surface (dA). This surface free energy (dG surf ) is given by
(3.1)
dG surf = γdA
Hydrophilic
moiety
SO4–Na+
Hydrophobic
(CH2)11 moiety
CH3
Air
Aqueous phase
Figure 3.2 A monolayer of the anionic surfactant sodium dodecylsulfate (SDS) at the
aqueous-air interface. The monolayer is in equilibrium with SDS molecules in the bulk aque-
ous phase. The hydrophobic moieties of SDS are pointing away from the aqueous phase and
the polar headgroups are buried in the aqueous phase. This orientation of the SDS molecules
on the surface stabilizes the air–water interface.
where the proportionality constant γ is the surface tension.

The units of surface tension are energy per unit area, J m–2
(or since energy can be thought of as applied force multiplied
by the distance moved, the units are also Newton’s per meter,
N m–1). These dimensions are also the same as the spring con-
stant (from Hooke’s law), which essentially measures the resis-
tance of a spring to stretch due to an applied force. Thus, we
can interpret surface tension as a measure of the resistance of
a surface to increase its area.
Example 3.1 Determining the Work Done

in Expanding a Liquid Film
Consider a simple experiment in which a film is withdrawn from a

soap solution as shown in Figure 3.3. By noting that work done is
force times displacement, and the fact that this work is the surface
(a)
dh
(b)
Figure 3.3 A film being withdrawn from a soap solution to a height dh by a

force F, using a wire frame of length L.
free energy given by Equation 3.1, show that surface tension is F/2L.
Prove that the units of surface tension are N m –1.
Solution Work must be done to pull up the frame and create a film
[Figure 3.3(b)].
Work done (dw = F dh) is equal to the surface free energy (γ dA),
where A is the area of the surface. The soap film has two sides or
surfaces, so in this example area A is actually 2A. Also, A is the
length L times the distance moved dh.
Therefore, F dh = γ 2dA = γ 2L dh.
Rearranging gives
F
γ= .
2L
Since force is measured in N, the units of γ are N m –1.

Table 3.1 Surface Tension Values of

Some Pure Liquids
γ / mN m–1
Liquid 25°C 50°C 75°C
Water 71.99 67.94 63.57
1-Decanol 28.51 26.68 24.85
Mercury 485.48 480.36 475.23
Ethanol 21.97 19.89 —
Bromine 40.95 36.40 —
Pyridine 36.56 33.29 30.03
Toluene 27.93 24.96 21.98
Benzene 28.22 25.00 21.77
Source: CRC Handbook of Chemistry and Physics,
Internet Version 2007, 87th ed., Taylor and
Francis, Boca Raton, 2007. With permission.
The larger the surface tension, the greater the resistance to

increase the surface area, which in turn fundamentally depends
on the interplay between various intermolecular interactions
occurring on the surface. Table 3.1 lists the surface tension val-
ues of a few common liquids at three different temperatures.
Water has one of the highest known surface tension values
(about 72 mN m–1 at room temperature) for the reasons men-
tioned above. However, when amphiphilic molecules are pres-
ent in water, they tend to aggregate on the surface and lower
the surface tension. For example, adding a milligram of SDS to
approximately 200 mL of pure water lowers the surface tension
by about 30 mN m–1. This decrease in surface tension occurs
because the amphiphilic molecules orient themselves at the
interface such that they expose their insoluble hydrophobic
tails to the air while keeping the polar head group buried in the
aqueous phase, thereby freeing up water molecules that would
otherwise be forced to remain unfavorably at the air-water
interface (Figure 3.2). These conditions stabilize the surface and
consequently lower the surface tension. The degree to which
the surface tension decreases depends on the structure of the
amphiphile and the packing density of the resulting monolayer.
If the concentration of the amphiphiles is low, then we can
expect that the number of molecules on the surface is also rela-
tively small. In a way, this dilute surface approximates a gas-
eous phase in which individual molecules are far apart and are
free to move over the surface in a random fashion. As the con-
centration increases, however, the packing of the amphiphilic
molecules at the surface becomes denser and the surface tension

consequently decreases. This behavior continues with rising
concentration until the point at which a saturated monolayer
is formed. Beyond this concentration the surface tension value
does not change. The phase behavior of amphiphiles at the air-
water interface is discussed in Chapter 5.
3.1.3 Contact Angles and Wetting Phenomena

When a drop of water is placed on a planar solid surface, it may
at one extreme completely spread to cover the entire surface, or at
the other extreme form a spherical droplet on the surface. These
situations represent either complete wetting or complete de-wet-
ting. Usually the degree of wetting is intermediate between these
extremes and depends largely on the interfacial energy (or sur-
face tension) between the liquid and the solid surface. The con-
tact angle is the angle at which the liquid-vapor interface meets
the solid surface (Figure 3.4). If the planar surface is horizon-
tal and the droplet is not moving, this angle is called the static
contact angle [Figure 3.4(a)]. If the liquid is in motion because
the surface is tilted, then we can identify two “dynamic” con-
tact angles, the advancing contact angle and the receding contact
angle [Figure 3.4(b)]. This picture is similar to a raindrop running
down the surface of a window. Usually the advancing contact
angle is much larger than the receding angle, and the difference
between the two values is called contact angle hysteresis.
Figure 3.4(a) shows a non-wetting drop (conventionally
called a “sessile drop”) on a planar solid surface making a
contact angle, θ, which must be greater than zero. The vari-
ous interfaces are described by their surface tensions: the liq-
uid-vapor tension (γLV), the solid-vapor tension (γSV), and the
solid-liquid tension (γSL). The Young equation provides a rela-
tionship between these various surface tensions and the static
contact angle:
γSV = γSL + γLV cosθ (3.2)
Since γLV and θ can easily be measured (see Chapter 4), the
value of γSV – γSL can be determined. If the liquid completely wet-
ted the solid (θ = 0°), then the value of γSL = 0. This is a hypothet-
ical situation in which there is no “tension” between these two
phases. In fact a plot of cosθ versus γLV is linear with a negative
slope. The line can be extrapolated to the value cosθ = 1 and
the corresponding surface tension measured (on the x-axis).
Cosθ = 1 corresponds to a contact angle of 0°, or complete wet-
ting. The corresponding surface tension is called the critical
Liquid
γLV
Vapor θ γSL γSV
Solid
(a)
θr
θa
(b)
(c)
Figure 3.4 A liquid drop on a solid surface. (a) The drop is non-wetting with a contact
angle θ. (b) The drop is running down the solid with advancing (θa) and receding (θr) contact
angles. (c) Complete spreading. The solid dark lines in (a) highlight the various interfaces
with different surface tensions: the liquid-vapor tension (γLV), the solid-vapor tension (γSV),
and the solid-liquid tension (γSL).
surface tension (γc). Since under these conditions of complete

wetting γSL = 0, the Young equation tells us that γc = γSV.
It is worth noting that surfaces can be hydrophobic (water
hating) or hydrophilic (water loving). The surface of glass,
for example, is considered hydrophilic and Teflon hydropho-
bic. Water will not spread on hydrophobic surfaces, but will
instead form a droplet with a relatively large contact angle.
Conversely, water will spread on a hydrophilic surface since it
has a strong affinity for that surface, forming a thin film with
a very small contact angle.
3.1.4 Nanomaterials and Superhydrophobic Surfaces

If the contact angle of a sessile drop approaches 180°, the
drop essentially adopts its spherical geometry on the surface
and moves around much like a frictionless bearing on the
Figure 3.5 A sessile drop on a superhydrophobic surface. (Reproduced with per-

mission from Ma, Hill, Lowery, Fridrikh, Rutledge. “Electrospun Poly(Styrene-block-
dimethylsiloxane) Block Copolymer Fibers Exhibiting Superhydrophobicity.” ACS Langmuir
2005, 21:5549–5554.)
surface. Figure 3.5 shows a spherical droplet of water on a

hydrophobic surface. In order for a water droplet to behave
this way, the surface needs to be one that minimizes the
area of the solid-water interface, or a “superhydrophobic”
surface. Creating such surfaces has been a fascinating chal-
lenge in nanoscience.
Superhydrophobic surfaces are not new technology by any
means—scientists have been experimenting with superhydro-
phobicity for nearly a century, and they have been entranced
by the superhydrophobicity of natural materials like the lotus
leaf for far longer. However, new developments as well as new
potential applications for their use have spurred the popular-
ity of superhydrophobic surfaces. It is important, therefore,
to understand the basic types of superhydrophobic surfaces,
their specific characteristics, and the potentials for the use
and development of each method of fabricating these surfaces.
All of these concepts were outlined in the popular review arti-
cle, “Progress in Superhydrophobic Surface Development,” by
Paul Roach, Neil Shirtcliffe, and Michael Newton, 2008.
First, a basic understanding of the mechanism behind supe-
rhydrophobicity is important in understanding its applica-
tions. There are two basic states of superhydrophobic surfaces,
each governed by separate equations and having separate
characteristics. These states are illustrated in Figure 3.6. The
(a) (b) (c)
Figure 3.6 The various wetting states on rough surfaces: (a) Wenzel state, (b) Cassie-
Baxter state, and (c) intermediate wetting.
first type was outlined by Wenzel in 1936 and describes a wet-

ting state in which water rests upon a surface whose morphol-
ogy has been altered so that in a given area, water is in contact
with more surface than if the surface were completely flat.
Wenzel states are described by the equation
cosθW = rcosθ (3.3)
where θ is the contact angle on an unmodified surface, θW is the

Wenzel contact angle (on the rough surface), and r is the ratio
of the actual surface area of the substrate to the projection of
that surface onto a horizontal surface. In other words, r is the
ratio of the actual surface area to what the surface area would
be if the substrate were completely smooth. This equation
essentially states that Wenzel wetting increases the contact
angle of a drop of water by creating more hydrophobic surface
with which the water can interact. However, if the surface is
hydrophilic (θ < 90°) then Wenzel wetting actually increases
the hydrophilic properties of the surface. Also, because water
is present between perturbations, water in a Wenzel state is
less likely to roll off the substrate than on a flat surface of the
same material. Thus, Wenzel wetting relies on two factors—an
already hydrophobic substrate, and an increase in the surface
area of that substrate.
The other major type of wetting was discovered by Cassie
and Baxter in 1944 and describes a wetting state in which
water rests upon nano- or micro-sized “pins” with air spaces
in between. Unlike Wenzel wetting, Cassie-Baxter states have
water not touching a significant portion of the substrate, but
rather relying on the air gaps in between the substrate to form
a superhydrophobic surface. In fact, Cassie-Baxter surfaces

can be fabricated with hydrophilic substrates and still display
superhydrophobicity. Unlike Wenzel surfaces, water in Cassie-
Baxter states rolls off surfaces much more easily than on a flat
surface of the same material. These properties are reflected in
the equation governing Cassie-Baxter wetting:
cosθC = φs(cosθ) + (1-φs)cosθX (3.4)
where φ represents the fraction of surface present at the top

of the protrusions (where the water is in contact with the sub-
strate), (1-φ) represents the fraction of air gaps, and θX is the
contact angle over the air gaps, which is approximated at 180°,
and θC is the Cassie-Baxter contact angle (on the rough sur-
face). Thus, there are two ways to increase superhydrophobic-
ity of a Cassie-Baxter surface—either increase the value of θ by
increasing the inherent superhydrophobicity of the substrate,
or decrease φs by making bigger air gaps.
Interconversion between these two states is possible, but
there is an energy barrier to overcome much like any transi-
tion state between two energy minima. This phenomenon
comes into play when Cassie-Baxter surfaces display Wenzel
wetting, which is possible if water falling from a great dis-
tance (i.e., rain) is forced into the crevices between roughness
features. The transition from one state to another is important
when discussing some methods of creating superhydrophobic
surfaces.
The final important feature of superhydrophobic surfaces
that must be mentioned before a more thorough description
of the specifics is the advantage of multiple-scale rough-
ness, in which perturbations range from the nanoscale to the
microscale. Recent study has shown this kind of surface to
increase the ease with which drops roll off the surface (in
other words, decreasing the contact-angle hysteresis), the pre-
vention of conversion from Cassie-Baxter to Wenzel states, and
the tendency to convert from Wenzel to Cassie-Baxter states.
This property is beneficial for a variety of reasons and will be
revisited later.
Because the review article “Progress in Superhydrophobic
Surface Development” specifically outlines the major methods
of creating superhydrophobicity, we will refrain from repeat-
ing a list of specific methods and rather focus on unifying
themes and methods of fabrication as well as the most salient
features of each major method. One such feature belongs to the
first method described by this article—fibrous hydrophobic
surfaces. This method can span anything from carbon nano-

tube-coated cotton to nanoscale polymer electrospinning. It is
versatile, as fibrous materials are inherently rough, and coat-
ing with hydrophobic materials or nano-rods can increase this
roughness. Very high contact angles have been reported using
these methods. Also, some researchers have even reported
surfaces having self-cleaning properties that are dirt and oil
resistant while producing high contact angles and low hyster-
esis. This technology appears to be promising for the devel-
opment of water- and contamination-resistant clothing that
would essentially wash itself.
Another interesting feature that is unique to a specific fab-
rication method is the ability of certain hydrophobic semicon-
ductors to be crystallized into a superhydrophobic surface that
exhibits superhydrophobicity in the dark but superhydrophi-
licity in the light. One likely explanation for this phenomenon
is that superhydrophilicity is generated by the excitation of
electrons on the surface, which suggests that there are surface
properties that can be changed by applying a voltage. Instead of
excitation due to ultraviolet (UV) radiation, a voltage could be
used, thus allowing for materials that can be dried instantly at
the flick of a switch, or wetted completely with another flick.
Several overarching themes appear in this article as well.
One of the most notable is the generation of fractal solids as
superhydrophobic surfaces, which may be achieved in many
ways. One common approach is crystal growth, in which a
material (both organic and inorganic crystalline growths have
been reported) is cooled or condensed into a fractal solid.
Another common method is the diffusion-limited growth pro-
cess, in which a material deposits so quickly onto a substrate
that the growth of the film is limited by the flow of material
over the substrate. This method, by randomness, creates small
disturbances on the surface of the flat substrate, which then
cause more particles to attach to these perturbations. The pro-
cess continues until large bumps form, which then have small
disturbances on their surfaces that create sub-bumps, and so
on. The greatest benefit of these kinds of assemblies is that,
being fractal, these surfaces have multiple-scale roughness.
The main bumps can act as Wenzel-type surfaces, while the
smaller bumps act as Cassie-Baxter surfaces. As is expected,
very low hysteresis and very large contact angles have been
reported on fractal solids. Another important benefit of fractal
solids is their ease of fabrication—because these surfaces are
usually created by chaotic movement of material, they require
less specialized machinery and can be made with a variety
of materials. In fact, fractal superhydrophobic surfaces have

been reported using semiconductive polymers, giving these
particular surfaces the ability to switch properties as men-
tioned before.
One very common downside of fractal solids, however,
is their tendency to be opaque. Their multi-scale roughness
is very beneficial to their superhydrophobic properties, but
because these features range from nanoscale to microscale,
they interact with visible light. Many superhydrophobic sur-
faces need not be transparent. However, this kind of surface
could very well be used for windshield coatings on cars, or
as protective layers for photoactive devices such as LEDs or
photovoltaic cells. All of these applications would be rendered
useless if no light could penetrate to or from the surface.
This problem is not present in another type of superhydro-
phobic surface fabrication—uniform nanostructures. These
structures can be generated in many ways, from carbon nano-
tube growth to lithographic etching. These structures are the
opposite of fractal patterns in that, though they are sometimes
random in arrangement, they are generally uniform in size.
These materials can be easily generated using laser lithography
as well as chemical etching to produce surface morphologies
of a specific predetermined shape and pattern. Also, these sur-
faces can be easily “decorated” with additional coatings and
features to produce increased hydrophobicity. Structures of
this type have been reported to reach contact angles up to 178°
using dodecanoic acid–coated, cobalt hydroxide nanopins. As
well, significant (168°) though not quite as substantial Wenzel
contact angles were reported using decorated carbon nano-
tubes. Perhaps most importantly, these kinds of structures are
of one size and thus do not often absorb in the visible range.
This feature makes these kinds of structures more viable for
the kinds of surfaces previously discussed. The drawback of
these kinds of materials is that they are much less effective at
repelling water than fractal structures. They lack multi-scale
roughness, and even methods that generate more random
arrays are less hydrophobic than fractal surfaces.
What has emerged in recent years has not been the develop-
ment of one superior form of constructing superhydrophobic
surfaces. Rather, many different methods have developed, each
with its own benefits and drawbacks. As more applications
for superhydrophobic surfaces appear—clothing, photoac-
tive devices, windows, and more—the drive is for a variety
of superhydrophobic materials that can be customized to fit
these multiple needs.
3.2 ADSORPTION PHENOMENA:
SELF-ASSEMBLED MONOLAYERS
A molecule approaching a surface may experience a net attrac-
tive force and consequently become trapped or confined at the
surface. Such a species is called the adsorbate, and adsorption
is the physical process by which adsorbate molecules accu-
mulate onto a solid surface. Desorption is the opposite pro-
cess in which molecules leave the surface and enter the bulk
phase. The solid surface in question is referred to either as the
substrate or an adsorbent. The former is usually a planar sur-
face and the latter can be a high-surface-area porous solid. The
sticking of a reactant molecule to a surface was first proposed
by Michael Faraday in 1834 as the initial step of a surface-cat-
alyzed reaction. In this section we are interested in the kind
of adsorption that results in a monolayer and has a thickness
on the order of nanometers. The surface itself does not need to
be flat but may be rough or even porous. The extent of adsorp-
tion depends not only upon the types of intermolecular forces
involved (van der Waals, electrostatic, hydrogen bonding), but
also on the surface area; the greater the surface area of the
substrate, the greater the extent of adsorption.
The best adsorbents are those with large total surface areas,
such as silica gel (SiO2, surface area >1000 m2/g) and activated
carbon. Silica gel is commonly used in chromatographic col-
umns to enhance the separation of solute mixtures by taking
advantage of the different degrees of adsorption of the various
components. An interesting example of an adsorption phe-
nomenon occurs in polar stratospheric clouds in the upper
atmosphere, which are highly porous and act as substrates for
the adsorption of gases such as HCl.
In this and the next section we focus on the adsorption of
molecules from solution or gaseous phase onto a solid sub-
strate. However, it is important to appreciate that adsorption
is a surface phenomena and can occur at any interface. For
example, molecules may adsorb from an aqueous solution to
the aqueous-air interface. Liquid-liquid interfaces, such as the
boundary between an oil and water phase, also represent regions
at which adsorption may occur. Surface adsorption plays a cen-
tral role in the formation of nanomaterials. Adsorbents pres-
ent themselves as platforms for the self-assembly of molecules
into nanostructures. Specific examples in which nanomateri-
als are synthesized this way are presented in Chapter 5. At
this point it is worth mentioning that solid substrates can be
chemically modified so that adsorption can be selective. This
modification is known as surface functionalization and may

be as simple as oxidizing a metal surface to change the surface
energy (or to render it hydrophilic), so that polar molecules are
spontaneously adsorbed to the surface.
Many solids have the property of adsorbing large quantities
of gases and solutes from liquid solutions. This process is gen-
erally very specific both with respect to the adsorbent and the
material adsorbed, driven largely by thermodynamic consider-
ations. Adsorption is usually an exothermic process and can
be divided into two kinds: chemisorption (chemical adsorp-
tion) and physisorption (physical adsorption). In general, if
adsorption is specific and if large amounts of heat are liber-
ated (greater than about 50 kJmol–1), the adsorption process is
referred to as chemisorption (first proposed by Irving Langmuir
in 1916). In this process, bonds are broken in the adsorbate mol-
ecules and new covalent bonds are formed between the adsor-
bent and the adsorbate until a complete monolayer has been
established. The resulting substrate-adsorbate bond strengths
range from 200 to 500 kJmol–1. The chemisorbed monolayer is
irreversibly bound to the solid surface and usually changes
the surface properties of the solid substrate. Thus, chemisorp-
tion is an excellent method of chemically functionalizing a
solid surface. As an example of chemisorption, consider the
molecule octadecanethiol, or ODT, shown in Figure 3.7. The
molecule contains 18 carbon atoms (17 methylene groups
and one methyl group) terminated by a thiol group (SH). The
thiol group is extremely reactive toward gold, resulting in a
strong Au-S covalent bond. Thus, by placing a gold-coated sub-
strate into a solution of ODT in a solvent such as chloroform,
CH3(CH2)17SH + Au surface SAM
CH3CH3CH3CH3CH3CH3CH3CH3CH3CH3
CH3
S S S S S S S S S S S
Gold coated substrate
SH
Figure 3.7 The formation of a self-assembled monolayer by the chemisorption of octa-

decane thiol on a gold surface.
molecules of ODT spontaneously chemisorb to the Au surface

and form a tightly packed monolayer within hours. A few fea-
tures regarding this process are worth noting. First, adsorp-
tion is rapid, irreversible, and stops after a complete monolayer
is formed. Second, in addition to the strong Au-S bonds that
are formed, the strong van der Waals interactions between the
neighboring alky chains in the ODT monolayer allow the mole-
cules to pack very tightly and force the hydrocarbon backbone
into an all-trans conformation (Figure 3.7). The resulting film
is solid in nature, has a thickness almost the length of a fully
extended ODT molecule, and is referred to as a self-assembled
monolayer, or SAM. Finally, the SAM renders the solid sur-
face hydrophobic because the surface now contains a very high
density of closely packed methyl groups. Consequently, a drop
of water on this surface has a large contact angle (>110°).
If adsorption is nonspecific and if only small amounts of
heat are liberated, comparable to the heat of vaporization of the
adsorbed material, the process is physisorption. The interac-
tion between the adsorbent and adsorbate is much weaker (~20
kJmol–1) in a physisorbed film compared to a chemisorbed film.
It should be noted that although physisorption is considered
nonspecific, thermodynamic considerations are important
when considering this process. As an example, we consider
the physisorption of a long-chain alcohol, such as dodecanol,
from an aqueous solution onto a solid hydrophobic surface.
Let the hydrophobic surface be composed of an ODT mono-
layer chemisorbed to a gold surface. The adsorption process
is driven by a strong tendency of the alcohol to avoid water
(dodecanol is only sparingly soluble in water) and to associate
with any accessible hydrophobic surface. Thus, the dodecanol
adsorbs to the ODT SAM (Figure 3.8). The resulting dodecanol
monolayer is closely packed with the OH headgroup orien-
tated toward the water phase. Like the ODT SAM, the alkyl
chains of dodecanol are closely associated with each other
due to favorable van der Waals interactions, and form a solid-
like SAM on the surface of the hydrophobic film. Unlike the
ODT, the adsorption of dodecanol is reversible. The molecules
in the film are in dynamic equilibrium with the dodecanol
molecules in the aqueous bulk phase. Depleting the number
of dodecanol molecules in the bulk phase reduces the packing
density in the monolayer at the solid-aqueous interface.
Adsorption characteristics such as temperature, bulk con-
centration dependence, and reversibility are often used to
distinguish chemisorption and physisorption. The energetic
difference between the two adsorption processes is nicely
CH3(CH2)17OH
OH OH OH OH OH OH OH
CH3 CH3 CH3 CH3 CH3 CH3 CH3
CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3
S S S S S S S S S S S
Figure 3.8 The formation of a physisorbed monolayer of dodecanol on top of a SAM

composed of octadecane thiol on a gold surface.
illustrated in the one-dimensional Lennard-Jones potential

energy curve for the adsorption of a diatomic molecule on a
planar surface. V(x) = 0 at large distances corresponds to zero
interaction between the substrate and adsorbate molecule at
infinite separation. As the molecule approaches the surface
(going from point A to B in Figure 3.9), there is a negative attrac-
tive potential between the surface and the adsorbate molecule.
The potential reaches a minimum at distance B in the case
of physisorption and at distance C in the case of chemisorp-
tion. These distances represent the corresponding equilib-
rium substrate-adsorbate bond lengths for the two processes.
At distances less than B for physisorption (or less than C in the
case of chemisorption), the attractive interaction lessens until
V(x) = 0, below which repulsive interactions start to become
C B A
V(r)
0 r
Physisorption
P
ΔE
Chemisorption
Figure 3.9 A one-dimensional Lennard-Jones potential energy curve for the chemisorp-
tion and physisorption of a molecule on a planar surface. A, B, and C represent various
distances between the surface and the molecule.
more important. The minimum potential (at distances B and

C) for chemisorption is at a smaller substrate-adsorbate dis-
tance compared to physisorption because the former processes
lead to a shorter bond distance between the substrate and
adsorbate. Furthermore, the potential well for chemisorption
is deeper than for physisorption due to a stronger substrate-
adsorbate bond strength than that formed in physisorption,
with the former process involving the breaking and making
of covalent bonds. Sometimes a molecule can be trapped in a
physisorbed state before being chemisorbed. In this case the
physisorbed molecule is a precursor to chemisorption. The
two potential energy curves shown in Figure 3.9 cross at a
distance represented by point P. This is the point where the
physisorbed precursor can “cross over” to the chemisorbed
state. ∆E represents the activation energy in going from the
physisorbed state to the chemisorbed state.
The adsorption capacity of solid surfaces is determined from
measurements of the mass (or moles) of material adsorbed and
the area available for adsorption per unit mass of adsorbent.
Usually the amount of solute adsorbed from solution per unit
mass of adsorbent depends upon the solute bulk concentra-
tion up to its saturation point. Also, the amount adsorbed per
unit mass of adsorbent for a given solute bulk concentration

decreases with increasing temperature. If the adsorption is
physical, it is reversible and the solute leaves the surface of
the adsorbent to reestablish equilibrium when the adsorbent
is removed from a solution in equilibrium and placed in a
solution of lower concentration. Such reversibility may not
be shown if the process is chemisorption. The following sec-
tion describes how adsorption capacity can be measured and
how information, such as monolayer coverage and substrate
surface area, can be extracted by measuring the adsorption
capacity.
3.2.1 Simple Adsorption Isotherms

An adsorption isotherm is a plot of surface coverage (i.e., the
fraction of the adsorbent surface covered by the adsorbate) ver-
sus the bulk concentration. The “isotherm” term means that
these measurements are carried out at constant temperature.
Figure 3.10 shows a few simple adsorption isotherms. The line
that trails off to a constant surface coverage [Figure 3.10(a)] rep-
resents a situation in which adsorption stops because the sur-
face is saturated by the adsorbate. This typically corresponds
to monolayer coverage. In Figure 3.10(b), adsorption seems
to continue indefinitely as a function of bulk concentration.
Surface Coverage
(a)
(c)
(b)
Bulk Concentration
Figure 3.10 Adsorption isotherms describing (a) Langmuir adsorption, (b) multilayer
adsorption, and (c) cooperative adsorption.
Adsorption in this case does not stop after a saturated

monolayer has formed; instead, adsorption continues and mul-
tilayers are formed on the surface. In Figure 3.10(c), adsorption
seems to increase exponentially with bulk concentration. This
situation occurs if the presence of the adsorbate at the surface
promotes further adsorption due to favorable intermolecular
interactions. This kind of adsorption is known as cooperative
adsorption. It may or may not eventually level off at very high
bulk concentrations.
If the bulk phase is a gas, then the adsorption isotherm will
be a plot of surface coverage as a function of gas pressure at a
constant temperature. Adsorption isotherms allow us to deter-
mine some important characteristics of the adsorption process
such as equilibrium constants, the number of adsorption sites
available for adsorption on the substrate, and the enthalpy of
adsorption.
The Langmuir adsorption isotherm, presented by Langmuir
in 1918, is the simplest adsorption isotherm. The Langmuir
adsorption isotherm assumes that there are a finite number
of adsorption sites on the solid substrate and that each site is
taken up by a single adsorbate molecule. During the adsorp-
tion processes, these sites are taken up until we reach a point
of saturation in which all sites have been taken up by adsorbate
molecules. This point represents complete monolayer cover-
age, and further adsorption ceases. Furthermore, the Langmuir
adsorption isotherm assumes that there are no interactions
between the adsorbate molecules during the adsorption pro-
cess, and that the enthalpy of adsorption is independent of
surface coverage. Figure 3.10(a) represents a Langmuir adsorp-
tion profile.
By representing the solid surface by S(s) and the adsorbate
molecules in a gas phase by A(g), we can write the following
equilibrium between the adsorbate and the substrate:
ka
( ) ()
A g + S s AS s
kd
() (3.5)

The constants ka and kd are the rate constants for the adsorp-
tion and desorption process, respectively. At equilibrium, the
forward and reverse rates are the same, and so from basic
kinetics,
ka[A][S] = kd[AS] (3.6)

or
ka [ AS ]
= =K (3.7)
kd [ A ][ S ]

where K is the equilibrium constant for the adsorption pro-

cess. The adsorbate surface contains a certain concentration
of sites available for adsorption. Let β be the concentration
of sites per square meter and let θ be the fraction of surface
sites occupied by the adsorbate molecules. The concentration
of occupied sites is therefore given by βθ, and the concentra-
tion of free available adsorption sites is given by β – θ β = (1-θ)
β. The rate of desorption (νd) is proportional to [AS], which in
turn is proportional to the number of occupied surface sites.
Furthermore, the rate of adsorption (νa) is proportional to [A]
[S], which in turn is proportional to the number of available
sites and the number density of molecules in the bulk phase.
Thus,
νd = kdθβ (3.8)
νa = ka(1–θ)[A] (3.9)
At equilibrium, these two rates are equal, so
kdθβ = ka(1–θ)[A] (3.10)
Rearranging gives
1 kd 1
= 1+ = 1+ (3.11)
θ ka [ A ] K [ A]

For a gas phase adsorption process, [A] represents the

concentration (number of molecules per unit volume) of the
adsorbate gas. Assuming an ideal adsorbate gas, the concen-
tration can be expressed as a pressure by use of the ideal gas
equation, PV = nRT.
number of molecules nN A nN A P N A P P
[ A] = = = = =
V V nRT RT kT
(3.12)
where P is the pressure of the adsorbate gas, and k (= R/NA) is

the Boltzmann constant. Equation 3.11 can be rearranged to
Equation 3.13.
1 kT 1
= 1+ = 1+ (3.13)
θ KP aP
where a = K/kT. Equation 3.13 is the Langmuir adsorption

equation. A plot of 1/θ versus 1/[A] or 1/P yields a straight line
with a slope equal to the reciprocal of the equilibrium con-
stant (or kT/K).
The fraction θ can be related to the following ratio:
V
θ= (3.14)
Vm

In this equation, V is the volume of the gas adsorbed to the
surface, and Vm represents the volume of gas corresponding
to monolayer coverage. Complete monolayer coverage corre-
sponds to θ = 1, and V = Vm. Equation 3.14 can be substituted
into Equation 3.13, giving Equation 3.15.
1 1 1
= + (3.15)
V aPVm Vm

Example 3.2 The Adsorption of Nitrogen onto a Mica

Surface Follows the Langmuir Adsorption Model
The following data were collected at 273.15 K. Determine the value

of a, Vm, and the total number of surface sites. What would the con-
centration of surface sites be if the mica substrate was a square of
length 2 cm?
P/10–12 torr 2.50 1.32 0.48 0.30 0.20

V/10–8 m3 3.40 2.92 2.00 1.54 1.25
Solution According to Equation 3.15, a plot of 1/V versus 1/P will

have a slope of 1/aVm and an intercept of 1/Vm. Figure 3.11 shows
this plot. A linear fit to the data yields the equation y = 1 × 10 –5 x
+ 3 × 107.
The intercept on the y-axis is equal to 3 × 107 = 1/Vm, and so
Vm = 3.3 × 10 –8 m3.
From the slope 1 × 10 –5 = 1/aVm, and so a = 3.0 × 1012 torr–1.
To determine the total number of sites on the mica surface, we first
need to determine the number of molecules corresponding to Vm.
9×107
6×107
7×107
8×107
5×107
1/[V(m3)]
4×107
3×107
2×107
1×107
0
0 1×1012 2×1012 3×1012 4×1012 5×1012 6×1012
1/[P(torr)]
Figure 3.11 A plot of 1/V versus 1/P for the adsorption of N2 on mica at
273.15 K.
First, we know that under these conditions, 1 mol of gas occupies

2.24 × 10 –2 m3.
The number of moles corresponding to Vm is
3.3 × 10−8 m3
= 1.47 × 10−6 mol
2.24 × 10−2m3mol-1
The number of molecules is simply NA times the number of moles.
(6.022 × 1023 mol–1)(1.47 × 10 –6 mol) = 8.85 × 1017 molecules
Now each N2 molecule occupies a single site on the mica surface.

Thus, there must be 8.85 × 1017 sites on the surface.
For a square of length 2 cm, the concentration of surface sites is
simply the ratio of the total number of sites to the total surface area
of the square.
8.85 × 1017molecules
= 2.21 × 1021m−2
(0.02m)2
3.2.2 Other Useful Adsorption Isotherms

The Langmuir isotherm makes a number of assumptions that
may not apply to all adsorbates and adsorbents. For instance,
strong intermolecular interactions may cause deviations
from Langmuir adsorption behavior. Many systems dis-
play multilayer adsorption. The Brunauer Emmett Teller, or
BET, isotherm is useful for modeling such behavior. The key
assumptions made in the BET isotherms are that physisorption
on a solid is infinite, that there are no inter-layer interactions
in the multilayer film, and that each layer can be described by
the Langmuir model. The BET isotherm takes the form shown
in Equation 3.16.
V cz
θ= = (3.16)
Vm ( ) ( )
1 − z ⎡⎣1 − 1 − c z ⎤⎦
The term z = P/P*, where P* is the vapor pressure above

the adsorbate before any adsorption occurs. The constant c is
related to the enthalpy of adsorption of the first layer (∆H1) and
the enthalpy of vaporization of the subsequent layers (∆Hvap)
(Equation 3.17).
⎛ ΔH 1 − ΔH vap ⎞
c = exp ⎜ ⎟ (3.17)
⎝ RT ⎠

Figure 3.12 shows how the form of the isotherm changes for

different values of c. When c >> 1, the isotherm simplifies to
Equation 3.16 and is useful in describing surfactant adsorp-
tion (Section 3.3).
V 1
= (3.18)
Vm 1 − z
There are many useful isotherms, each one modeling

a specific kind of adsorption. For example, the Temkin iso-
therm, θ = c1 ln(c2P), where c1 and c2 are constants, describes
adsorption for systems in which ∆H1 changes with pressure
or concentration of the adsorbate. The Freundlich isotherm,
θ = c1 P1/C2, describes adsorption involving strong intermo-
lecular interactions. In Section 3.3.2, the Gibb’s adsorption
equation is introduced, which allows adsorption at the solid-
aqueous interface to be modeled reasonably well.
C1 > C2 > C3
c1
c2
V/Vm
c3
z = P/P*
Figure 3.12 The effect of changing the parameter c on the BET isotherm.
3.3 SURFACTANT CHEMISTRY

The word surfactant is an acronym for surface-active agent.
Surfactants describe a class of molecules that have a tendency
to adsorb to surfaces and interfaces and lower the interfacial
tension. For instance, the air-water interface has a surface ten-
sion of 72.8 mN/m. The addition of 10% NaOH to the water
increases this value to 78 mN/m. However, adding a typical
surfactant to the water at relatively low concentrations (~mM)
can lower the surface tension to 20 mN/m. Interestingly, the
surface tension between an aqueous surfactant solution and
an oil such as heptane can be as low as 1 mN/m.
To rationalize the effect surfactants have on surface tension,
we need to consider the structure of these molecules. They are
amphiphilic, which means that one part of the molecule is sol-
uble in a specified fluid (the lyophilic part) and the other part
of the molecule is insoluble (the lyophobic part). If the fluid is
water, then the soluble part is called the hydrophobic part, and
the insoluble part is the hydrophilic part. Thus, the structure
of a surfactant has a region that is largely non-polar, typically
hydrocarbon or fluorocarbon in nature, and another region that
is polar, charged, and interacts strongly with water. These two
regions are commonly referred to as the hydrophobic tail (or
chain) and the hydrophilic head group, respectively. Figure 3.13
shows the structure of some common surfactants, including
some naturally occurring ones such as phospholipids (DMPC).
Hydrophobic chain Hydrophilic head
O–Na+
O S O Anionic (SDS)
+
N(CH3)2Br– Cationic (DDAB)
O–
Zwitterionic
N OS (DDAPS)
+ O
[–O(CH2)2]3OH Non-ionic(C12 E3)
Cationic Germini surfactant
+ +
Br–(CH3)2NCH2 CH2N(CH3)2Br–
CnH2n+1 CnH2n+1
O– O
+ O
N P O
O
O O Zwitterionic
(DMPC)
O
Figure 3.13 The molecular structure of some common surfactants, with the hydro-
phobic and hydrophilic moieties indicated. Note that the hydrocarbon chains are shown
in all-trans conformation. This conformation is rarely adopted in micellar structures or in
adsorbed surfactant films at interfaces. The surfactants shown are sodium dodecylsulfate
(SDS), didodecyl dimethylammonium bromide (DDAB), zwitterionic dodecyl-N,N-dimethyl-3-
ammonio-1-proponate-sulfonate (DDAPS), non-ionic C12E3, cationic Germini surfactant, and
the zwitterionic lipid molecule 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC).
The amphiphilic nature of surfactant molecules gives them

some interesting properties in water. The hydrophobic region
is insoluble and the surfactant molecules are forced to accu-
mulate on the surface and expose these non-polar chains away
from the water and toward the air. As Figure 3.14 illustrates,
the number density of the molecules at the surface increases
with bulk concentration. There comes a concentration beyond
which no more surfactant molecules can pack at the interface
Air-water
interface
(a)
Increasing surfactant concentration
(b)
(c)
(d)
Figure 3.14 Surfactant molecules in the bulk aqueous phase are in equilibrium with a
monolayer at the air-water interface. The population of the molecules increases with increas-
ing bulk concentration (a)–(c). At the CMC (d) a saturated monolayer is formed and micelles
are present in the bulk phase.
and a saturated monolayer is formed. The packing density of

surfactant molecules within the monolayer will depend on the
intermolecular interactions between neighboring head groups
and the hydrophobic interactions between the tail groups.
The formation of the monolayer represents a thermodynami-
cally favorable situation: polar head groups are buried in the
aqueous phase and the hydrophobic tails are avoiding con-
tact with water. This scenario explains why surfactants lower
the surface tension of the air-water interface. It is easy to see
that surface tension decreases as the surfactant concentration
increases, but then levels off to a constant value when a satu-
rated monolayer is formed. In Chapter 4 we see how the sur-
face tensions of surfactant solutions are measured.
3.3.1 Micelle and Microemulsion Formation

What happens if we continue to increase the surfactant con-
centration in an aqueous solution beyond that required to
form a monolayer? The molecules are unable to adsorb to the
air-water interface. Instead they self-assemble in the bulk
phase into nanostructures or aggregates known as micelles.
The shape and size of these micelles depends on the structure
of the surfactant. For example, SDS forms spherical micelles;
each micelle is 4 nm in diameter and is composed of about
60 molecules. The number of molecules making a micelle is
known as the aggregation number.
The concentration at which a full monolayer is present at
the interface and micelles begin to form in the bulk solution is
known as the critical micelle concentration, or CMC. The pro-
cess of micellization is in some ways similar to precipitation,
but the precipitate itself has a very narrow size distribution
and is stable and soluble in water. This property is because
the micelle structure is such that the hydrophobic chains are
aggregated in the core while the polar head groups form the
exterior part of the structure [Figure 3.14(d)]. The delicate bal-
ance between inter-head group interactions and the geometry
of the surfactant molecule result in a micelle with a particular
shape, size, and aggregation number.
The CMC of a particular surfactant solution is very sensi-
tive to impurities and other physical conditions. For example,
increasing the temperature of an aqueous surfactant solution
increases the CMC. Thermal agitation makes it more difficult
for the molecules to self-assemble into micelles, so a higher
concentration is required to reach the CMC. The addition of
salt decreases the CMC of ionic surfactant solutions because
the added ions screen the charged head groups of the sur-
factant, thus making it easier to form micelles. As the chain
length of the surfactant increases, the CMC decreases due to
the reduction in solubility of the surfactant. Table 3.2 lists the
CMCs of some common surfactants.
In a non-polar solvent, surfactant molecules can self-
a ssemble to form “reverse micelles.” In this situation, the
molecules are aggregated in such a way that the hydropho-
bic moieties form the exterior of the micelle and the polar
regions of molecules form the core of the micelle. These
micelles are thermodynamically stable in a non-polar solvent
such as hexane.
Micelles in aqueous solution are capable of solubilizing
small amounts of oil added to the aqueous surfactant solution
Table 3.2 The CMC Values of Various Surfactants

Surfactant CMC (mol dm–3)
Dodecylammonium chloride 1.47 × 10–2
Dodecyltrimethylammonium bromide 1.56 × 10–2
Decyltrimethylammonium bromide 6.5 × 10–2
Sodium dodecyl sulfate 8.3 × 10–3
Sodium tetradecyl sulfate 2.1 × 10–3
Sodium decyl sulfate 3.3 × 10–2
Sodium octyl sulfate 1.33 × 10–1
CH3(CH2)9(OCH2CH)6OH 3 × 10–4
CH3(CH2)9(OCH2CH)9OH 1.3 × 10–3
CH3(CH2)11(OCH2CH)6OH 8.7 × 10–5
CH3(CH2)7C6H4(OCH2CH)6OH 2.05 × 10–4
Source: Holmberg, K.; Jönsson, B.; Kronberg, B.; Lindman, B. 2003.
Surfactants and Polymers in Aqueous Solution, 2nd ed., John
Wiley & Sons, Chichester, West Sussex, England. With
permission.
Oil phase
Oil-water
interface
Aqueous phase
Figure 3.15 When an aqueous surfactant phase above its CMC is brought in contact
with a pure oil phase, the micelles are able to solubilize small oil droplets. The aqueous phase
is a microemulsion. A monolayer of the surfactant likely populates the oil-water interface.
The diagram also shows a “reverse” micelle in the oil phase containing a water droplet.
above its CMC. As a result the micelles swell with oil and
increase in size (Figure 3.15). This swollen micellar phase
is thermodynamically stable and is known as a microemul-
sion. In essence, tiny oil droplets are solubilized in water.
Microemulsion phases are generally made by putting an
aqueous surfactant phase in contact with an oil phase. The
two phases don’t mix completely, but a small equilibrium

amount of the oil will enter the aqueous phase and turn it
into a microemulsion. In fact, a complex equilibrium is estab-
lished between the oil phase, a monolayer at the oil-water
interface, and the aqueous micellar solution. It is also pos-
sible for reverse microemulsion aggregates to be present in
the oil phase.
The amphiphilic nature of surfactants is the reason why
these molecules self-assemble into nanostructures, both as
micelles and microemulsions, and as monolayers. The abil-
ity to adsorb at interfaces and lower surface tension is the
main reason why surfactants are used in almost all detergent
and cleaning products. Microemulsions, with their ability to
solubilize oil, have shown promise in tertiary oil recovery.
Furthermore, manipulating the molecular structure while
still preserving the amphiphilic nature allows the possibil-
ity of introducing a desirable functionality into the surfactant
molecules and then exploiting its ability to self-assemble into
novel materials.
3.3.2 The Determination of Surface Excess: The CMC

and the Cross-Sectional Area per Molecule
Surface tension measurements of aqueous amphiphilic solu-
tions allow one to determine the bulk concentration at which
a saturated monolayer is reached. This concentration also cor-
responds to the CMC if a surfactant is used. Furthermore, one
can use surface tension measurements to determine the cross-
sectional area occupied by each molecule on the surface. Let’s
first consider a series of aqueous solutions of the cationic sur-
factant cetyl trimethylammonium bromide [(C16H33)N(CH3)3Br],
or CTAB. CTAB is an effective antiseptic agent against bacte-
ria and fungi and is used in hair conditioning products. After
measuring the concentration of each solution, we can gener-
ate a surface tension versus concentration plot, as shown in
Figure 3.16. The surface tension values drop from pure water
(~72 mN m–1) to ~40 mN m–1 when the bulk aqueous concen-
tration of CTAB is 1 mM. The decrease in surface tension in
this range is due to the increase in packing density as the bulk
concentration is increased. Beyond 1 mM, the surface tension
remains constant at 40 mN m–1, which indicates that the pack-
ing density of CTAB molecules at the air-water interface is not
increasing as the bulk concentration increases. At 1 mM and
beyond, the surface has been saturated with CTAB molecules
Surface Tension
Bulk Surfactant Concentration
CMC
Figure 3.16 The drop in surface tension of air-water interface with increasing sur-
factant concentration. The intersection of the horizontal (constant surface tension) and the
sloped lines corresponds to the CMC of the surfactant.
and a close-packed, highly oriented, monolayer is present on

the surface.
What happens to the excess surfactant molecules in the
bulk phase when the concentration exceeds that required to
form a complete monolayer on the surface (1 mM in the case of
CTAB)? They are unable to remain as individual molecules in
solution and instead self-assemble into micellar nano-struc-
tures. This is shown in Figure 3.16, where a break point in
a plot of the surface tension versus surfactant concentration
marks the CMC. The CMC of an aqueous solution of CTAB is
1 mM.
Surface tension measurements also allow the determina-
tion of a quantity known as surface excess. Consider an aque-
ous solution containing surfactant molecules. Let’s assume
that we can determine the concentration of molecules in a
two-dimensional plane somewhere in the bulk phase below
the air-water interface. We can then compare this bulk phase
concentration to the concentration of molecules on the actual
surface of the solution. Since we are considering a surfactant
solution, we expect the surface concentration of surfactant
molecules to be much higher than the bulk phase concen-

tration. The difference between these two concentrations is
known as the surface excess. The value tells us how much
extra surfactant we have on the surface compared to a two-
dimensional plane in the bulk phase. Surface excess is, to a
good approximation, a measure of the concentration of mol-
ecules on the surface.
For non-ionic amphiphiles, the Gibb’s adsorption equation
relates surface tension (γ) to surface excess (Γ) at a given tem-
perature (T), and is given by
1 dγ
Γ=− (3.19)
RT d ln C
where C is the surfactant concentration in solution below

the CMC and R is the molar gas constant in J K–1 mol–1. The
surface excess is obtained from the slope of a plot of the
surface tension versus the logarithm of the concentration.
For ionic amphiphiles, a slightly different equation is used
(Equation 3.20) because the presence of the counterion has to
be taken into account and the surface as a whole must be con-
sidered electrically neutral.
1 dγ
Γ=− (3.20)
2RT d ln C
Figure 3.17(b) shows a plot of the surface tension values

versus the logarithm of the concentration for aqueous CTAB
solutions. Below the CMC, the change in the slope reflects the
different surface excess values at the various concentrations.
At (i), the slope (or the value of dγ/dlnC) is relatively small,
indicating a small surface excess. At the higher concentration
indicated by (ii), the derivative dγ/dlnC becomes larger, and
according to Equation 3.19, the surface excess becomes larger.
At higher concentrations like point (iii), the surface excess
increases even more. The largest slope occurs at the CMC and
the maximum possible surface excess is achieved, thus indi-
cating the presence of a saturated monolayer at the aqueous-air
interface. Equations 3.19 and 3.20 cannot be applied beyond
the CMC.
65 65
60 60 (i)
55 55
(ii)
γ/mN m–1
γ/mN m–1
50 50
Saturated (iii) Saturated
45 monolayer 45 monolayer
40 40
35 35
CMC CMC
30 30
0 1 2 3 4 5 6 7 8 9 10 –2.5 –1.5 –0.5 0.5 1.5 2.5
[CTAB]/mM Log(C)
(a) (b)
Figure 3.17 The surface tension of aqueous CTAB solutions versus (a) concentration
and (b) logarithm of the concentration. The surface tension values were obtained at 20°C
using the Wilhelmy plate method. Beyond the concentration corresponding to the CMC, the
surface tension values become constant, indicating a saturated monolayer at the aqueous-
air interface.
Example 3.3 Calculating Surface Excess
Consider a concentrated aqueous solution of n-decanol. A plot of

the surface tension versus the logarithm of the concentration gives
a slope of –3.23 mN/m. Use dimensional analyses to determine the
units of surface excess and then calculate the surface excess. Using
this value, determine the cross-sectional area per molecule of n-de-
canol at the air-water interface.
Solution Since the logarithm term in Equation 3.19 is unitless, the

derivative dγ/lnC has the same units as surface tension. The units of
R are J K–1 mol–1 and we know that 1 N = 1 Jm–1. Thus, according to
Equation 3.19, the units of Γ are
1 1
Nm −1 = Jm−1m−1 = mol m−2
(JK−1mol−1 ) ⋅ K (JK−1mol−1 ) ⋅ K
The surface excess of the n-decanol solution is
1 dγ 1
Γ=− =− ( −3.23 × 10−3Nm−1 )
RT d lnC ( 8. 314 JK )(
−1mol−1 298K
)
= 1.30 × 10−6 mol m−2

Assuming the solution of n-decanol is close to saturation, we

have close to a complete monolayer of the alcohol at the aqueous-
air interface. The cross-sectional area per adsorbed molecule is
inversely proportional to the adsorbed amount. If the surface excess
is expressed in mol/m2, then the area per molecule, σ, is
1
σ (m2 /molecule ) =
NAΓ
σ is usually expressed in nm2/molecule. Since NA = 6.023 × 1023

mol–1, we have
σ (nm2 /molecule ) =
(109 nmm )2
6.023 × 1023 molecule
mol
s
⋅Γ
nm2 ⋅mol
1.6603 × 10−6 m2 ⋅molecules
=
Γ
nm2 ⋅mol
1.6603 × 10−6 m2 ⋅molecules
=
1.30 × 10 −6 mol
m2
= 1.28nm2 /molecule
Thus, each n-decanol molecule occupies an area of 1.28 nm2 on

the surface of water.

• Evans, D. F. and Wennerström, H. The Colloidal Domain,
2nd ed., 1999, Wiley-VCH, New York. pp. 99–153, 217–
295. This is one of the finest books written on colloidal
systems and includes a thorough discussion of surfac-
tant chemistry, monolayers, and microemulsions. Also
Chapters 3 and 5 are particularly relevant to the inter-
molecular interactions in nanomaterials.
• Holmberg, K., Jönsson, B., Kronberg, B., and Lindman,
B. Surfactants and Polymers in Aqueous Solution, 2nd
ed., 2003, John Wiley & Sons, Chichester, West Sussex,
England. This is certainly one of the most thorough
and comprehensible books on surfactant chemistry. It
is very light on mathematical rigor and provides very
thorough descriptions and examples.
• Rosen, M. J. Surfactants and Interfacial Phenomena,
3rd ed., 2004, John Wiley & Sons, New York. This book
is an excellent source of data on many surfactants,
including industrial synthetic surfactants, “green”
surfactants, and amphiphilic molecules relevant to

living systems.
• Clint, J. H. Surfactant Aggregation, 1992, Blackie &
Son. This book contains some excellent chapters on
micelles and microemulsion. It is easy to follow and is
highly recommended for anyone interested in surfac-
tant science.
• Roach, P., Shirtcliffe, N. J., and Newton, M. I. “Progress
in Superhydrophobic Surface Development,” Soft
Matter, 2008, 4, 224–240. This is an excellent review
article on superhydrophobic surfaces and their rela-
tion to hard nanomaterials. The review contains many
examples and references to methods used to construct
superhydrophobic surfaces.
• Hartland, S., Ed. Surface and Interfacial Tension:
Measurement, Theory, and Applications, 2004,
Surfactant Science Series, Volume 119. Marcel Dekker,
Santa Barbara, CA. This book provides a rigorous treat-
ment of interfacial tension, film stability, and wetting
phenomena. This book is recommended only for the
serious student.
• Tóth, J., Ed. Adsorption: Theory, Modeling, and
Analysis, 2002, Surfactant Science Series, Volume 119.
Marcel Dekker, Santa Barbara, CA. This book contains
many papers describing adsorption isotherms, includ-
ing the adsorption behavior of biological molecules.
The book is mathematically rigorous and is recom-
mended only to those seriously interested in adsorp-
tion phenomena.

1. Calculate the work done when the surface of water
increases by 50 nm2.
2. Surfactant molecules on the surface of water, if suf-
ficiently dilute, can be described as a two-dimensional
gas phase. If there are no intermolecular interactions
between the surfactant molecules, the gas can be con-
sidered ideal. Instead of the familiar PV = nRT, this
ideal will obey a two-dimensional ideal gas equation
Πσ = RT, where Π is the “surface pressure” due to the
surfactant molecules, and σ is the surface area per mol-
ecule. Derive an expression for the reversible isother-
mal work due to expansion of this two-dimensional
gas. What is the work done when the gas expands
and the area per molecule increases from 20 nm2 to

40 nm2? Clue: The work can be determined by solving
the following integral:

∫
w = − ∏ dσ
3. The modified van der Waals equation is a more realis-

tic equation describing a monolayer of lipid molecules
at the air-water interface (see Question 2). This equa-
tion can be described as
KT α
Π= −
σ − β σ2

where the surface pressure (Π ) is a function of the

independent variables temperature (T) and the surface
area per lipid molecule (σ), i.e., Π (T,σ). K, α, and β are
constants. Derive an expression describing the revers-
ible isothermal work due to the expansion of a lipid
monolayer obeying this equation of state.
4. Charcoal is an excellent adsorbate for organic mol-
ecules. The amount of dodecanol adsorbed on this
material from a toluene solution was measured at room
temperature. The following data gives the equilibrium
amount adsorbed on charcoal and the corresponding
equilibrium concentration of dodecanol in the bulk
phase.
Bulk concentration 0.010 0.035 0.061 0.104 0.149

(mol dm–3)
Amount adsorbed 24.0 50.3 70.0 81.2 90.8
(µmol g–1)
(a) By means of a graph, show that the data fit the

Langmuir adsorption model. Calculate the area
occupied by each adsorbed dodecanol molecule
at saturation coverage. Take the adsorption area
of the charcoal to be 100 m2 g–1.
(b) How would the adsorption isotherm change if
the dodecanol was adsorbed from a slightly more
polar solution compared to toluene?
5. This question concerns the application of Young’s equa-

tion, which considers the equilibrium state of a drop
of liquid on a solid surface in terms of the various sur-
face tensions. Consider a drop of oil (an n-alkane) on
the surface of solid polytetrafluotoethylene (PTFE). As
expected, the oil spreads to some degree and forms a
drop on the surface. The following oils were placed on
the solid surface and the corresponding contact angles
measured. The surface tension of each oil is also given.
Oil (n-alkane), n cosθ γ (mN/m)

6 0.95 18
8 0.87 22
12 0.78 25
16 0.72 27
Estimate the critical surface tension (γc) correspond-

ing to complete wetting. Show that its value is simply
γSV. Do you think it is true that all liquids with a surface
tension less than γc will spread on PTFE? If so, explain.
6. The Gibb’s adsorption equation can be used to obtain
values of surface excess at various concentrations. The
following surface tensions were measured for aqueous
solutions of n-pentanol at 20°C (see graph below). A
polynomial function fit to the data is also shown along
with the equation representing the best fit.
75
y = –1E+07x5 + 4E+06x4 – 500001x3 + 28330x2 – 1031.3x + 72.587
70 2
R =1
Surface Tension (mN/m)
65
60
55
50
45
40
0 0.02 0.04 0.06 0.08 0.1
Concentration (mol/L)
(a) Calculate the surface excess concentrations and

the average area occupied by each adsorbed mol-
ecule for bulk concentrations of 0.01, 0.02, and
0.04 mol dm–3.
(b) Plot a surface pressure versus area per molecule
(Π-σ) curve for the adsorbed n-pentanol monolayer
and compare it with the corresponding curve for
an ideal gaseous film. You should say something
about how the molecules are interacting.
Clue: The compressibility of the film is
Πσ
Z=
kT

This should be equal to 1 at all surface pressures
for an ideal gas.
7. Consider the following plot of surface tension versus
SDS concentration. The squares represent the data
from a purified sample of SDS (>99.99%, recrystallized
thrice from ethyl acetate), and the circles show the
data from a batch used as received from a popular ven-
dor (>99.0%). The latter sample contains trace amounts
of the corresponding alcohol (dodecanol).
80
70
Surface Tension (mN/m)
60
50
40
30
20
0 2 4 6 8 10 12 14
SDS0 (mM)
(a) Estimate the approximate CMC of pure SDS.

Explain how this CMC is affected by (i) a decrease
in temperature, (ii) an increase in ionic strength,
(iii) an increase in pH, (iv) the addition of a

polycation.
(b) With the aid of diagrams explain the observation
of the minimum in surface tension for the con-
taminated SDS sample. Explain clearly why the
data differs from the purified sample.
8. In a recent paper in the Journal of Chemical Education,
Bresler and Hagen reported an interesting laboratory
project to investigate surfactant adsorption (Bresler,
M. R.; Hagen, J. P. J. Chem. Edu. 2008, 85, 269-271). This
problem will go through some of the steps these authors
took to obtain an expression for the surface tension of a
surfactant solution.
(a) Rearrange the Gibb’s adsorption equation
(Equation 3.19) to the form
RT Γ
dγ = − dC
C
(b) The surface excess Γ can be assumed to follow

the Langmuir adsorption isotherm, especially
when considering a non-ionic surfactant. Express
Equation 3.13 in the form
K ⎡⎣ A ⎤⎦
θ=
1 + K ⎡⎣ A ⎤⎦

and argue that
K adC
σΓ =
1 + K adC

where σ is the cross-sectional area of the surfac-

tant molecules at the surface, and Kad is the equi-
librium constant for adsorption. What assumption
has been made in saying that surfactant adsorp-
tion follows the Langmuir adsorption isotherm?
(c) Solve
K adC
σΓ =
1 + K adC

for the surface excess, and then substitute the

result into
RT Γ
dγ = − dC
C
Integrate the resulting expression to obtain the

Szyszkowski equation,
RT
γ = γ0 −
σ
(
ln 1 + K adC )

where γ0 is the surface tension of pure water.

9. The Szyszkowski equation in Question 8 can be used
to obtain the standard Gibb’s energy of adsorption
o
( �Gad ). By plotting surface tension versus bulk con-
centration below the CMC and then curve fitting the
data, one can obtain Kad. This value is related to the
standard Gibb’s energy of adsorption by the equation
o
ΔGad = − RT ln K ad . The following data was collected for
a non-ionic surfactant.
C / µmolL-1 0.0 2.5 5.2 8.0 13.0 17.5 21.0 31.5

γ / mNm-1 72.8 53.6 49.7 45.2 42.6 40.8 40.7 40.8
(a) Plot γ versus lnC and determine the CMC of the

surfactant.
(b) Plot γ versus C and determine the parameters γ
and Kad using nonlinear curve fitting.
(c) Determine the standard Gibb’s energy of adsorption.
(d) The standard Gibb’s energy of micellization is
o o
given by ΔGmic = RT ln CMC . Determine �Gmic .
10. Berberan-Santos commented on Bresler and Hagen’s
paper (Questions 8 and 9) in a letter published in the
same journal (J. Chem. Ed. 2009, 86, 433). Berberan-
Santos pointed out that the Szyszkowski equation can
be reduced to the following equation if the surfactant
concentration is sufficiently low:
⎛ o ⎞
ΔGad RT
γ = ⎜γ0 + ⎟ − ln C
⎝ σ ⎠ σ

Derive this equation by considering a dilute surfactant

solution. Use the data given in Question 9 to plot γ ver-
sus lnC. From the linear plot obtain the values of σ,
o
�Gad , and Kad. Berberan-Santos stated that nonlinear
fitting is preferable in general (Question 9), but may not
be mandatory. Discuss this in the context of the data
presented in Question 9.
11. This question concerns the stability of spherical
“nano-bubbles.” Consider a bubble, like that formed
when a soapy film on a ring is blown. The bubble
has an internal pressure P1, and a radius r. Po is the
external pressure (e.g., the pressure of the surround-
ing air). At equilibrium, the bubble is stable and dG/
dr = 0, where dr is the infinitesimal decrease in bubble
radius. If P1 > Po, work must be done to ensure dr = 0.
The change in Gibb’s energy due to the change in sur-
face area is approximately equal to
dG = –8πr dr γ + ∆P 4πr2 dr
where γ is the surface tension of the bubble and ∆P

= P1 – Po. The first term in the above equation is a
measure of change in Gibb’s energy due to surface ten-
sion, and the second term describes the mechanical
work done against the pressure difference across the
bubble surface.
Use the equilibrium condition dG/dr = 0 to derive the
Laplace equation:
2γ
ΔP =
r
The Laplace equation essentially gives the pressure

inside a stable bubble.
(a) The Young-Laplace equation
⎛ 1 1⎞
ΔP = γ ⎜ + ⎟
⎝ R1 R2 ⎠

can be applied to nonspherical bubbles. In this

expression, R1 and R2 represent the two principal
radii of curvature. Show that the Young-Laplace
equation reduces to the Laplace equation for a

spherical bubble.
(b) Calculate the Laplace pressure (∆P) in units of
bars for bubbles of radius 1 nm, 2 nm, 10 nm, and
1000 nm. Comment on how ∆P changes with the
size of these “nano-bubbles.”
(c) Using your understanding of Laplace pressure,
explain why small boiling chips are often added
to hot reaction mixtures.
Four
Characterization at
the Nanoscale
CHAPTER OVERVIEW
This chapter provides a broad survey of some of the common
techniques used to characterize nanomaterials. Novel tech-
niques and methodologies to probe nanoscale dimensions are
being developed continually, and many of these are based on
variations of common tools and practices used in surface science
and materials research laboratories. The aim of this chapter is
to provide a modest level of background in a sufficient range of
techniques to allow one to make rational choices in determin-
ing which method or combination of methods are best suited to
characterize a particular feature of a nanomaterial. The meth-
ods introduced in this chapter range from traditional surface
science tools, spectroscopic methods, and gravimetric tech-
niques to more specialized characterization approaches such as
nonlinear optical methods and interferometic techniques. The
chapter focuses on the principles behind the techniques and
the interpretation of data. Specific examples of the application
of these methods to nanomaterials are found in Chapter 5.
4.1 SURFACE TENSIOMETRY: THE

SURFACE TENSIOMETER
There are a number of methods used to determine the sur-
face tension of liquids. One important method is to study the
movement of the fluid up through a capillary tube, and this
is probably the most accurate method of determining surface
tension values. The basic setup is shown in Figure 4.1. From
this setup, the surface tension of a particular fluid of density ρ
can be calculated directly using the height h to which the liq-
uid rises through the narrow capillary of radius r as described
by Equation 4.1.
97
2r
Figure 4.1 A fluid moving through a narrow capillary tube. The distance h the fluid
travels depends on the surface tension (γ) and the contact angle (ϕ). The distance h also
depends on the diameter 2r of the capillary.
rhΔρg
γ=
2cos φ (4.1)
The term ∆ρ represents the difference between the density of

the liquid (e.g., bulk water) and the density of the vapor (usu-
ally air). The angle ϕ is the contact angle the liquid makes
against the capillary surface, and g is the acceleration due to
gravity. For a narrow capillary the contact angle approaches
zero, so the cosϕ term in Equation 4.1 is generally set equal
to 1. The surface tension then depends on the fluid density,
the radius of the capillary, and the height the fluid travels
up the tube. Other methods for determining surface ten-
sion include measuring the volume of a drop detached from
a narrow tube, analyzing an image of a pendant drop, and
observing a jet of liquid emerging from a nozzle of ellipti-
cal cross-section. A thorough treatment of these various
methods can be found in volume 119 of the surfactant sci-
ence series, Surface and Interfacial Tension: Measurement,
Theory, and Applications.
The Wilhelmy plate method is another common technique
used to determine surface tension values and is the one to
which we confine our attention for the remainder of this sec-
tion. The Wilhelmy plate method involves measuring the
Characterization at the Nanoscale 99
Force balance
h
φ w
Figure 4.2 The Wilhelmy plate method. A platinum or paper plate of length I and width
w is immersed into a fluid to a depth h. The forces acting on the plate are measured using
a force balance. ϕ indicates the contact angle the fluid makes against the plate. If a paper
plate is used, this angle can be reduced to zero.
forces acting on a “plate,” usually a very thin piece of platinum

or paper, at the liquid-air interface (Figure 4.2).
If a plate with dimensions l = length, w = width, t = thick-
ness, and of density ρP is immersed to a depth h into a fluid of
density ρL, then the forces acting on the plate are its weight,
the upthrust on the submerged part of the plate due to buoy-
ancy, and the surface tension of the liquid on the plate. The
total force on the plate can be written as
F = (ρPlwt)g – (ρLhwt)g + 2(w+t)γcosϕ (4.2)
where g is the acceleration due to gravity. In this expression,

(ρPlwt)g is the weight of the plate, (ρLhwt)g is the buoyant
upthrust, and the surface tension contribution to the force is
2(w+t)γcosϕ, where ϕ is the contact angle that the liquid makes
on the plate as indicated in Figure 4.2.
Example 4.1 Measuring the Density from

the Force Acting on a Plate
A clean plate of width w and thickness t is submerged into an aque-

ous surfactant solution. The solution has a surface tension value of
γ mN/m. The force acting on the plate is measured as a function
of immersion depth h. At h = 2 cm, the force is F1, and at a depth of
h = 4 cm, the force is F2. Show how this information can be used to
estimate the density of the surfactant solution.
Solution We can substitute the values of h and F into Equation 4.2

for the two cases and write,
(a) (ρPlwt)g – 0.02(ρLwt)g + 2(w+t) γcosϕ = F1

(b) (ρPlwt)g – 0.04(ρLwt)g + 2(w+t) γcosϕ = F2
Subtracting these two equations gives 0.02(ρLwt)g = F1 – F2.

Thus, by measuring the difference in force for a plate of a given
width and thickness, ρL can be obtained.
The total force on the plate can be measured accurately by

connecting the plate to a sensitive force balance. However,
in order to accurately calculate the surface tension from the
total force using Equation 4.2, we would need to know (also
very accurately) the dimensions of the plate as well as the
densities of both the plate and the liquid. Fortunately, the
expression of the measured total force in Equation 4.2 can
be simplified somewhat through careful manipulation of the
procedure used to measure the total force on the plate, as
follows.
First, before making any measurements and prior to sub-
merging the plate into the liquid, the force balance is reduced
to zero. This reduction eliminates the weight term from
Equation 4.2. The plate is then submerged into the fluid and
slowly raised until its lower edge is level with the liquid sur-
face, thus eliminating the upthrust term because h = 0. In this
position, Equation 4.2 is reduced to
F = 2(w+t)γcosϕ (4.3)
Using Equation 4.3, therefore, we can determine the surface

tension provided that we know the contact angle the liquid
makes to the plate. This is often not necessary, however,
because the expression can be further simplified by realizing
that as we slowly pull the plate away from the liquid surface,
ϕ decreases and becomes zero (or cosϕ = 1) just before detach-
ment. Thus, if we measure the force just before detachment of
the plate from the liquid, we can determine the surface ten-
sion from the dimensions of the plate:
F
γ= (4.4)
2(w + t )
In order to obtain an accurate surface tension value, it is

important to eliminate sources of contamination and work
at a constant temperature (±0.5°C), since surface tension is

temperature dependent. Furthermore, in order for Equation
4.4 to be used, a zero contact angle is required to exist between
the plate and the liquid. This is usually achieved by using pre-
cisely cut paper plates, as the liquid can completely soak the
paper, thus ensuring a zero contact angle. Platinum plates,
even when cleaned thoroughly, can have their surfaces con-
taminated by the liquid during the first immersion, and their
contact angle can vary during subsequent immersions. Flame
cleaning the plate is usually recommended when working
with platinum plates.
Surface tension experiments have been pivotal to our
understanding of the properties of monolayers. The technique
provides direct information on packing density and the area
occupied by a molecule in a monolayer. Although the area per
molecule gives a qualitative indication of the thickness of the
film, surface tension cannot be used to give the absolute film
thickness or the absolute mass of the monolayer. Fortunately,
film thicknesses as low as a fraction of a nanometer can be
measured using optical methods such as ellipsometry (Section
4.3) or dual beam polarization interferometry (Section 4.5).
Gravimetric techniques also exist that can measure the mass
of a monolayer on solid surfaces. The quartz crystal microbal-
ance is a popular method used to determine the mass of a
monolayer.
4.2 QUARTZ CRYSTAL MICROBALANCE

All gravimetric analyses rely on the determination of the mass
of a material, or in our context, the mass of a nanofilm that
has been deposited on a solid support. Since mass can usu-
ally be measured accurately, gravimetric analysis is generally
a precise analytical method. However, when working with
nanofilms, the mass in question may well be as low as a few
nanograms per square centimeter, and traditional methods
for weighing such samples are not possible. The next few sec-
tions discuss methods that allow the direct measurement of
the mass and the thickness of a nanofilm (or, more often, the
measurement of mass-related and thickness-related para
meters). These methods are the quartz crystal microbalance
with dissipation monitoring (QCM-D), ellipsometry, dual
beam polarization interferometry (DPI), and surface plasmon
resonance (SPR). These techniques are routinely used in many
research laboratories and can measure the changes in mass
and thickness during nanofilm growth.
First, let’s focus on the quartz crystal microbalance (QCM).

QCM is a powerful technique that can measure the mass of
material as small as a few nanograms adsorbed to a surface.
What makes this technique particularly appealing is that mod-
ern QCM instruments allow one to follow the mass deposition
process as a function of time. In other words, the formation of
a thin nanofilm, such as a model membrane, can be observed
in real time. The quartz crystal microbalance is based on the
piezoelectric characteristics of quartz, so in order to properly
understand how QCM operates, let us first discuss the piezo-
electric effect.
4.2.1 The Piezoelectric Effect

Since early times it has been known that an electric field could
be induced in certain types of crystals if they underwent a
change in temperature (i.e., were heated or cooled). This
phenomenon was named pyroelectricity. A Greek scientist,
Theophrastus, first noticed the pyroelectric effect in 340 b.c.
in experiments with tourmaline, a crystal silicate contain-
ing various metallic elements. In 1707, Johann Georg Schmidt
showed that by heating or cooling tourmaline (as well as sev-
eral other classes of crystals of proper geometries), a positive
or negative electric field could be induced in the crystal.
In 1880, scientist-brothers Jacques and Pierre Curie dis-
covered that they could also induce an electric field in cer-
tain classes of crystals, not by heating or cooling them, but
by applying a mechanical stress. In addition, they realized
that by changing the direction of the mechanical stress (i.e., a
compression or expansion) they could control the sign of the
electric field created. Their prior knowledge of pyroelectricity
led them to this discovery, and they dubbed the phenomenon
the piezoelectric effect, from the Greek piezein, meaning “to
push.” Today this is known as the direct piezoelectric effect—
the ability of a crystal to produce an electric field in response
to mechanical stress.
In subsequent years, the Curie brothers and other scientists
began mapping out the crystalline structure requirements for
piezoelectricity. They found that crystals possessing piezo-
electric qualities have neither an absolute center nor a plane or
axis of symmetry perpendicular to the axis of electric activity
(the plane on which the electric field arises). There are cur-
rently 21 known classes of crystals that exhibit the piezoelec-
tric effect, including quartz, cane sugar, and bone.
Another important advance in the understanding of piezo-
electricity came with the discovery that a piezoelectric crystal
would be physically deformed (e.g., expand or contract) in

response to an applied voltage. This effect was effectively
the opposite of the direct piezoelectric effect and was subse-
quently named the converse piezoelectric effect. It was also
discovered that the converse effect could be used to induce
vibrations along a crystal. By applying a precise alternat-
ing current, the deformations of the crystal could be rapidly
changed between expansion and contraction, causing the
crystal to oscillate.
The piezoelectric effect is the foundation of a host of tech-
nologies. One of the original applications was a piezoelectric
device used during the development of sonar in which piezo-
electric quartz crystals were used as transducers to detect
echoes returning from underwater objects. Piezoelectric
devices are also commonly used in microphones—the air
pressure fluctuations caused by a sound distort the piezoelectric
device in the microphone, producing an electrical signal.
The converse piezoelectric effect is also used for the extraor-
dinarily fine and accurate adjustment of devices such as the
needle in a scanning tunneling microscope.
4.2.2 QCM Principles

A quartz crystal microbalance operates on the principles of
piezoelectricity. As we’ve discussed, a piezoelectric quartz
crystal will oscillate at a specific resonant frequency in
response to an AC voltage. (In fact, the resulting resonant fre-
quency is so reliable that quartz crystals are used in many
precision time-keeping devices). This resonant frequency is
dependent upon the size and mass of the quartz crystal. If the
mass of the quartz crystal is altered (such as when a mole-
cule adsorbs to its surface), the resonant frequency at which
the crystal is oscillating shifts slightly [Figure 4.3(a)]. If the
adsorbed substance is not too thick or floppy, the change
in mass on the crystal’s surface is proportional to the shift
in resonant frequency, as given by the Sauerbrey equation:
C ⋅ Δf
Δm = − (4.5)
n
where m is mass, f is frequency, n is the overtone number of
the crystal (n = 1, 3, 5, 7, …), and C is a constant that depends
upon the specific quartz crystal.
A quartz crystal microbalance detects minute changes
in mass by applying an AC potential across a quartz crys-
tal to induce its resonance frequency and then monitoring
104
Crystal oscillating
with frequency
induced by AC
Pure crystal With material After AC potential
potential
oscillation adsorbed on the surface is removed
Time Time
Amplitude
Amplitude
(a) (b)
Figure 4.3 (a) The resonant frequency of a piezoelectric quartz crystal in a QCM system will oscillate more slowly when a material is adsorbed to its surface. Thus
a decrease in the resonant frequency is observed during an increase in mass adsorbed to the surface. (b) Dissipation is a measure of how quickly the QCM crystal stops
oscillating after the AC potential is removed. As such, it describes the relative thickness and rigidity of thin films adsorbed to the surface of the quartz crystal. The decaying
oscillations of the crystal after the circuit is broken obey an exponentially decaying sinusoidal curve. The numerical data of the decay can be fitted to Equation 4.7, which
allows for the calculation of a time constant τ, which in turn can be used to calculate the dissipation of the crystal using Equation 4.8.
Understanding Nanomaterials
the changes in that resonant frequency that result from the

adsorption of molecules to the crystal’s surface. Since those
resonant frequency shifts are proportional to mass under
the conditions described above, and because even very small
shifts in the resonant frequency can be detected with modern
electrical equipment, QCM operates as a sensitive balance, or
gravimetric device (hence the name microbalance). By moni-
toring the shifts in frequency as a function of time, QCM is
able to monitor the formation of a thin nanofilm in real time. It
should be noted, however, that the Sauerbrey relation shown
in Equation 4.5 is often not exact when applied to surfaces at
the solid-liquid interface because it was developed for oscil-
lations in air and only applies to rigid masses attached to the
crystal. It generally underestimates the mass adsorbed to the
crystal surface under a liquid phase. Therefore, resonant fre-
quency shifts detected using QCM should be considered a
mass-related parameter, not an absolute measurement of mass
adsorbed to a surface.
Example 4.2 What Are the Detection Limits of QCM?
A typical QCM-D instrument uses 5-MHz quartz crystals (meaning

the crystal’s fundamental resonant frequency is ~5 MHz). These
crystals have a Sauerbrey constant of C = 17.7 ng Hz–1 cm –2. If the
QCM-D instrument is capable of detecting changes in resonant fre-
quency of ~0.1 Hz, what is the detection limit of a typical QCM-D
instrument at its fundamental resonant frequency (n = 1)?
Solution We utilize the Sauerbrey relation to find
C ⋅ Δf (17.7ng ⋅ Hz −1 ⋅ cm −2 )(~ ± 0.1Hz )

Δm = − =− =~ ± 2ng ⋅ cm −2
n 1
Therefore, the detection limit of a typical QCM-D instrument is

~2ng cm –2. Indeed, the QCM is an incredibly sensitive balance!
4.2.3 QCM and Dissipation (D)

The quartz crystal microbalance with dissipation monitoring
(QCM-D), as compared with traditional QCM, offers additional
information called dissipation about the materials adsorbed to
the QCM crystal surface. Dissipation is a measure of the abil-
ity of the adsorbed material to release or dissipate the energy
of the oscillating QCM-D crystal. As such, it provides insight
into characteristics of the adsorbed film such as its density,
thickness, and viscosity. Collectively, these properties are
sometimes referred to as viscoelastic properties. Dissipation is

classically defined as
Edissipated
D= (4.6)
2π E stored
where Edissipated is the energy lost during one oscillation cycle

and Estored is the energy stored in the crystal.
In practice, dissipation is determined by monitoring the
time decay of the quartz crystal’s oscillation when the AC
potential is removed. The decay in the crystal’s oscillations is
an exponentially decaying sinusoidal of the form
A(t ) = Aoe − t /τ sin(2πft + ϕ ) (4.7)
where t is time, τ is the decay constant, f is frequency, and φ is

the phase angle. By numerically fitting the observed decay of
the crystal’s oscillations to Equation 4.7, the time constant τ can
be obtained, from which dissipation D can be calculated as
1
D=
π⋅ f ⋅τ (4.8)
In other words, dissipation can be thought of as a measure

of how quickly the crystal stops oscillating when the elec-
trical circuit is broken [Figure 4.3(b)]. If the adsorbed thin
film on the crystal’s surface is thick and “floppy” (or not very
rigid), then it is decoupled from the crystal’s oscillations and
efficiently dissipates the energy of the crystal. Consequently,
the crystal stops oscillating quickly and a high value of dis-
sipation is reported. Conversely, if the film is thin and rigid, it
oscillates together with the crystal and does not dissipate the
crystal’s energy effectively. Hence, it takes a longer amount
of time for the crystal to stop oscillating and a low dissipa-
tion value is reported. In this way, dissipation is a measure
of the “floppiness” of the adsorbed film (or better, its lack of
rigidity).
By using a suitable mathematical model, the dissipa-
tion and frequency shifts resulting from the adsorption of a
nanofilm to a QCM-D crystal surface can be used to calculate
the viscoelastic properties of the film. Viscoelasticity is the
property of materials that exhibit both viscous and elastic
characteristics when undergoing some kind of deformation.

Viscous materials resist flow when a stress is applied. Elastic
materials when stretched very rapidly return to their original
state once the stress is removed. Viscoelastic materials have
elements of both of these properties. While the amount of valu-
able information obtained using such a mathematical model is
impressive, caution should be observed in its use as the model
inevitably requires the input of one known parameter (such
as thickness, density, etc.) in order to calculate the other vis-
coelastic properties. If this required parameter is unknown or
not known precisely, then the resulting calculated viscoelastic
properties are unreliable. For these reasons, it is often helpful
to perform a QCM-D experiment in conjunction with measure-
ments taken from ellipsometry, SPR, or DPI.
4.2.4 Modern QCM-D Setup

Traditional QCM was developed in the 1960s as a method to
detect the adsorption of gas molecules to surfaces. With more
recent advances, the capabilities of QCM have been extended
to detect surface adsorption at the solid-liquid interface, and
QCM-D has become a valuable tool in the characterization of
thin films under solution.
A typical QCM-D setup involves the quartz crystal being
mounted in a flow cell with electrodes mounted on either side
of the crystal, as shown in Figure 4.4. The QCM-D crystal is
generally prepared with an active sensor surface such as gold,
hydroxyapatite, SiO2, etc. The resonant frequency of the QCM
QCM liquid cell
Fluid in Fluid out
Nano-assembly
QCM Coating on crystal

O-ring seal crystal
Applied voltage
Figure 4.4 Typical architecture of a QCM-D instrument. A QCM crystal is exposed to

solution inside a flow cell. Changes to the crystal surface (such as film adsorption) are
monitored via changes in the crystal’s oscillations that have been induced by the applied
voltage.
crystal is monitored as a function of time during the exposure

of the crystal to a given solution under continuous flow con-
ditions of ~0.100–0.300 mL/min. If the solution contains
materials that have some sort of affinity for the QCM crystal
surface, they adsorb preferentially, increasing the mass of the
quartz crystal and consequently producing a negative shift
in the measured frequency. If the adsorbed material is suffi-
ciently well behaved (i.e., is relatively thin and rigid), then the
Sauerbrey relation can be applied, allowing us to calculate the
mass of the adsorbed thin film.
QCM-D has a wide variety of applications. It is widely
used in a biophysical context to detect interactions between
biomacromolecules, such as between an enzyme and its sub-
strate. It is also commonly used to examine the characteris-
tics of polymer thin film formation as well as film behaviors
under different conditions. Another application is determi-
nation of the effectiveness of different detergents to clean a
given surface.
4.3 ELLIPSOMETRY
There are many approaches to determining the thickness of a
nanofilm. Measuring the absorbance of light through a nano-
film is one way to determine its thickness. However, as will
be shown, determining thickness from absorbance measure-
ments requires knowledge of both molar absorptivity and the
concentration of the material comprising the film. It may be
challenging to obtain the concentration of the molecules in
such a film.
Dragging a stylus over the surface of the substrate is another
approach to determining film thickness. This method is the
basis of the technique known as profilometry. A profilometer
can be used to measure a film’s thickness and roughness. In
this method, a diamond stylus is moved vertically in contact
with the surface of the nanofilm, then moved laterally across
the surface for a specified distance and specified contact force.
A profilometer can measure small surface variations in vertical
stylus displacement as a function of position. A typical profi-
lometer can measure small vertical features ranging in height
from 10 nm to 50 µm. In fact, most of the world’s surface finish
standards are written for contact profilometers. Unfortunately,
the technique is “invasive” in that the stylus makes contact
with the surface and may well damage the film. There are non-
contact profilometers that use light as a way of measuring the
height of surface features. However, non-contact profilometry
cannot measure the thickness of films on the order of 10 –10 m

(angstroms).
A technique known as ellipsometry has been used exten-
sively to measure the film thickness as low as a few angstroms.
Ellipsometry is a non-contact and nondestructive technique
used for measuring the thickness and refractive index of a thin
film on a solid surface. Thin films ranging in thickness from
a few angstroms to ~1 µm can be measured accurately and
quickly using this method. Ellipsometry analyzes the state
of polarization of light reflecting from a surface and uses the
laws of electromagnetism (specifically, Maxwell’s equations)
to resolve the thickness and refractive index of the nanofilm.
4.3.1 Basic Principles of Electromagnetic

Theory and Polarized Light
Let’s begin by describing the classical representation of light as
an electromagnetic wave. Ellipsometry measures the change
in the polarization of a light beam after its reflection from the
solid surface of the sample being characterized. Although
detailed coverage of the theory of ellipsometry is beyond
the scope of this book, we include a brief overview of elec-
tromagnetic theory in order to better understand the basis of
ellipsometry. The material also serves as background to other
methods covered in this chapter such as DPI, SPR, and spec-
troscopic methods.
Light can be viewed as an oscillating electromagnetic field
propagating through space. The oscillating field of light has
two components: an electric field and a magnetic field that are
mutually perpendicular to each other and also perpendicular
to the direction of light propagation. Only the electric field
component is considered here since the magnetic component
does not interact appreciably with molecules. Thus, we con-
sider light as an oscillating electric field whose amplitude and
orientation can be represented by a line that we call the elec-
tric field vector. The orientation of the electric field vector at a
given moment in time is defined as the “polarization axis” of
the light. This model of light is shown in Figure 4.5, where the
length and direction of the solid arrows indicate the strength
of the field and its orientation, respectively.
We now turn to the polarization of light. Usually, light emit-
ted by most sources consists of photons whose electric fields are
oriented in all directions that are perpendicular to the direc-
tion of propagation. This is unpolarized light. Conversely, lin-
early polarized light consists of photons whose electric fields
are oriented in only one direction.
x
x y
y
z
Amplitude
z x
Position
(a) (b)
Figure 4.5 Classical representation of light as an electric field. The solid arrows’ vec-
tors represent the field’s orientation and intensity. The magnitude of the vectors oscillates
in time, and their orientation defines the polarization axis. (a) is a wave representation
of the light and (b) is a linear representation showing to which axis the electric field is
confined.
In the simplest case of linearly polarized light, we can use a

line to show the electric field oscillating along a single axis as
shown in Figure 4.5(b). In fact, linearly polarized light, propa-
gating along the z-axis can have its polarization axis oriented
in any direction in the xy-plane. In other words, any source
that produces linearly polarized light in the xy-plane can be
thought of as being the linear combination of two vector com-
ponents oriented along the x- and y-axes. The polarization axis
of linearly polarized light is determined by the relative mag-
nitudes of the two components. For linearly polarized light,
the component light sources must have identical frequency
and must also be in-phase with one another. By “in-phase”
we mean that the minima and maxima of the electric field
oscillations for each component must line up. A representa-
tion of linearly polarized light formed by a linear combination
of two mutually perpendicular vector components is shown
in Figure 4.6.
While we are on the topic of plane-polarized light, it is appro-
priate to introduce some common terminology when consider-
ing the reflection of light from a planar surface (Figure 4.7).
For light incident upon this surface, we can define the plane
of incidence as the plane that contains the surface normal and
the light beam before and after the reflection (called the inci-
dent beam and reflected beam, respectively). If the electric field
is polarized along the plane of incidence, we refer to the light
as being p-polarized. Conversely, if the electric field is polar-
ized along a plane perpendicular to the plane of incidence, we
call the light s-polarized. A combination of s- and p-polarized
electric fields can produce linearly polarized light, but only
when there is a 0° phase difference between the two compo-
nents. For example, if two s- and p-polarized light beams are
(i) (ii) (iii)

y y y
x + x = x
(a)
x y x y y
z + z = z
(b)
Figure 4.6 (a) Linear and (b) wave representations of linearly polarized light of the
same amplitude but various orientations. The polarizations are along (i) the x-axis and
(ii) the y-axis. Vector addition of (i) and (ii) creates a linear polarization 45° from the
x-axis (iii).
Surface normal
Ep Plane of incidence
Es
φi φr
Figure 4.7 Light reflection off a planar surface. The plane of incidence contains the
incident beam, the surface normal, and the reflected beam. If the polarization of the light is
along the plane of incidence, then it is called p-polarized light. If the polarization vector is
perpendicular to the plane of incidence, then the light is called s-polarized light.
completely in-phase and if they each have electric field vectors

of the same magnitude, then upon addition the resulting beam
would be linearly polarized with a polarization axis that
would be 45° with respect to the plane of incidence.
Let’s now consider what would happen if the two electric
field vector components making up a polarized light beam
were not in-phase. First, let’s talk about a special case—when
the two components are 90° out of phase, or when one of the
oscillations is a quarter of a wavelength ahead of the other.
When viewed along the direction of propagation, the resulting

polarization has constant magnitude (the length of the vec-
tor line is the same), but its orientation changes in time such
that the tip of the polarization vector traces a circular path in
time. This kind of light is referred to as circularly polarized
light and is depicted in Figure 4.8(a) and (b). Depending on the
relative phases of the electric field components, the tip of the
polarization vector may trace out either a left- or right-handed
screw and the corresponding light is described as right- or left-
circularly polarized.
If two out-of-phase light waves combine (and not with a
phase difference of 90°n where n is an integer), then ellipti-
cally polarized light is formed. The linear combination of the
two vector components results in a polarization vector that
traces an ellipse in time when viewed along the axis of propa-
gation. Elliptically polarized light is depicted in Figure 4.8(c).
x x
z z
(a)
y y y
90° out of phase
x + x x
(b)
y y y
Phase difference
is not 90° n
x + x x
(c)
Figure 4.8 (a) Two mutually perpendicular electric field components that are 90° out
of phase will combine to generate circularly polarized light. (b) Vector addition of the two
components leading to circularly polarized light. (c) Vector addition of the two components
leading to elliptically polarized light.
As with circularly polarized light, elliptically polarized light

can be left or right handed depending on the relative phases
of the components. The “fatness” of the polarization ellipse
(the ratio of the short axis of the ellipse to the long axis of the
ellipse) is referred to as the ellipticity of the light. It should
also be mentioned that elliptically polarized light can be pro-
duced by the linear combination of orthogonal components of
circularly polarized light as well as linearly polarized light. In
fact, linearly and circularly polarized light can be considered
special cases of elliptically polarized light.
4.3.2 Basic Principles of Ellipsometry

Now that we’ve discussed the electromagnetic nature of light
and polarization, let’s turn to our discussion of ellipsometry.
Ellipsometry involves the reflection of light from a surface or
the interface between two mediums. Consider a model reflec-
tive surface coated with a nanofilm of thickness d (Figure 4.9).
For now we will limit our consideration to the interaction of
an incoming light beam from a light source with the first sur-
face (i.e., the interface between n1 and n2). In this model sys-
tem, when light passes from one medium to a second medium,
several phenomena occur at the interface. Some of the light
is reflected from the surface and some enters the second
medium.
When linearly polarized light reflects off a surface, there
is a phase shift in both the parallel and perpendicular com-
ponents (i.e., the s- and p-polarized components). As before,
parallel and perpendicular are in relation to the plane of inci-
dence of the incident light beam.
φ1
n1
n2 d φ2
φ3
n33
Figure 4.9 The reflection, refraction, and transmission of light through a model mul-
tiple interface system. In our case, the layer of thickness d can be a nanofilm assembled on
a solid support. The ni represent the refractive index of each phase.
There may also be an amplitude difference between incident

and reflected beams in both components. In fact, the phase
shift and amplitude differences are usually not the same for
both components. As a result, the reflected beam is elliptically
polarized. The ellipticity depends on the optical properties of
the substrate (reflecting surface) as well as the optical proper-
ties and thickness of any overlying films. One optical property
of importance is the refractive index (RI), also denoted as n.
The refractive index of a medium is defined as the ratio of the
speed of light in a vacuum (3.0 × 108 m/s) to the speed of light
in the medium.
It is useful to obtain a parameter that describes how much
each of the s- and p-components of the incident light are
reflected or transmitted. This information is given by the
Fresnel reflection coefficient, r, which is the ratio of the ampli-
tude of the reflected wave to the amplitude of the incident
wave. For a single interface, the Fresnel coefficients for s- and
p-polarized light are equal and are given by
n2 cos φ1 − n2 cos φ2
r12p = (4.9)
n2 cos φ1 + n2 cos φ2
n2 cos φ1 − n2 cos φ2
r12s = (4.10)
n2 cos φ1 + n2 cos φ2
In the two equations, the subscripts refer to medium 1 and

medium 2 and the light transmitted through the interface is
ignored.
So far, our discussion of ellipsometric principles has focused
on the interaction of light with two media and a single inter-
face. Now let’s consider the rest of Figure 4.9, which shows a
model of light interacting with multiple interfaces, in this case
a reflective surface coated with a nanofilm of thickness d. The
situation is only a bit more complex than for a single inter-
face. The main source of complexity is that at each interface a
light wave encounters, some of the light is reflected back from
the interface and some is transmitted through it. Referring to
Figure 4.9, we can realize that the result of this partial reflec-
tion/partial transmission is that some of the light that enters
the thin nanofilm is internally reflected, “bouncing” between
medium 1 and the reflective substrate surface. Furthermore, the
intensity of the light inside the thin nanofilm eventually decays
as it releases some light back into medium 1 at each “bounce.”

Each of these transmissions back into medium 1 is successively
smaller and leads to a series of partial waves that combine to
give a resultant total reflected wave. Therefore, our calculation
of the amount of light reflected back into medium 1 from a sys-
tem of multiple interfaces such as in Figure 4.9 must be a mea-
surement of this total reflected wave and account for all of the
small partial waves. This calculation is performed by modify-
ing the Fresnel coefficients into total reflection coefficients R for
multiple interfaces and these coefficients are given by
r12p + r23p e − i 2α
Rp = (4.11)
1 + r12p r23p e − i 2α
r12s + r23s e − i 2α
Rs = (4.12)
1 + r12s r23s e − i 2α
where
⎛d⎞
α = 2π ⎜ ⎟ n2 cos φ2, i = −1
⎝ λ⎠

and λ is the wavelength of the incident light in a vacuum.

The total reflection coefficients for each component (p or s)
are the ratios of the reflected wave amplitude to the incident
wave amplitude. We see that the equations for these total
reflection coefficients each incorporate an exponential decay
term, e –i2α , which accounts for the increasingly smaller par-
tial waves that are produced by the partial reflection/partial
transmission at the nanofilm/medium 1 interface. We also
notice that if d = 0 (or if there is no nanofilm on the surface)
then Equations 4.11 and 4.12 reduce to Equations 4.9 and 4.10,
as expected.
4.3.3 Obtaining the Thickness of Films: Optical

Parameters Del (Δ) and Psi (ψ)
Section 4.3.2 covered the necessary background needed to
understand how ellipsometry can be used to yield thickness
values for thin nanofilms. At this point, we need to define two
parameters (δ1 and δ2) that describe the change in phase as
light is reflected off a surface. Let δ1 be the phase difference
between the p-polarized component and the s-polarized

component of the incident light. Let δ2 be the phase difference
between the p-polarized and the s-polarized component of
the reflected light. We can now define one of the most impor-
tant optical parameters used in ellipsometry, the parameter
Del (Δ), as the phase difference between the p-polarized and
s-polarized components of the incident light upon reflection.
In other words, Δ is the resulting change in the phase differ-
ence between the s and p waves as the light is reflected from
the sample (Equation 4.13).
Δ = δ1 − δ 2 (4.13)
The two components (p and s) making up the incident light

each have a given amplitude (length of the electric field vec-
tor), and those amplitudes may also change upon reflection.
These amplitude changes are given by the total reflection coef-
ficients in Equations 4.11 and 4.12. The second fundamental
ellipsometric optical parameter Psi (ψ) can be defined in terms
of these coefficients and is given by the equation
Rp
tan ψ =
Rs (4.14)
where ψ is defined as the angle whose tangent is the ratio of

the magnitudes of the total reflection coefficients. We define
the additional complex quantity ρ as the ratio of the total
reflection coefficients, or
Rp
ρ= (4.15)
Rs
Using these three equations, we can now present the funda-

mental equation of ellipsometry:
Rp
ρ= = tan ψe iΔ = tan ψ (cos Δ + i sin Δ) (4.16)
Rs
Δ and ψ are the experimental quantities measured by the

ellipsometer and fitted to a computer model to yield refractive
index and film thickness (which are embedded in the equa-
tions for RP and RS).
Most ellipsometry experiments are performed on nanofilms

at the solid-air interface. The underlying substrate is usually
silicon. The determination of refractive index and film thick-
ness for these samples is often done as follows. Suppose we
have a substrate covered by a film as in Figure 4.9. Generally,
Δ and ψ are obtained for the bare silicon substrate and the
instrument determines the refractive index information of
silicon and the native SiO2 layer directly above it. This is
done prior to any film deposition. A table of Δ and ψ values
as a function of film thickness (called a Del/Psi trajectory) is
determined. The Del/Psi trajectory is obtained using a com-
puter program separate from the ellipsometer, although some
instruments have programs to compute Del/Psi trajectories
integrated with the instrumentation. The film is then depos-
ited on the substrate and Δ and ψ are obtained for the substrate
and film. The unknown thickness of the deposited film may
then be obtained by comparison of the Δ and ψ values with the
calculated Del/Psi trajectory for the bare substrate. The optical
constants of the thin film, such as the refractive index, must
be input in order for the program to produce the trajectory.
The user makes an educated guess at these values.
Although the ellipsometer accurately determines Δ and
ψ, these values are meaningless unless the program used to
calculate the Del/Psi trajectory assumes the correct model.
The model used is typically a two-layer model as shown in
Figure 4.9 for a silicon dioxide/silicon wafer substrate.
4.3.4 The Ellipsometer

The layout of a typical ellipsometer is illustrated in Figure 4.10.
In short, a laser beam is appropriately polarized and then
reflected off the substrate at an angle of 70°. The polarization
Light source
Polarizer
Detector
φi φr
Analyzer
Quarter
wave plate
Sample
Figure 4.10 Layout of a typical ellipsometer. The angle of incidence (ϕi) is typically 70°.
The light source is a helium-neon laser, and the detector is a photomultiplier tube.
state of the reflected beam is measured using the appropriate

optical equipment and its intensity is measured using a light-
detecting photomultiplier tube.
The polarized light source is a helium-neon laser that emits
a coherent beam of red light at a wavelength of 632.8 nm. Most
common ellipsometers are usually equipped with a single-
wavelength light source. However, because it is crucial that
the nanofilm being studied does not absorb the incident light,
spectroscopic ellipsometers have been developed that can
produce light beams of several different wavelengths to avoid
such problems.
The polarizer linearly polarizes the coherent unpolarized
light beam from the helium-neon laser. The polarizer is a spe-
cial filter that transmits light only if the polarization axis of
the light lines up with that of the analyzer. If the light does
not line up with the polarization axis of the polarizer, then
the light is separated into its parallel and perpendicular com-
ponents relative to the polarization axis and only the parallel
component is transmitted. The linearly polarized beam then
passes through another optical element, the quarter-wave
plate. The quarter-wave plate is also commonly referred to as
a retarder. Its purpose is to transform the linearly polarized
beam into an elliptically polarized beam. The wave plate is an
anisotropic material whose refractive index depends on the
orientation of the propagating wave. P-polarized waves and
s-polarized waves travel with different speeds through such
a material. The thickness of the wave plate can be chosen to
yield a beam whose components are exactly 90° out of phase
with one another.
When linearly polarized light reflects off a surface, as previ-
ously explained, elliptically polarized light is typically pro-
duced. The rotating null ellipsometer produces an incident
beam of varying ellipticities by varying the polarizer angle
prior to the wave plate. When the beam reflects off the sam-
ple, the ellipticity of the beam changes. If the ellipticity is
just right, then the change produced in the polarization of the
beam by reflection from the sample produces a linearly polar-
ized beam. The analyzer, which is identical to the polarizer, is
then rotated until the polarization axis of the analyzer is per-
pendicular to the polarization axis of the reflected beam. The
polarizer and analyzer angles are rotated sequentially until a
null is located. These angles are used to determine the opti-
cal parameters Δ and ψ for the sample. In practice, the instru-
ment completes this computation and the values for Δ and ψ
are output.
4
Ellipsometric Thickness (nm)
0
0 1 2 3 4 5 6 7 8 9 10
Bilayer Number
Figure 4.11 Thickness data as determined by ellipsometry for the layer-by-layer con-
struction of a typical polyelectrolyte nanoassembly. The bilayer represents a layer of poly
cation complexed with a layer of polyanion.
A beautiful illustration of ellipsometry involves charac-

terizing the sequential deposition of a polycation and a poly-
anion on a silicon surface. Polyelectrolyte nanoassemblies
(the protocols and applications) are discussed in Chapter 5.
Essentially, a clean silicon substrate is immersed into a solu-
tion of a polycation for about 5 minutes. It is then removed,
washed, dried, and then immersed into a solution of a polyca-
tion. The adsorption of the polyelectrolytes is driven by elec-
trostatic attraction for each other, and the resulting assembly
can be described as a polyelectrolyte bilayer. The procedure
can be repeated to form many layers.
After the construction of each layer, ellipsometry can
be used to determine the film thickness. The thickness can
be determined to about 0.2 nm. Usually some effort has to be
made to ensure that the thickness is determined from the same
area on the substrate. Figure 4.11 shows how the thickness of
a layer-by-layer polyelectrolyte nanoassembly changes with
each successive layer.
4.4 SURFACE PLASMON RESONANCE

Other techniques aside from ellipsometry exist to determine
the refractive index and thickness of a thin nanofilm. A sur-
face plasmon resonance (SPR) sensor is another optical method
that is employed to detect changes in thickness and refrac-
tive index of very thin organic films adsorbed to a metallic
surface. The SPR is often used to detect interactions between

molecules, and it has emerged as perhaps the most widespread
“surface method” for detecting and quantifying interactions
between biological macromolecules at the nanoscale.
4.4.1 Principles of SPR

When incident light strikes the interface between a substance
with a high index of refraction and another substance with
a lower index of refraction, the light is completely reflected
as long as the angle of the incident light is greater than that
of the critical angle (see section on Evanescent Waves in the
Appendix). In fact, the critical angle θcritical for the interface
between any two substances can be calculated by applying
Snell’s law:
⎛n ⎞
θcritical = arcsin ⎜ 2 ⎟ (4.17)
⎝ n1 ⎠
where n1 and n2 are the indices of refraction of the more

dense and less dense substances, respectively (n1 > n2). This
phenomenon of complete reflection of the incident light is
termed total internal reflection. Total internal reflection
is normally observed when visible light is shone upon the
interface between a glass prism (n = ~1.5) and water (n = ~1.3)
at θincidence > θcritical. However, if the surface of the prism fac-
ing the aqueous solution is coated with a thin layer of silver
or gold as shown in Figure 4.12, then total internal reflection
Glass prism
Incident Reflected
light light
Metallic layer
(Au or Ag)
Surface plasmon
wave
Aqueous solution
Figure 4.12 The generation of a surface plasmon wave. When polarized light is shone
at the correct angle through a glass prism onto a metal-water interface, then a surface
plasmon wave is generated that propagates along the interface. This results in a reduction
of intensity of the reflected light.
is not always observed. This loss of total internal reflection

occurs because some of the incident light is “channeled”
into the metal-water interface where it generates oscillating
waves of surface charge density that move along that inter-
face. These oscillating waves of surface charge density are
called surface plasma waves or surface plasmons, and the
phenomenon of their creation serves as the basis for SPR
sensors.
The creation of these surface plasmons is angle depen-
dent, meaning that there exists an angle of the incident light
(greater than θcritical) at which the generation of the surface
plasmons reaches a maximum. This angle is defined as the
surface plasmon resonance angle θspr (or SPR angle). The
SPR angle therefore represents the greatest “channeling” of
the incident light into the metal-water interface and conse-
quently the greatest reduction in the intensity of the reflected
light. So by measuring the reduction in reflectance as a func-
tion of the angle of incidence, the SPR angle (or angle at
which a global minimum in reflectance is observed) can be
determined.
Of more practical use, it turns out that the electric field asso-
ciated with the surface plasmons is not completely contained
in the metal-water interface, but stretches slightly into each
medium, decaying exponentially as it extends away from the
interface. Because of this penetration into each medium, any
changes to the refractive index near the metallic surface alter
the properties of the electric field, which in turn alters the
SPR effect. Therefore, the SPR angle is sensitive to even very
small changes in the index of refraction near the metal-water
interface, such as might occur when a protein or polymer film
is adsorbed to the metallic surface. In fact, for a given wave-
length λ of light, the change in θspr is related to the change in
refractive index at the surface ∆nsurface and the change in the
thickness of the thin film ∆d by
Δθspr (λ) = c1Δnsurface + c2Δd (4.18)
where c1 and c2 are constants.

An SPR sensor typically operates by detecting changes in
the SPR angle during the adsorption of molecules to the metal
surface. Because ∆θspr is a function of ∆nsurface and ∆d, which
in turn are a result of molecular interactions at the surface,
an SPR sensor serves as a sensitive method to detect binding
events of molecules to a surface. It should also be noted that
∆θspr is typically reported in resonance units (RU), where 1000

RU corresponds to a change of 0.1°. For common analytes such
as proteins, the correlation between ∆θspr and the amount of
substance adsorbed to the surface has been determined and a
1000-RU shift is approximately equivalent to an adsorption of
0.1 ng/mm2 for most proteins.
It should be noted from Equation 4.18 that the SPR angle
is a function of both the thickness d and the index of refrac-
tion n of the thin film, meaning that neither parameter can be
extracted by itself without making certain assumptions (i.e.,
a change in thickness could be extracted if it were assumed
that the refractive index remained constant). For this reason,
the change in SPR angle is said to measure the effective refrac-
tive index or a thickness- and RI-related parameter. In order
to resolve thickness and absolute refractive index simultane-
ously, another method must be used, such as dual polarization
interferometry.
4.4.2 SPR Instrument Setup

A schematic diagram of a common SPR sensor (such as the
prevalent Biacore instrument) is shown in Figure 4.13. The
primary component is a prism that has been coated with a gold
sensor surface so that surface plasmons are generated when
Sample To waste
Flow cell
Gold sensor surface
Glass prism
Polarized
Detector
light source
Figure 4.13 Typical setup of an SPR instrument. A polarized light source is used to
excite the SPR effect at the gold-water interface. The detector monitors the changes in the
intensity of the reflected light as a function of angle of incidence in order to detect changes
in the SPR angle that may result from the adsorption of a thin nanofilm to the gold sensor
surface. A flow cell allows for the easy exposure of the gold sensor surface to the desired
sample solution.
polarized light of a particular wavelength is shone upon it at

the SPR angle. This gold sensor surface serves as the bottom
of a flow cell through which solutions of various analytes can
be passed. If the analytes interact with the sensor surface (or
with molecules that have been pre-adsorbed to the surface),
then a change in the SPR angle is observed by the detector and
recorded by the computer software. The flow cell itself is typi-
cally very small (20–60 nL) and common flow rates are on the
order of 1–100 µL/min.
While SPR sensors can be employed for a variety of pur-
poses that involve the characterization of ultra-thin films,
they are primarily used as biosensors—to detect interactions
between proteins and substrates, between strands of DNA,
and between drug molecules and protein targets. SPR can also
be used to monitor the formation of self-assembled monolay-
ers on gold surfaces, particularly the formation of alkyl-thiol
monolayers.
The SPR effect is not limited to planar surfaces, but can
also be observed using gold or silver nanoparticles in solution.
In this case, the nanoparticles have a wavelength-dependent
absorbance that is a result of the excitation of surface plasmons
within the nanoparticles, and this wavelength is measured
rather than an angle-dependent response as with planar SPR
(which would be difficult to measure for a nanoparticle). The
wavelength at which the maximum absorbance is observed is
a function of the particle size and particle shape. For example,
spherical gold nanoparticles with a diameter of ~13 nm exhibit
a maximum absorbance at 520 nm.
The wavelength at which the SPR effect is maximized
also depends on the dielectric environment (i.e., the refrac-
tive index) near the nanoparticle surface and on the distance
between neighboring nanoparticles. These two properties
have allowed gold nanoparticles to be used as sensitive bio-
sensors. For example, short pieces of DNA called oligonucle-
otides have been covalently attached to gold nanoparticles
through thiol chemistry, and when complementary DNA is
introduced into the solution, the oligonucleotide-coated nano-
particles aggregate as they hybridize to the strands in solution.
This aggregation causes the distance between neighboring
nanoparticles to decrease and a blue-shift in the wavelength
of maximum absorbance is observed. In this manner, gold
nanoparticles can be used to determine the complementar-
ity of two DNA sequences. Similar methods have also been
developed to examine the changes in refractive index that
occur on the surface of a gold nanoparticle when something
adsorbs to its surface—for example, when an antigen binds to

a surface-bound antibody.
4.5 DUAL POLARIZATION INTERFEROMETRY

Unlike other surface characterization techniques such as
QCM-D and SPR whose measurements offer only indirect esti-
mations of the mass, density, or thickness of a thin film, dual
polarized interferometry (DPI) is an optical technique that
actually provides an exact method to measure these param-
eters at the solid-liquid interface. Since DPI provides two
independent measurements of the effective refractive index
of a thin nanofilm, it is able to simultaneously determine the
mass, density, and thickness of that nanofilm. A variety of
other waveguide spectroscopic techniques exist that are simi-
lar to DPI, such as optical waveguide lightmode spectroscopy
(OWLS) and coupled plasmon waveguide resonance (CPWR)
spectroscopy, but only DPI is discussed here.
4.5.1 Waveguide Basics

As may be recalled from our discussion of the critical angle
of reflection, if light is shone on the boundary between two
substances, the light is totally reflected within one substance
as long as it is shone at an angle greater than the critical
angle (θcritical) and if the refractive index of the second sub-
stance is less than that of the first (as shown in Equation A.2
in the Appendix). Such a phenomenon is termed total internal
reflection.
With that in mind, suppose we were to construct a layer of a
substance with one refractive index and sandwich it between
two layers of a substance with a lower refractive index. In such
a setup, we would expect that light shone at an angle greater
than θcritical could undergo total internal reflection inside the
middle layer, alternately reflecting off the top and bottom lay-
ers as it passed through. Indeed, such behavior is observed and
the setup is called a waveguide, with the middle layer termed
the waveguide core and the top and bottom layers called the
cladding regions, as shown in Figure 4.14. When light is shone
upon the edge of a waveguide, total internal reflection can
occur and the light passes through the core and emerges from
the other side of the waveguide. As a side note, a similar appli-
cation of total internal reflection is responsible for the ability
of fiber optic cables to transmit light over long distances (the
interior of the fiber optic cable is of a higher refractive index
than the surrounding material).
Evanescent
Cladding region field (~100 nm)
Light source Waveguide

Waveguide
Core
core
Cladding region Evanescent

field (~100 nm)
Figure 4.14 The setup of a basic waveguide. A waveguide is typically composed of

a waveguide core sandwiched between two cladding regions. Light is transmitted through
the waveguide core via total internal reflection. At each surface between the core and the
cladding regions, an evanescent field is generated that extends ~100 nm into the cladding
region.
When light undergoes total internal reflection within the

waveguide, it turns out that the light is not completely con-
tained in the core, but it extends slightly into the cladding
regions. This extension is called the evanescent field (see
Equation A.1 in the Appendix). The evanescent field decays
exponentially, moving away from the surface of the waveguide
core, and at a distance of ~100 nm into the cladding region its
presence is negligible. Because of the interaction that occurs
between the evanescent field and the cladding region, the
speed at which the light propagates through the core of the
waveguide depends slightly on the refractive index of the area
of the cladding regions near the core’s surface. Thus, light
traveling through a waveguide can be a sensitive measure of
any changes to the refractive index that occur within the first
hundred nanometers of the core’s surface, such as when a thin
nanofilm is deposited upon the surface of the waveguide core.
The problem then becomes how to accurately measure the
speed at which the light travels through the waveguide (or in
other words, how to measure the refractive index of the film
deposited on the core’s surface). To solve this issue, we turn to
interferometry.
4.5.2 Waveguide Interferometry and the

Effective Refractive Index
Interferometry is the study of the ways in which light waves
interact or interfere with each other. Perhaps the most well-
known example of early scientific interferometry is the dif-
fraction pattern observed by Thomas Young in his double-slit
experiments. Young noticed that if he shone a light beam on
a screen that contained two closely spaced slits, then a char-
acteristic interference pattern of alternating light and dark
bands could be observed on a second screen located behind

the first, as shown in Figure 4.15. This interference pattern
could be explained by appealing to the wave nature of light—
the light bands represented regions where the light waves
emerging from each slit interfered with each other construc-
tively, whereas the dark bands were the result of destructive
interference of the light waves.
It is possible to construct a waveguide-based interferometer
analogous to Young’s interferometer using a setup as shown
in Figure 4.16. In this type of waveguide, there are two wave-
guide cores rather than one, and the cores are separated by a
thin cladding region. The top waveguide core is termed the
sensing waveguide and the bottom core is termed the refer-
ence waveguide. If a broad beam of light is shone on the edge
of this type of “waveguide stack,” the light is totally internally
Screen
Light source
Interference fringes
Slit in the far field
Double slit
Figure 4.15 The classic Young’s double-slit experiments. Interference between the two
wavefronts emerging from the double slit produce an interference pattern on a screen placed
some distance away.
Solution in Solution out

CCD
Sensor waveguide
Laser Reference waveguide
Silicon oxide
Interference fringes
in the far field
Figure 4.16 Typical architecture of a DPI flow cell. Light enters the stacked waveguide
and upon exiting generates an interference pattern in the far field due to a phase shift. The
phase shift occurs due to the adsorption of material onto the sensor waveguide.
reflected through both of the core regions and emerges from

the other side to generate an interference pattern, as shown in
Figure 4.16.
This waveguide interferometer can be used to determine
the refractive index of a small nanofilm if the top cladding
region is replaced by a sensing region (generally an aqueous
solution maintained by a fluidic cell, as shown in Figure 4.16).
In this setup, light travels through the sensing waveguide
at a speed that depends in part on the refractive index of
the area within ~100 nm of the sensing waveguide surface
(i.e., within the evanescent field). Hence, any thin films or
molecules that are adsorbed to the sensing waveguide sur-
face change the speed at which the light travels through the
sensing waveguide (because the adsorbed molecules change
the refractive index of the sensing region within the evanes-
cent field). On the other hand, light always travels through
the reference waveguide at the same speed because the clad-
ding regions on either side are fixed. And, since the relative
speeds of the light emerging from the two waveguides affects
the interference pattern produced, the interference pattern
changes as a thin film is adsorbed to the sensor surface. More
specifically, the relative positions of the light and dark bands
(or fringes) shift as the film adsorbs to the surface. By using
Maxwell’s equations of electromagnetism and after applying
some rather complex mathematics, the shifts in the fringe
pattern can be used to calculate the refractive index of the
film adsorbed to the surface of the sensing waveguide. Thus,
a waveguide interferometer is able to characterize a thin
nanofilm.
The refractive index measured by an interferometer
such as described above is not the true refractive index of
the film, but rather the effective refractive index, which
is a complicated function of both the absolute refractive
index and the thickness of the film. Incidentally, this kind
of data is the same as that obtained using SPR, where the
SPR angle change is a function of the change in absolute
refractive index and the change in thickness of the film (see
Equation 4.18 in Section 4.4.1). Normally the effective refractive
index cannot be dissected into the absolute refractive index
and thickness; however, if two different types of measure-
ments of the effective refractive index are made, then it is
possible to mathematically calculate both parameters. DPI
makes these two measurements by using two different types
of polarized light.
4.5.3 Principles of Dual Polarization Interferometry

DPI uses a waveguide interferometer setup as described
above, only the light source is modified such that it alter-
nately produces two different types of linearly polarized
light via a polarizer switch (see Figure 4.16). One type of
polarized light, named the transverse magnetic (TM) polar-
ized mode, is composed of light waves with an electric field
that oscillates perpendicular to the direction of the wave-
guide core, as shown in Figure 4.16. The other type, called
the transverse electric (TE) mode, has light waves with an
electric field that oscillates parallel to the direction of the
waveguide. Incidentally, if we define the waveguide core as
the plane of incidence, then the TM and TE modes directly
correspond to the s- and p-polarized light modes described
in our discussion of ellipsometry. Both of these polariza-
tion modes generate evanescent fields that extend into the
cladding or sensing regions, but the fields produced by each
mode are of different intensities and decay at different rates.
Therefore, each polarized mode generates its own interfer-
ence pattern on the detection screen and consequently each
polarized mode provides a separate calculation of the effec-
tive refractive index.
It is important not to be misled by the direction of oscil-
lation of the two polarized modes into thinking that no eva-
nescent field would be generated by the TE mode, which
oscillates parallel to the direction of the waveguide core.
Regardless of polarization, the light waves still undergo total
internal reflection (meaning the light beams still “bounce”
off the cladding and/or sensing regions) and the two polar-
ized modes still generate evanescent fields in the surround-
ing regions.
For the effective RI determined by each polarized mode, a
large number of absolute refractive index and thickness val-
ues can be calculated that could possibly yield the observed
effective RI, as shown in Figure 4.17. However, there is only
one unique pair of absolute RI and thickness values that may
generate the observed effective RI for both polarized modes.
This pair represents the actual value of the absolute RI and
the thickness of the film on the sensing waveguide surface.
Therefore, by using two different polarized modes of light
and by calculating the unique solution pair, DPI can be used
to determine the actual RI and thickness of thin nanofilms.
The technique is so sensitive that thickness changes of less
than 1 angstrom are detectable. Furthermore, if one can
(a) Fringe view (TM, TE)
(b) Sensor phase response

Polarization 1
Polarization 2
Response
Time
Maxwell’s
equations
(c) Equivalent homogeneous layer

Thickness
Polarization 2
Polarization 1
Absolute RI (density)
Figure 4.17 (a) Typical representation of the fringe pattern observed in a DPI experi-
ment. (b) The sensor phase response for the two polarization states. (c) Effective refractive
index plots showing a unique solution for TE and TM modes. The point at which the curves
cross gives the absolute refractive index and thickness of the film.
assume that the RI of the thin film is a linear function of

the density of its contents (a good assumption for many thin
films), then the RI can be manipulated to yield density δ of
the film according to
n film − nbuffer
δ= (4.19)
dn film /dc
where dnfilm/dc is the change in refractive index of the film

per change in content density and nfilm and nbuffer are the
refractive indexes of the film and buffer (or solvent), respec-
tively. In order to get the mass per unit area of the film, one
need merely multiply the calculated density value by the
average thickness of the film. Thus, DPI can be used to cal-
culate the average density, mass, and thickness of the film
simultaneously.
4.5.4 Parameters of a DPI Instrument

and Common Applications
A commonly used dual polarized interferometer is manufac-
tured by Farfield Group Ltd, Manchester, UK. In their instru-
ment, the DPI sensor surface is composed of silicon oxynitride,
which allows for a wide range of covalent surface modifica-
tions through siloxane chemistry (discussed in Chapter 5).
The flow cell that exists above the sensor surface has a volume
of 2 µL, and typical flow rates are on the order of tens of µL/
min. Their instrument boasts thickness detection limits of <1
angstrom and mass detection limits of 100 fg/mm2. In addi-
tion, it can make multiple measurements per second, which
allows it to monitor surface alterations in real time.
DPI has been widely marketed as a biosensor to measure
the interactions between proteins and their substrates and
to study the formation of membranes, supported lipid bilay-
ers, and the interaction of surfaces with lipid vesicles. A
non-biological application of DPI is the characterization of
the physisorption of nanospheres to hard surfaces. Because
it is able to simultaneously determine both thickness and RI
(and consequently density and mass), DPI is also routinely
used to monitor the changes in surface morphology of vari-
ous thin films. For example, DPI can be used to estimate the
shape of a protein adsorbed to a surface under conditions of
differing pH.
4.6 SPECTROSCOPIC METHODS

Spectroscopy is the study of the ways in which electromagnetic
radiation interacts with matter. Spectroscopy is an invalu-
able field of chemical analysis, and different spectroscopic
methods are able to determine identity, concentration, and
structural information of chemical compounds, among other
useful information. Because the electromagnetic spectrum is
so broad, different spectroscopic methods are employed for
different regions of the spectrum. Figure 4.18 depicts a listing
Increasing energy
Wavelength (m)
10–13 10–11 10–9 10–7 10–5 10–3 10–1 101
Microwave
X-ray Ultraviolet
Infrared
Gamma ray
Radio
Visible
1021 1019 1017 1015 1013 1011 109 107

Frequency (Hz)
3 × 1010 3 × 108 3 × 106 3 × 104 3 × 102 3 × 100 3 × 10–2 3 × 10–4

Wavenumber (cm–1)
Spectroscopy Typical Range Transition Studied
X-ray absorption
0.1–100 Å Inner electron energy states
and diffraction
UV-vis absorption Outer (bonding) electron

200–800 nm
and fluorescence energy states
IR absorption and 0.8–300 μm

Vibrational energy states
raman scattering (14000 to 30 cm–1)
Nuclear magnetic Spin of nuclei in magnetic

0.6–10 m
resonance field
Figure 4.18 The electromagnetic spectrum and a listing of common spectroscopic

methods used to interrogate those regions of the spectrum. Note that the energy of light
increases at smaller wavelengths and higher frequencies.
of the common regions of the electromagnetic spectrum as a

function of wavelength and frequency. It also lists the most
common spectroscopic methods employed in each region
of the spectrum and the kinds of transitions those methods
probe.
The following sections examine the application of a variety
of spectroscopic techniques to the study of nanomaterials. In
order to better understand these methods, it is appropriate to
begin with a brief discussion of the interactions between light
and matter.
4.6.1 Interactions Between Light and Matter

As described earlier (see Section 4.3 on ellipsometry), in many
instances light is best thought of as an oscillating electro-
magnetic wave with a characteristic energy E. The energy of
a photon of this electromagnetic radiation is a function of its
frequency v (or its wavelength λ), and can be calculated by
Einstein’s famous equation
hc
E = hv = (4.20)
λ
where h is Planck’s constant and c is the speed of light. This

equation was introduced in Chapter 2; because of its impor-
tance we repeat it here.
Molecules also have characteristic energy states, but it
turns out that a given molecule cannot exist in any arbitrary
energy state. Rather, the possible energy states for a molecule
are quantized, meaning that the molecule is only allowed to
exist in a finite number of discrete energy states. These dis-
crete allowable energy states are the sum of several quantized
aspects of the molecule, such as the energies of its electrons
around their respective nuclei, the interatomic vibrations that
exist in the molecule, and the rotations of the molecule around
its center of mass. We can also say that a given molecule has
quantized (or discrete) electronic, vibrational, and rotational
energy states and that it can only exist at those energy states
(or at a sum of those energy states). The lowest energy state
of a molecule is termed the ground state and higher energy
states are referred to as excited states. When energy is chan-
neled into a molecule (say, by heating it or allowing it to absorb
radiation), it may enter an excited state. Following this excita-
tion, the molecule may relax from the excited state to a lower
excited state or to the ground state. This relaxation is often
accomplished by the emission of electromagnetic radiation
(and consequently the frequency and the wavelength) of the
emitted light is the exact difference between the upper and
lower energy states. Using Einstein’s equation (Equation 4.20),
this process can be written as
hc
E1 − E 0 = hv = (4.21)
λ
where E1 – E0 is the difference in energy between the higher and

lower energy states. Therefore, by measuring the wavelength
of the emitted light, one can calculate the energy difference

between the higher and lower states of the molecule.
On the other hand, when electromagnetic radiation is shone
upon a substance (rather than being emitted from it), several
processes may take place (see Figure 4.19). Typically, the
Non-radiative
relaxation
Excited Electronic State

and associated
vibrational states
Eincoming light = ΔE1 ΔE1 ΔE2 Eemitted light = ΔE2
Ground Electronic State

and associated
vibrational states
Absorption Fluorescence
(a)
Non-radiative
relaxation
Intersystem
crossing
Excited Electronic State
and associated
vibrational states
Triplet state
Eincoming light = ΔE1 ΔE1 Eemitted light = ΔE2
ΔE2
Ground Electronic State

and associated
vibrational states
Absorption Phosphorescence
(b)
Figure 4.19 Energy diagrams depicting the transitions that occur during fluores-
cence (a) and phosphoresence (b). Note that the energies of the photons (either those being
absorbed or those being emitted) are identical to the corresponding difference in energy
states of the molecule.
radiation is transmitted through the matter at a velocity that

depends upon the refractive index of the medium. However,
if the frequency (or energy) of the incoming light happens to
match the exact difference between two of the allowed energy
states in a molecule, then the incoming light may be absorbed
by the molecule. This outcome, in turn, promotes the molecule
to an excited energy state and decreases the intensity of the
light at that frequency according to the proportion absorbed by
the molecule. This entire process is called absorption, and the
absorption of different frequencies of electromagnetic radia-
tion corresponds to different types of energy transitions in a
molecule. For example, absorption of light in the ultraviolet
and visible region of the electromagnetic spectrum typically
causes transitions between electronic energy states of a mol-
ecule. Alternatively, absorbed infrared radiation causes a tran-
sition in the vibrational states of the molecule. It is important to
emphasize that the incoming radiation is only absorbed by the
molecule if the frequency of the radiation matches the exact
energy difference between allowed energy states of the mole-
cule. Otherwise, the radiation is only transmitted through the
substance. Therefore, if one is able to measure the frequencies
at which a molecule or substance absorbs the incoming radia-
tion, then valuable information about that molecule can be
determined, such as the type and strength of different bonds
that exist in the molecule.
After the molecule has absorbed the incoming radiation, it
ordinarily returns to the ground state after some short amount
of time due to one of several relaxation processes. For some
molecules, fluorescence is a common relaxation process in
which the molecule relaxes by reemitting light, generally of
a lower energy (longer wavelength) than the absorbed light.
Fluorescence is typically observed in the ultraviolet-visible
region of the electromagnetic spectrum and is depicted sche-
matically in Figure 4.19. Nonradiative relaxation is the most
common relaxation process, in which the excited molecule
relaxes without reemitting any electromagnetic radiation.
This process typically occurs in small steps by the conversion
of the excited energy into kinetic energy through collisions
with nearby molecules, producing heat. Nonradiative relax-
ation is also depicted in Figure 4.19.
Phosphorescence is another relaxation process that typi-
cally occurs in the UV-visible region. It is observed when an
excited electron undergoes nonradiative intersystem crossing
to a more rare and slightly more stable excited state called
a triplet state. When the excited electron then relaxes back

to a state of lower energy, a photon is produced at a longer
wavelength (smaller frequency) than the absorbed light.
Phosphorescence is a much more rare phenomenon than fluo-
rescence and is consequently less important to our discussion
of spectroscopic techniques.
In summary, molecules may exist only at discrete, quan-
tized energy states. By examining the ways in which a given
molecule interacts with light, one can gather information
about its energy states, thus providing valuable insight into the
identity of the molecule, the strength and type of its chemical
bonds, and the concentration of the molecule in the substance
being studied. For our elementary purposes, those observa-
tions serve as the basis of spectroscopy. Let’s consider several
important types of spectroscopy and their applications to the
study of nanomaterials.
4.6.2 UV-Visible Spectroscopy

4.6.2.1 Principles of UV-Visible Spectroscopy
UV-visible spectroscopy uses the transmission of visible and/
or ultraviolet light through a sample to determine the pres-
ence and/or the amount of material that absorbs light within
the sample. As mentioned in our discussion of the interactions
between light and matter, absorption of a photon in the ultra-
violet and visible wavelength range (~190 nm to 800 nm) typi-
cally produces an electronic transition within the absorbing
molecule. The absorption of a photon increases the energy of
the molecule, causing the promotion of the molecule out of a
low-energy ground state electron configuration into an excited
energy state. The region of a molecule that absorbs the light and
undergoes the electronic transition is called the chromophore.
For example, the amine group –NH2 absorbs light at a wave-
length of about 190 nm. The absorption causes an electron from
the lone pair on nitrogen to be excited into an antibonding
molecular orbital. This transition is denoted as n → σ*. For our
purposes, we consider only the amount of light absorbed and
the corresponding absorption wavelength (λmax). The exact
electronic nature of the transitions is less important. It should
be noted that the actual absorption of chromophores in a given
molecule depends on the exact electronic environment of the
chromophore within the molecule. As a matter of convention,
we also note that values for the UV-visible spectrum are typi-
cally reported in wavelengths of light (generally nm).
When absorption takes place in the sample, there is a

decrease in the intensity, or radiant power, of the light beam
passing through the sample. The transmittance T of the sample
is defined as the ratio of the intensity of the beam after pass-
ing through the sample (I) to the original intensity of the beam
(Io). The absorbance A of the sample is defined as the negative
logarithm of transmittance, as shown in Equation 4.22.
⎛I ⎞
A = log ⎜ o ⎟ = − log T
⎝ I ⎠
(4.22)
For samples with relatively low concentrations, the absor-

bance of the sample can be directly related to its concentra-
tion. This relationship is known as the Beer-Lambert law and
is given in Equation 4.23.
A = cεl
(4.23)
where c is the concentration of the sample, l is the sample

path length, and ε is a parameter known as the extinction
coefficient and is also the molar absorptivity of the sample.
The molar absorptivity is an important parameter that gives
the probability of absorbing a photon for a particular mol-
ecule at a specific wavelength of light. Its value is specific to
a particular chromophore and does not depend on concen-
tration or path length. If the path length and ε of a sample
are known, then the density of the sample’s contents can be
calculated from the absorbance value. Alternatively, if the
content density is known, then the path length can be cal-
culated. In the context of a thin film (Figure 4.20), the path
length is simply the film thickness (d), and in this manner
the thickness of a nanofilm can be determined using UV-vis
spectroscopy.
4.6.2.2 Setup of a UV-Visible Spectrophotometer

At a basic level, a UV-vis spectrophotometer consists of a light
source, a dispersive device, a sample holder, and a detector,
along with several optical components. Common kinds of
spectrophotometers are double-beam spectrophotometers,
which use a beam splitter to allow the light beam to pass
through a sample cell and a reference cell alternately through-
out the measurement process. The beam is diverted into the
reference cell by the splitter several times per second, and the
Incident light Transmitted
from light light to
source detector
Characterization at the Nanoscale
Thin film
Transparent
substrate
d
Figure 4.20 Transmission mode UV-vis absorption spectroscopy can be used to determine the thickness of a thin nanofilm. Beer’s law is used just as with bulk phase
measurements, but the thickness of the film replaces the path length of the sample cell in the equation, as shown.
137
Reference cell
Beam
chopper
Light source
Detector
Sample cell
Figure 4.21 A schematic of a common double-beam spectrophotometer used to make

bulk phase measurements.
absorbance of the reference cell is measured and automatically

compared with the absorbance of the sample cell to produce
a more accurate value of the absorbance of the species being
measured by correcting for fluctuations or variations in the
lamp intensity over time. A schematic of a double-beam spec-
trophotometer is shown in Figure 4.21.
In practice, UV-visible spectrometers contain two differ-
ent light sources, one to produce light in the UV range and
one to produce visible light. The light sources are never run
simultaneously. Instead, as the spectrophotometer measures
the absorbance of a sample across a predetermined range of
wavelengths, the light source switches from the visible source
to the ultraviolet source. The light change is generally set to
occur around 360 nm to allow for the highest light intensity at
each wavelength.
4.6.3 The Absorption of Visible Light by a Nanofilm

As an example of the application of UV-visible spectroscopy
to characterize the growth of a nanofilm, consider the anionic
polymer PAZO whose structure is shown in Figure 4.22. The
PAZO monomer unit contains two phenyl groups bridged by
an azo group (–N=N–). This unit functions as a chromophore
and absorbs light at about 360 nm due to a π → π* transition.
Thus, an aqueous solution of PAZO appears orange because
of the broad absorption of light centered at 360 nm. The poly-
mer polyethylenimine (PEI) encountered previously contains
amine groups that absorb light at about 190 nm. There is no
absorption in the visible region and thus an aqueous solution
of PEI is colorless. Both PAZO and PEI are commonly used to
0.40
n
0.35 HN
SO2
0.30 PAZO
0.25
Absorbance
0.20 N
N
0.15 λmax
0.10 CO2–
OH
0.05
0.00
250 350 450 550 650
Wavelength (nm)
(a)
0.5
0.4
Absorbance at λmax
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
Bilayer Number
(b)
Figure 4.22 (a) The absorption spectrum of PAZO (structure shown in the inset). λmax
represents the wavelength at which the absorbance is a maximum. (b) A plot of absorbance
at λmax as a function of bilayer number. The bilayer represents a layer of polycation com-
plexed with a layer of polyanion.
create multilayer films for a variety of applications. We will

discuss these applications and the methods used to fabricate
the multilayers in Chapter 5. About 1-nm thick layers of PEI
and PAZO can be constructed on a glass substrate in a layer-
by-layer fashion. These layers are held together by strong elec-
trostatic interactions between the positively charged PEI and
the negatively charged PAZO. Each PEI and PAZO pair in the
film can be described as a “bilayer.” Figure 4.22(b) shows a
UV-visible spectrum of a PEI and PAZO multilayer film as a

function of the bilayer number. Only the absorbance value at
~360 nm is shown. In this range, only the chromophores
associated with PAZO are able to absorb light (centered at a
wavelength of 360 nm). The PEI is essentially transparent,
and therefore the increase in the absorbance in Figure 4.22(b)
reflects the increase in the amount of total PAZO in the film
after each bilayer.
Example 4.6 Determining Absorbance Values
Estimate the slope of the line of the graph shown in Figure 4.22 and
use it to predict an absorbance value of a film composed of 15 bilay-
ers. How would you determine the concentration of PAZO in the film
from the PAZO’s molar absorptivity? What other technique would be
useful in this determination?
Solution A linear fit to the data yields a slope of ~0.02 (Figure 4.23).
The intercept is close to zero. The slight negative number in the
intercept is likely due to a baseline shift. Thus the equation of the
line is y = 0.02x, where y = A. When x = 15, the absorbance (A)
is 0.75. From Beer’s law, A = cεl, where l represents path length,
or in this case the thickness of the film. We can estimate l using
ellipsometry, and if we know ε (molar absorptivity), we can obtain
the concentration, c.
0.40
0.35
0.30
0.25
Absorbance
0.20
0.15
0.10
0.05 y = 0.0247x – 0.0192
0.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Bilayer Number
Figure 4.23 The absorbance (at λmax) data of a polyelectrolyte multilayer film.
The bilayer represents a layer of polycation complexed with a layer of polyanion.
4.6.4 Molecular Fluorescence Spectroscopy

4.6.4.1 Principles of Fluorescence and
Fluorescence Quantum Yield
As described in Section 4.6.1 on the interactions between light
and matter, fluorescence is the process whereby a molecule
that has been excited by the absorption of radiation relaxes to
a lower energy state by emitting a photon of lower energy than
the photon that was absorbed. Some types of molecules do not
exhibit fluorescence; molecules that do fluoresce are called
fluorophores. For most fluorophores, fluorescence is typically
observed in the ultraviolet to visible range of the electromag-
netic spectrum, meaning that it corresponds to transitions in
the electronic state of the molecule (much like UV-vis absorp-
tion spectroscopy). As with UV-vis spectroscopy, the wave-
length of light (generally nm) is used to describe fluorescence
spectra rather than wavenumber (cm–1) or frequency (Hz)
units.
Each fluorophore typically has a characteristic absorption
profile, which is identical to the UV-vis absorption spectrum
for the molecule. Many of the wavelengths at which the fluo-
rophore absorbs UV-vis light excite the molecule to a sufficient
degree such that it fluoresces upon relaxation. Wavelengths
that produce fluorescence when absorbed are referred to as
excitation wavelengths. However, not all excitation wave-
lengths cause the fluorophore to fluoresce to the same degree
or even to produce fluorescent light of the same wavelength.
Therefore, for each excitation wavelength, the fluorophore has
a characteristic fluorescence emission profile, which is the
range and intensity of wavelengths that are produced when
that excitation wavelength is used to excite the fluorophore.
Alternatively, for each emission wavelength there also exists
an excitation profile, which is the range of wavelengths that
produce fluorescence of that emission wavelength and the
intensities of emission associated with each of those excita-
tion wavelengths. Both emission and excitation profiles pro-
vide useful information about the fluorophore and its local
environment, so fluorometers can often be set to determine
either the emission spectrum at a set excitation wavelength
or the excitation spectrum at a set emission wavelength. It
should be noted that because fluorescence-emitted photons
are typically of lower energy (i.e., longer wavelength) than
their excitation photon counterparts, it is not surprising that
the emission spectrum for a given fluorophore usually occurs
at longer wavelengths than its excitation spectrum. Typical
emission and excitation spectra for an optically active poly-

mer are shown in Figure 4.24. The polymer is a derivative of
poly(phenylene vinylene) (PPV), and such materials are being
developed for applications in nano-photovoltaics.
As discussed previously (Section 4.6.1), there are several
alternative relaxation processes that may occur in lieu of
fluorescence, such as non-radiative relaxation and intersys-
tem crossing. Furthermore, reabsorption of the fluorescent
photon by a neighboring fluorophore may occur, especially
at higher concentrations of fluorophore. Therefore, it is con-
venient to define a parameter that represents how often a
given fluorophore successfully produces a fluorescent photon.
Fluorescence quantum yield (Φf ) is the usual measure of such
a parameter and is defined as the ratio of photons absorbed
to photons emitted for a given fluorescing species. As can be
imagined, the correct determination of Φf is especially impor-
tant in the characterization of certain emissive nanomaterials
such as those designed for photoluminescent devices.
120 0.20
0.18
(a) (b)
100
0.16
0.14
80
Fluorescence Intensity
0.12
Absorbance
60 0.10
0.08
40
0.06
0.04
20
0.02
0 0.00
350 450 550 650 750
Wavelength (nm)
Figure 4.24 Typical absorption (a) and fluorescence emission (b) spectra of PPV, an
optically active polymer. The fluorescence intensity is in arbitrary units. The molecules are
excited at ~475 nm (the wavelength corresponding to the peak absorbance) and emits light
at ~575 nm (the wavelength corresponding to the peak fluorescence).
Determination of Φf for a given material A can normally be

accomplished in a rather straightforward manner by compar-
ison with a standard material of known quantum yield Φstd
according to
g a na2
Φ fa = Φ std 2
(4.24)
g std nstd
In this equation na and nstd are the refractive indexes of the

solvents in which the material and the standard were dis-
solved, respectively. gx is effectively defined as the deriva-
tive of the total fluorescence intensity of a fluorophore with
respect to its absorbance of the excitation light (i.e., a measure
of how the total fluorescence intensity changes with respect
to absorbance). For low concentrations of the fluorophore, the
relationship is usually linear (i.e., gx is constant), so gx can be
calculated by integrating the total fluorescence spectrum for a
certain excitation wavelength and then plotting integrated flu-
orescence intensity versus absorbance of the excitation wave-
length for a number of different concentrations.
4.6.4.2 Setup of a Fluorometer for Bulk Phase and

Thin Film Fluorescence Measurements
Figure 4.25 shows the typical setup of a fluorometer. The
detector is usually placed at ~90° from the incident light beam
so as to not confuse light that has been transmitted through
the sample with fluorescent radiation. Optical filters can
Sample cell
Light source
Transmitted light
Emitted light
Detector
Figure 4.25 A schematic of a typical fluorometer.

also be placed between the sample and the detector in order

to ensure that only the fluorescence is detected. The fluores-
cence emission spectrum is gathered by shining a beam of a
certain excitation wavelength on the sample and then scan-
ning the output (emission) wavelength range with the detector
while recording the intensity of fluorescence at each emission
wavelength. Alternatively, the excitation spectrum can be
obtained by setting the detector to monitor a specific emission
wavelength and recording the intensity of fluorescence while
scanning the input (excitation) wavelength range.
In some circumstances the intensity of the emission is too
low to be recorded without amplification. In these situations
it is best to use a photomultiplier tube. It is worth noting that
photomultiplier tubes should be used only in cases of low-
power radiation; otherwise, they will be damaged. A photo-
multiplier tube contains a photocathode surface that emits
electrons when exposed to radiation. It also has multiple other
electrodes, called dynodes. Each photoelectron that strikes a
dynode causes many more electrons to be emitted. This cre-
ates an avalanche effect that results in millions of electrons
for each initial photoelectron to be collected at an anode. The
resulting current is measured in terms of voltage and is related
to the wavelength being scanned.
For bulk phase measurements, the solution phase sample
is generally placed into a fluorescence cuvette and fluores-
cence measurements are made at 90° to the incident beam. For
thin films, a substrate on which the thin film has been built is
placed so that the incident light shines upon the surface of the
substrate at a large angle of incidence (i.e., more parallel to the
surface) and fluorescence is detected at a different angle than
the angle of reflection. Each of these two setups is depicted in
Figure 4.26.
4.6.5 Vibrational Spectroscopy Methods

4.6.5.1 Introduction to Vibrational Modes
When a molecule absorbs infrared (IR) radiation, it is excited
to a higher energy state, as described in Section 4.6.1. As
with the absorption of other types of electromagnetic radia-
tion, the IR absorption process is quantized and the molecule
can absorb only particular frequencies of IR radiation. The
absorption of IR radiation corresponds to a change in energy
of approximately 20 kJ/mol. This is the amount of energy
required to cause covalent bonds to stretch, bend, and twist,
particular combinations of which are called the vibrational
Transmitted light Thin film

Substrate
Reflected light
from surface
Em
Emitted light itte
d ligh
t
Detector
Excitation light Detector

Excitation light from
from light source
light source
(a) (b)
Figure 4.26 Methods for monitoring the fluorescence of samples in a cuvette in solu-
tion phase (a) or in a thin film (b). The detector is strategically placed so that it will inter-
cept the maximum amount of fluorescent light without picking up any of the transmitted or
reflected light.
modes of a molecule. Only those frequencies of IR radiation

that match the natural vibrational frequencies of the covalent
bonds of the molecule lead to IR absorption by the molecule.
The energy absorbed leads to an increase in the amplitude
of the vibrational motions of the bonds in the molecule.
Furthermore, in order for a molecule to absorb IR radiation, it
must possess a dipole moment that changes during the course
of the vibration. A time-varying dipole moment means that
the electron density between a covalent bond is changing.
IR energy is transferred to the molecule when the frequency
of this electrical oscillation is the same as the frequency of
the incoming IR light. Therefore, symmetric molecules that
have no dipole moment such as N2, H2, and O2 do not absorb
IR radiation. The natural frequency of vibration of a bond is
given by
1 k
ν= (4.25)
2πc μ
which is essentially the same as Hooke’s law used to describe

a spring undergoing harmonic oscillation. In this equation,
the constant k is called the force constant of the bond and its
units are N/m. The magnitude of k gives a direct measure of
the stiffness of a covalent bond. µ is known as the reduced
mass. For a simple diatomic molecule containing two atoms of

mass m1 and m2, the reduced mass is given by
m1m2
μ= (4.26)
m1 + m2
As a simple example, consider the structure of H2O. The cen-

tral oxygen atom is sp3 hybridized, meaning that the molecule
is bent with single bonds between the oxygen and hydrogen
atoms. The molecule has a net dipole moment and is thus able
to absorb IR radiation. Figure 4.27 shows three possible ways
that H2O can vibrate. When the two O-H bonds are simulta-
neously increasing and decreasing in length we say that the
vibration is symmetric. If one O-H bond length increases
while the other decreases, then we have an asymmetric vibra-
tion. The molecule can also undergo a bending motion. All
three modes of vibration lead to changes in the dipole moment
and therefore result in the absorption of IR radiation at char-
acteristic frequencies. Since it is easier to bend the molecule
than to stretch it, the bending mode has a relatively small k
value. The implication is that, of the three observed IR absorp-
tion frequencies, the bending mode corresponds to the lowest
frequency. If we replace the H atoms in H2O with deuterium
atoms (D2O), then the reduced mass of D2O is greater than that
of H2O. According to Equation 4.25, all three IR absorption
frequencies for D2O are smaller than those for H2O.
Let’s consider the simple molecule CO2. The central carbon
atom is sp2 hybridized, meaning that CO2 is a linear molecule
with double bonds between the carbon and oxygen atoms.
Asymmetric Symmetric
Bend
Stretch Stretch
O O O
H H H H H H
IR Absorption 1600 cm–1 3760 cm–1 3650 cm–1
O C O O C O O C O
IR Absorption 667 cm–1 2349 cm–1 IR inactive
Figure 4.27 The vibrational modes of H2O and CO2 and the IR absorption associated
with each vibrational mode. The symmetric stretch in CO2 does not produce a net change in
the dipole moment of the molecule and is therefore IR inactive.
Since the molecule is linear and symmetric around the central

C atom, CO2 has a zero net dipole moment (the dipole moments
from the two C=O bonds cancel each other out). When this
molecule vibrates symmetrically, as shown in Figure 4.27, the
net dipole moment remains zero. Thus, the symmetric vibra-
tion of CO2 does not lead to the absorption of IR radiation.
However, the asymmetric and bending modes of CO2 do result
in a net dipole moment, and therefore they lead to the absorp-
tion of IR radiation.
Organic molecules usually contain a variety of functional
groups. Each functional group absorbs a specific IR frequency
that is characteristic of that group. For example, the carbonyl
group (C=O) absorbs around 1700 cm–1. The exact value depends
on the local environment of the functional group, such as the
identity of other atoms that are attached to the carbonyl func-
tional group. Figure 4.28 shows a number of functional group
vibrations and their characteristic IR absorption frequencies.
As a matter of convention, we note that IR spectroscopists typi-
cally report IR values in wavenumbers (cm–1), a unit that is
directly proportional to frequency ν and defined as
ν 1 (4.27)
wavenumber = =
c λ
Functional Group Group Frequency (cm–1)

-C-H (stretch) 2850-2960
=C-H (stretch) 3000-3100
≡C-H (stretch) ~3300
C=C (stretch) 1620-1680
C≡C (stretch) 2100-2260
-O-H (alcohols, H-bonded, stretch) 3200-3600
-O-H (carboxylic acids, H-bonded, stretch) 2500-3000
-N-H (stretch) 3300-3500
-N-H (bend) ~1600
C=O (stretch) 1670-1820
C≡N (stretch) 2220-2260
-S-H (stretch) 2550-2600
-S-S- (stretch) 470-620
Si-O-Si (stretch) 1020-1095
Si-O-C (stretch) 1080-1110
-N=N- (stretch) 1575-1630
Figure 4.28 A listing of common functional groups and their IR frequencies.

where c is the speed of light (in cm s–1) and λ is the wavelength

of light.
By plotting absorption or transmission versus incident IR
frequency, one can map out the exact absorption frequencies
for a sample containing the molecule of interest. This infrared
spectrum provides a fingerprint of the molecular structure of
the molecule.
Infrared spectroscopy is routinely used to elucidate the
structure of organic molecules. We now consider a few spe-
cialized applications of IR spectroscopy suited for the analysis
of nanomaterials.
4.6.5.2 Attenuated Total Reflection IR Spectroscopy

Attenuated Total Reflection Fourier Transform Infrared (ATR-
FTIR) spectroscopy is a powerful spectroscopic technique
used to investigate the structure of adsorbates confined to the
solid-air or solid-liquid interface. The technique is essentially
IR spectroscopy of molecules present at the surface of a solid.
It offers several advantages over its counterpart transmittance-
mode IR absorption spectroscopy. ATR-FTIR offers near sur-
face selectivity with only a minimal amount of sample and
allows for the detection of samples of mass on the order of
nanograms. Furthermore, using an infrared polarizer, it is
easy to determine the orientation of anisotropic (well-ordered)
samples such as self-assembled surfactant or lipid monolayers.
The sample can also be exposed to external conditions (e.g.,
various solvents or different pH conditions), making in situ
studies feasible.
ATR-FTIR makes use of evanescent waves in a manner that
is similar to DPI, as both are able to measure the intensity
of the waves that are formed from total internal reflection
within the surface-sensor medium. One significant differ-
ence is that ATR-FTIR monitors the absorption of IR light of
an adsorbate rather than changes in its thickness and refrac-
tive index like DPI. A system of mirrors is set up so that IR
light is shone upon a crystal called an Internal Reflection
Element (IRE) in such a way that total internal reflection
occurs, as shown in Figure 4.29. As with DPI, the total inter-
nal reflection of the IR beam results in the creation of an
evanescent wave that penetrates the sample above the crys-
tal and decays exponentially moving away from the crystal
surface. (For a refresher on total internal reflection and eva-
nescent waves, see Appendix A). For IR light, the penetration
depth of the evanescent wave is typically on the order of a
few micrometers (0.5–5 µm).
Evanescent field
IRE Crystal
IR light source Detector
Figure 4.29 A schematic diagram of an ATR setup for an FTIR spectrometer. The eva-
nescent wave that is generated at the surface of the IRE can penetrate the overlying region
up to several microns.
If the sample on the surface of the IRE happens to absorb the

frequency of IR light that is being shone through the crystal,
then the IR beam that emerges from the IRE and travels to the
detector has a diminished intensity (or is attenuated). One can
imagine that the adsorbate takes up the energy of the IR beam
through the evanescent wave, but only at those IR frequencies
that match the vibrational modes of the sample. Therefore, it
is possible to obtain an IR absorbance spectrum for an adsor-
bate by scanning the entire IR spectrum and monitoring the
frequencies at which the IR beam is attenuated upon its emer-
gence from the IRE.
There are several common ways to set up an ATR-FTIR sys-
tem to monitor the IR spectrum of a thin nanofilm. Perhaps
the most straightforward method is to build the thin nanofilm
on an external substrate and then clamp the substrate on top
of the IRE, sandwiching the thin nanofilm between the sub-
strate and the IRE. An IR absorbance spectrum of the nanofilm
can then be obtained as described above. While this method is
easy to implement, it does not allow for continuous monitor-
ing of the thin nanofilm during its creation. To acquire that
type of data, it is necessary to construct a flow cell on top of
the IRE and build the thin nanofilm on the surface of the IRE
itself. Using this setup, the IR spectrum can be monitored con-
tinuously, and changes that occur in the spectrum during the
creation of the thin film can be recorded.
For the purposes of determining the IR spectrum of a thin
nanofilm as described above, a horizontal ATR (HATR) setup is
commonly employed. In this setup, the IRE is a parallel-sided
crystal plate with dimensions on the order of 1 cm by 5 cm. For

an IRE of this size, the IR beam is typically reflected between
5 to 10 times at each surface, depending on the exact dimen-
sions of the crystal and the angle of incidence. The upper face
of the crystal is exposed to the thin film through a clamping
method or a flow cell setup, as described. The most common
crystal materials used for HATR setups are zinc selenide
(ZnSe) and germanium. Diamond is a more robust material
that can be used as an ATR crystal, although its higher cost is
a drawback.
4.6.5.3 Reflection Absorption IR Spectroscopy

Another surface-sensitive IR spectroscopy is Reflection
Absorption Infrared Spectroscopy, or RAIRS. In RAIRS, infra-
red light is shone onto a metal surface at a grazing angle of
incidence. The vibrational spectrum of molecules adsorbed
onto the surface is obtained by comparing the intensity of the
reflected light from a clean surface to one covered by a thin
film. The vibrational modes that are observable in RAIRS are
governed by a certain metal surface selection rule.
This states that only modes that have a component of their
transition dipole moment, the complex vector quantity that
includes the factors that change between two states, perpendic-
ular to the metal surface appear in a RAIRS spectrum. Hence,
the appearance of a RAIRS spectrum indicates that the transi-
tion dipole moments of specific functional groups in the film
molecules have a component parallel to the surface normal.
4.6.6 Raman Spectroscopy

4.6.6.1 Rayleigh and Raman Light Scattering
In the opening discussion of the interactions between light
and matter (see Section 4.6.1), we focused on processes such
as absorption, emission, and fluorescence. However, another
interaction called light scattering occurs commonly when
light encounters nanoscale matter. In order to understand the
scope and utility of Raman spectroscopy, a brief introduction
to light scattering is provided.
When light passes through a solution or suspension, much
of the light is transmitted directly through the solution, but
some of the light is scattered in different directions. To better
understand this scattering process, it is useful to model the
scattering of light. Suppose that when a photon interacts with
a molecule, the molecule is promoted to a virtual excited state
as shown in Figure 4.30. This virtual state is short-lived, so
we expect that the molecule almost immediately returns to its
Rayleigh Scattering Raman Scattering
Virtual excited Virtual excited
state state
= E0 ± ΔE
red = E0
tte red
E sc a E scatte
E0 = hν0 E0 = hν0
Etc.
2nd Vibrational state ΔE
1st Vibrational state
Energy state Ground state ΔE
Anti-stokes shift Stokes shift
Figure 4.30 Energy diagrams modeling the Rayleigh and Raman scattering processes. A photon interacts with a molecule, promoting it to a non-quantized virtual
state. The molecule quickly returns to a lower energy state, emitting a scattered photon. If the molecule returns to the same state in which it began, Rayleigh scattering has
occurred. If the molecule returns to a higher or lower vibrational energy state, then Raman scattering has occurred.
151
ground state, reemitting the photon. However, the reemitted

photon may not always be reemitted in the same direction as
it was traveling before interacting with the molecule, which is
scattering. It is important to note that this model of the scat-
tering process should not be confused with absorbance and
reemission of the photon. The incoming photon is not absorbed
by the molecule according to our model. The promotion to the
virtual excited state is a result of a momentary interaction with
the photon and the entire process is not quantized (meaning
it can happen for any energy of light). Absorbance only occurs
when the frequency of the incoming light exactly matches the
difference between allowed energy states of the molecule.
Most commonly, light is scattered elastically (or without los-
ing or gaining any energy), in a process described as Rayleigh
scattering. From Einstein’s equation (Equation 4.20), we know
that the energy of light is directly related to its frequency (or
wavelength). Therefore, since Rayleigh scattering is elastic,
the scattered light has the same frequency (or wavelength)
as before it was scattered. However, due to the wave nature
of light, Rayleigh scattering is a wavelength-dependent pro-
cess, meaning that some wavelengths are scattered to a greater
extent than others. Indeed, it is this aspect of Rayleigh scatter-
ing that accounts for the blue color of the sky. As light from the
sun interacts with particles in the atmosphere, it undergoes
Rayleigh scattering. Blue light is scattered more than the other
wavelengths of light, because Rayleigh scattering is strongest
for the shorter wavelengths of light that are closer in size to the
air molecules responsible for scattering, so it appears as though
the sky is a blue color. Rayleigh scattering is depicted schemat-
ically using our model of light scattering in Figure 4.30.
Raman scattering occurs when light is scattered inelasti-
cally by a molecule. In other words, Raman scattering happens
when the scattered light is of a higher or lower energy after it
has been scattered than before. This increase or decrease in
energy is generally due to a change in the vibrational energy
of the molecule. According to our model of light scattering,
this type of Raman scattering might occur when the molecule
relaxes from the virtual excited state to a vibrational state that
is higher or lower in energy than the state at which the mol-
ecule was previously. In this case, the reemitted (or scattered)
photon is a slightly different frequency (or wavelength) than
before it interacted with the molecule. The shift in the fre-
quency of the scattered light from its original value directly
matches the energy difference between the vibrational states
of the molecule.
In our diagram (see Figure 4.30), Raman scattering is rep-

resented as follows. Suppose the molecule happened to be in
the first vibrational state above the ground state at the time of
interaction with the photon. The molecule would be excited
to the virtual excited state and then immediately relax, pro-
ducing a scattered photon as described previously. Suppose,
however, that the molecule would relax to either the second
vibrational state or to the ground state rather than the first
vibrational state. In this case, the reemitted photon would have
either gained or lost energy according to the exact difference in
energies between the vibrational states, represented by ∆E in
the figure. And, because the energy of light is a function of its
frequency, the higher or lower energy light has a frequency
that is slightly shifted from its original value and Raman scat-
tering has occurred. Not only has the Raman scattered light
changed direction, but it has also shifted its frequency.
Because the frequency shifts of Raman scattered light are a
direct result of changes in vibrational states of the molecule,
it seems reasonable to assume that one could gather simi-
lar information to IR spectroscopy by simply measuring the
shifts in frequency of Raman scattered light. Indeed, this is the
basic idea behind Raman spectroscopy, and frequency shifts
in a Raman spectrum are directly analogous to IR absorption
frequencies. In practice, shifts in frequency of Raman scat-
tered light are monitored and those shifts are matched to spe-
cific bond vibrations within the molecule being studied. At a
basic level, this allows one to identify the functional groups
of the molecule being studied as well as their local chemical
environment.
As a matter of naming convention, one should note that a
frequency shift of Raman scattered light to a lower energy (or
lower frequency) is called a Stokes shift. A frequency shift
to a higher energy (or higher frequency) is called an anti-
Stokes shift. Typically, Stokes shifts are more probable than
anti-Stokes shifts, so their signals are stronger. As might be
expected, Raman scattering is much less likely to occur than
Rayleigh scattering for a given molecule and a given wave-
length of light. Therefore, a Raman spectrum that maps inten-
sity versus frequency of the scattered light has a centralized
and strong peak that represents the Rayleigh scattering of the
incident light. The Rayleigh peak is symmetrically surrounded
by a distribution of much smaller peaks that represent the
Stokes lines at the lower frequencies and the anti-Stokes lines
at the higher frequencies. This peak is often blocked from the
detector by a filter.
4.6.6.2 Surface Enhanced Raman Spectroscopy

Typically, the effects of Raman scattering are very weak.
Under normal conditions, the intensities of the Raman
scattered peaks in a spectrum are ~0.001% of the incident
light. Therefore, if a Raman spectrum is desired for an
analyte with a relatively small concentration (such as for
a thin film deposited on a surface), some method of signal
enhancement must be employed in order to gather any use-
ful information.
One method that is commonly used to enhance the Raman
signal is adsorption of the analyte onto a rough metal surface
or colloidal particle to obtain the Raman spectrum from the
adsorbed species. Most often silver is used, but gold and cop-
per have also been observed to produce similar effects. This
type of signal enhancement is called surface enhanced Raman
spectroscopy (SERS). SERS is a convenient method that allows
amplification of the Raman signal while simultaneously prob-
ing the region near a metallic surface, thus allowing Raman
spectra to be obtained for low quantities of a substance that
are adsorbed to a metallic surface.
The exact mechanism by which adsorption to a metallic
surface produces enhancement of the Raman signal is not
fully understood. Two predominant theories have emerged
to explain the effect. They are called the electromagnetic
enhancement theory and the chemical enhancement theory,
respectively. Each theory is able to explain some observations,
but not others.
The electromagnetic theory claims that the generation of
surface plasmons by the incident light is the means by which
Raman scattering is enhanced. As may be recalled from our
discussion of SPR (Section 4.4), a surface plasmon is a collec-
tive oscillation of the electrons at the surface of a conducting
metal (typically gold or silver) that is induced when light is
shone upon its surface. The generation of surface plasmons
greatly enhances the electromagnetic field near the metal-
lic surface and therefore would be expected to enhance the
strength of the Raman scattered light. This, then, is the funda-
mental idea behind the electromagnetic enhancement theory.
The incident light produces surface plasmons, which enhance
the electromagnetic field near the surface, which in turn pro-
duces a stronger Raman signal.
The chemical enhancement theory appeals to a charge-
transfer mechanism to explain the enhancement of Raman sig-
nal upon adsorption to a metallic surface. The charge-transfer
mechanism essentially states that new electronic states are

made available to the molecule due to its interaction with
the metallic surface. These new allowable electronic states
can be of intermediate energy between the highest occupied
molecular orbital of the molecule and its lowest unoccupied
molecular orbital. Hence, the new electronic states can serve
as resonant intermediaries in the process of Raman scat-
tering. Therefore, excitations of the molecule that use the
charge-transfer mechanism occur at much lower energies
than are normally required for the molecule to be excited.
Thus, Raman scattering can happen more readily and the sig-
nal is enhanced.
SERS is typically carried out on a “rough” metallic surface,
with roughnesses on the order of 10–100 nm. Such surfaces
are commonly manufactured by sputtering or evaporation of
the metal onto a substrate or by roughening of a metallic elec-
trode surface during oxidation-reduction cycles. Colloidal
particles of the metal are also used to produce SERS. These
colloids may either be suspended in solution or adsorbed
onto a substrate. On these types of “rough” metallic sur-
faces, the SERS effect extends up to tens of nanometers from
the surface, allowing for the effective probing of ultra-thin
nanofilms.
4.7 NONLINEAR SPECTROSCOPIC METHODS

4.7.1 An Introduction to Nonlinear Optics
Nonlinear optics deals with the interaction of intense light
with matter. The light source is usually a pulsed laser that
can generate radiation with a strong electric field strength.
Although the theory of nonlinear optics is beyond the scope of
this book, a cursory treatment is provided to facilitate under-
standing of how nonlinear optical effects can be used to probe
the properties of molecules in nanoassemblies.
This chapter has focused on linear optical effects such
as reflection, refraction, absorption, and interference. These
effects are observed from all light sources, regardless of inten-
sity. Linear optical effects are based on a linear relationship
between an oscillating electric field and an induced dipole
moment in a molecule. Thus, when the oscillating electric
field of light interacts with a molecule, the electron cloud
in the molecule also begins to oscillate. This electron den-
sity oscillation sets up an oscillating dipole moment in the
molecule. The strength of this dipole (µ) depends linearly on
the strength of the incoming electric field (E) according to

Equation 4.28:
μ = αE (4.28)
The constant α is known as the molecular polarizability. It

measures the degree of electron density distortion due to the
electric field. One consequence of this equation is that when
light of frequency ω interacts with a material, the frequency
is unchanged. For example, when incident light of frequency
ω bounces off a surface, the reflected light will also be of fre-
quency ω. Now Equation 4.28 is actually a power series equa-
tion, with quadratic and higher order terms:
μ = αE + βE 2 + γE 3 + … (4.29)
The higher order terms become important only if E is large.

The constants β and γ are known as the first and second hyper-
polarizability, respectively. These constants are essentially
zero when ordinary low-intensity light interacts with a mole-
cule. However, when light from an intense pulsed laser is used,
these nonlinear terms become significant. Equation 4.29 is the
general equation describing the effect of light on a molecule. It
is important to realize that all materials are nonlinear; it is a
matter of the magnitude of the perturbing electric field that is
needed to set in the anharmonicity and to drive it to nonlinear
behavior. Extending this treatment to a bulk material, a simi-
lar equation is obtained, where the dipole (µ) is now replaced
by the average polarization (P) of the bulk material:
P = χ1E + χ2E 2 + χ3E 3 + … (4.30)
The constants χ1, χ2, and χ3 are known as first-, second-, and
third-order susceptibilities. They are related to the sum com-
ponents of the corresponding hyperpolarizabilities averaged
over orientations of the molecules in the bulk material. The
complete mathematical descriptions of the various χ terms
are ignored here. We can separate each polarization term in
Equation 4.30, such that P1 = χ1E represents the linear polar-
ization, and P2 = χ2E2 represents the second-order nonlinear
polarization, etc. In fact, many of the nonlinear optical effects
used to study nanomaterials are based on the second-order
nonlinear polarization. We focus on this term.
Let’s consider two incident intense laser sources with fre-

quencies ω1 and ω2 incident on a material. These two frequen-
cies have two oscillatory electric fields (with fields E1 and E2)
simultaneously acting on the material, given by the equation
E = E1 cos ω1t + E2 cos ω 2t (4.31)
Substituting this equation into the expression for the sec-

ond-order nonlinear polarization, P2 = χ2E2, yields the follow-
ing interesting result:

P2 =
χ2
2
{( E 2
1 ) ( ) (
+ E22 + E12 cos 2ω1t + E22 cos 2ω 2t )
(4.32)
( ) (
+ 2E1E2 cos ω1 + ω 2 t + 2E1E2 cos ω1 − ω 2 t )}
This is an interesting result because of the terms 2ω1, 2ω2,
ω1 + ω2, and ω1 – ω2, which tell us that when two intense light
frequencies (ω1 and ω2) interact with a material, we can pro-
duce resulting light of doubled frequencies (2ω1, 2ω2), and even
resulting light of combined (ω1 + ω2) and subtracted (ω1 – ω2)
frequencies. The doubled frequency production is known as
second-harmonic generation (SHG), and the latter two prod-
ucts are usually referred to as sum-frequency and difference-
frequency generation (SFG and DFG). Two photons can pass
through a material and combine in energy to produce a sin-
gle photon. If we had only one frequency of light, then two
photons would combine to produce a single photon of twice
the energy (and hence frequency) of the incident photon. In
this case, there is no distinction between SHG and SFG.
We are combining two photons to produce one, so the conversion
efficiency of this process has a theoretical upper limit of 50%.
Not all materials generate a second-order polarization, and
those that produce this effect do so at varying efficiency. The
key determining factor is χ2, the second-order susceptibility.
The mathematical properties of χ2 depend on factors such as
molecular orientation, and these properties ultimately deter-
mine the nonlinear optical conversion efficiency. The proper-
ties of χ2 are discussed in the next section.
Example 4.7 Combining Photon Energies
A pulsed laser light source composed of two wavelengths, (a) green

light (500 nm) and (b) infrared light (900 nm), is passed through a
nonlinear optically active material. Determine the color of the second

harmonic and sum-frequency light emitted from the material.
Solution Convert all wavelengths to frequencies:
(a) ν = c = 2.998 × 10 ms = 5.996 × 1014Hz

8 −1
λ 500 × 10 m
− 9
(b) ν = c = 2.998 × 10 ms = 3.331 × 1014Hz

8 −1
λ 900 × 10 m
− 9
Combine the frequencies:
SHG:
(a) 2(5.996 × 1014) = 1.199 × 1015 Hz, or ~250 nm (deep UV light)
(b) 2(3.331 × 1014) = 6.662 × 1015 Hz, or ~450 nm (blue light)
SFG: 5.996 × 1014 + 3.331 × 1014 = 9.327 × 1014 Hz, or ~320 nm
(UV light)
Two factors determine the magnitude of the second-order

susceptibility. The first is symmetry, and the second is the
permanent dipole moment in the molecule (force required to
polarize the molecule). Generally speaking, conjugated mol-
ecules are easier to polarize than non-conjugated molecules
because the electrons are delocalized along the carbon back-
bone. Furthermore, if a conjugated molecule possesses elec-
tron donor and acceptor groups on opposite ends, it makes it
even easier to polarize the molecule. Such molecular configu-
rations increase the nonlinear second-order effect.
χ2 is zero in a centrosymmetric environment. If a molecule
possesses inversion symmetry, χ2 is reduced to zero. Inversion
involves passing each atom in a molecule through the center
of the molecule and placing it on the opposite side of the mol-
ecule. Let’s consider the ethylene and ethylyne. After inver-
sion, the ethylene molecule looks different, but the position of
the hydrogens is the same for ethylyne after inversion. We say
the ethylene is non-centrosymmetric, or lacks inversion sym-
metry. Interestingly, a molecule at an interface always lacks
inversion symmetry, even ethylyne (Figure 4.31). Therefore,
surfaces and interfaces are inherently non-centrosymmetric
regions, and molecules confined to such regions produce
SHG and SFG light. This fact allows SFG and SHG to be used
as surface-specific probes to investigate molecular structure
H H H H H H
C
C C C
(a) Inversion Inversion
C (b)
C C C
C
C
H H
H H H H
H
H
C Inversion C
(c)
C
C
H
H
Figure 4.31 An inversion operation on simple molecules. (a) Ethylene does not have
inversion symmetry because the positions of all the hydrogen atoms are changed after
the operation. (b) All atoms in ethylyne are unchanged after an inversion. This molecule
has inversion symmetry. If ethylyne is on a surface, as in (c), the center of symmetry
is broken. Surfaces represent non-centrosymmetric regions where there is no inversion
symmetry.
at interfaces. Many nanomaterials are supported on surfaces

and can be probed using such nonlinear optical techniques.
4.7.2 Second-Harmonic Generation

SHG involves the conversion of frequency ω to 2ω. The effect
is strongest in non-centrosymmetric regions and in mol-
ecules that are highly polarizable. Consider an amphiphile
containing a highly polarizable headgroup (Figure 4.32).
The polarizable headgroup can be considered as a nonlin-
ear optically active (NLO) chromophore. The NLO-active
chromophore in such molecular building blocks contains
an electron acceptor (A) and an electron donor (D), bridged
by a conjugated azobenzene unit. This D-π-A system is cru-
cial for a strong NLO response. In Figure 4.32, the acceptor
group is a carboxylic acid group and the electron donor is
the oxygen atom. A monolayer comprised of this molecule
produces a highly oriented assembly with a large net dipole.
These conditions meet the requirement of a strong SHG sig-
nal. An additional layer can be added to the assembly either
with the same orientation (asymmetric assembly) or with the
opposite orientation (symmetric assembly). The latter reduces
the SHG signal because of the centrosymmetric nature of the
symmetric bilayer.
Acceptor group COOH
π-conjugated “bridge” N
Donor group O
(CH2)10CH3
Figure 4.32 An amphiphile containing an NLO-active headgroup. This moiety is com-

prised of an electron donor (the oxygen atom) connected to an electron acceptor (the car-
boxylic acid group). The azobenzene group serves as a π-conjugated bridge between the
electron donor and acceptor groups. This headgroup is highly polarizable.
Constructing nanoassemblies that have a large NLO

response is important to a number of technologies, including
optoelectronics (using photons instead of electrons in elec-
tronic devices), communications, and data storage. Materials
with efficient frequency doubling properties can also be
used in lasers to generate light of different wavelengths. Self-
assembly methods can be used to guide molecular building
blocks into preferred orientations that yield new materials
with NLO-active properties. Even though this field is lead-
ing the development of next-generation electronic and com-
putational breakthroughs, fundamental limitations exist in
fabricating functional nanoassemblies for NLO applications.
These limitations arise from the inherent difficulty in obtain-
ing asymmetrically oriented NLO chromophores in layered
thin-film assemblies. As a result, the NLO response does
not increase quadratically with the film thickness, render-
ing the assembly useless. Layer-by-layer methods, covered
in Chapter 5, seem to be the most promising approaches to
creating multilayered films containing NLO blocks through-
out the 3-D assembly. An ideal NLO-active film would con-
tain hundreds of discrete layers of highly oriented molecules
self-organized on a solid support. In the case of NLO materi-
als, if electron donor-acceptor chromophores were organized
in such a way that the dipole moments were oriented in the

same direction, then a quadratic increase of the optical sec-
ond-harmonic signal would be observed with increasing film
thickness (number of bilayers) as predicted by Equation 4.33,
where L is the film thickness, Iω is the intensity of incident
light, n ω is the film’s refractive index at the incident light
frequency, and n2ω is the refractive index at the doubled
frequency.
2ω 2deff
2 2
L sin2( ΔkL /2)
I 2ω = (I ω )2 (4.33)
c 3ε o nω2 n2ω ( ΔkL /2)2
The Langmuir-Blodgett (LB) method is a commonly used

technique to organize amphiphilic molecules into densely
packed 2-D structures at the air-water interface. The 2-D
structure can then be conveniently transferred to a substrate.
This method is discussed further in Chapter 5. By repeating
the transfer process, ordered 3-D nanostructures of desired
thickness can be obtained. If asymmetric structural order
is retained throughout the assembly, one can then achieve
a sum effect of the functionality represented by an individ-
ual layer. However, the idea of relying on the sum effect of
individual LB layers to obtain an enhanced NLO functional-
ity in 3-D materials is seriously challenged by studies that
demonstrate a certain level of intermixing between layers.
Intermixing randomizes the chromophore orientation, result-
ing in a reduction of the SHG intensity. We are interested in
an asymmetric structure with a large optical nonlinearity.
This type of deposition is usually accomplished using two
different species (discussed further in Chapter 5). Disorder in
LB assembled materials is due to the presence of kinked alkyl
tails within the assembly, resulting in a decreasing packing
density. Furthermore, disorder may also arise from a lack of
in-plane ordering due to unfavorable dipole-dipole interac-
tion. In a typical LB assembly involving two different spe-
cies, both materials from a given deposition cycle may occupy
defects in underlying layers. Materials of either species have
a tendency to spread in a direction perpendicular to the sub-
strate, leading to enhanced interpenetration, despite the main-
tenance of equivalent bilayer thicknesses over many layers. In
other words, equivalence in deposition amount and thickness
over many layers may indicate similar net surface conditions
in each layer, but care must be taken in using the notion of a
similar surface to infer consistency in molecular order within

layers. Even disordered systems can display seemingly regu-
lar deposition patterns—a film with disorder on the molec-
ular level may appear markedly ordered at the macroscale.
Such considerations are far reaching in a field largely focused
on not only the order, but also the chromophore orientation at
the molecular level.
In addition to measuring the NLO response from a nano-
material, SHG can also be used to follow the assembly process
in real time, obtain the density of NLO-active chromophores
within the assembly, and even image ordered domains in
nanofilms. By obtaining the relative amount of SHG generated
from s- and p-polarized light, the orientation of the NLO-active
chromophore with respect to the surface normal can also be
calculated.
4.7.3 Sum-Frequency Generation Spectroscopy

Photons of two different frequencies (ω1 and ω2) can com-
bine to generate light at the sum frequency (ωSFG = ω1 + ω2).
Similar to SHG, SFG is produced from non-centrosymmetric
environments. SFG is routinely used as an IR spectroscopic
method to obtain molecular structure information of nano-
films confined at interfaces. In this technique one of the two
frequencies is fixed at a visible wavelength, typically 532 nm
(ωvis). The other frequency is variable in the IR region (ωIR).
The two light beams are overlapped on a surface and the
intensity of the SFG beam is measured. It turns out that this
intensity is greatly enhanced when ωIR is in resonance with
a vibrational mode of the molecule at the surface. Thus, a
plot of the SHG intensity versus ωIR provides a vibrational
spectrum of molecules on the surface. It is worth noting that
in contrast to ATR-FTIR spectroscopy, SFG spectroscopy is
surface specific, whereas ATR methods probe regions within
as much as a few microns from the surface. A theoretical
expression for the magnitude of the SFG signal is given by
Equation 4.34:
⎛ω ⎞ 2 2⎛ I I ⎞
I SFG = 128π 3 ⎜ SFG K SFG K vis K IR χ2 ⎜ vis IR ⎟ (4.34)
⎝ hc 3 ⎟⎠ ⎝ AT ⎠
The signal intensity, ISFG, is the number of SFG photons pro-

duced per laser pulse and depends on the beam intensities (Ivis
and IIR), the laser pulse length (T), the area of overlap of the
Infrared laser
(variable frequency)
Visible laser
(fixed frequency) Sum-frequency
emission
Fluid in Fluid out
Substrate coated with nanofilm

ωIR
θI ωSFG
ωVIS
R θSFG
θVIS
Figure 4.33 Typical geometry of an SFG experiment in which a prism is used to couple
the various laser beams onto a surface. The various angles of the IR, visible, and SFG beams
are indicated in the circle.
beams (A), the various geometric Fresnel factors (K), and the
second-order susceptibility.
In practice, in order to generate SFG light, precise incidence
angles of the fixed frequency visible and tunable IR light have
to be used. Furthermore, the emitted SFG light is observed
at a precise angle. This is known as phase matching, and the
appropriate angles can be calculated using Equation 4.35.
Figure 4.33 shows the typical geometry of an SFG experiment,
indicating the various angles shown in Equation 4.35.
ω SFG sin θSFG = ω vis sin θvis − ω IR sin θIR (4.35)
SFG spectroscopy has been used extensively to study the

conformation of alkyl chains in monolayers assembled at
interfaces. If the alkyl chain is in an all-trans conformation,
then all the CH2 groups are in a locally centrosymmetric
environment, as illustrated in Figure 4.34. The presence of
a gauche defect breaks this centrosymmetry and renders the
H H
H H
C C
H H
Inversion
C C
H H
C H C H
H H
C C
H H
H H
Figure 4.34 Portion of an all-trans alkyl chain. The CH2 are locally centrosymmet-
ric because after an inversion operation, the positions of all atoms remain unchanged.
Therefore, these groups in this conformation will not produce SHG or SFG light. A “kink” in
the chain or a gauche defect will create a non-centrosymmetric environment of CH2 groups,
rendering these groups SFG active.
CH2 sum-frequency active. The power of SFG spectroscopy

is illustrated by the following example. Consider a mono-
layer of dodecane thiol (DDT) chemisorbed to a gold surface
(Figure 4.35) and under water. Details of this type of mono-
layer are discussed further in Chapter 5. This monolayer is
tightly packed with all the CH2 groups in an all-trans con-
formation. Only the terminal CH3 groups are non-centrosym-
metric. The SFG spectrum of DDT shows only features that
are attributed to the vibrational modes of the terminal methyl
group. In comparison, Figure 4.36 shows the SFG spectrum of
a mercaptododecanoic acid (MDA). This molecule is structur-
ally similar to DDT, except that it contains a large carboxylic
headgroup instead of a CH3 group. The SFG spectrum of MDA
shows no CH3 features, but does show prominent vibrational
features that arise from CH2 groups. The presence of the meth-
ylene features indicates a significant amount of gauche defects
in the alkyl chains of the MDA monolayer, which is not sur-
prising since, compared to DDT, MDA has a bulky headgroup
that prevents the monolayer from packing tightly. This charac-
teristic leads to a greater amount of space for the alkyl chain to
occupy, resulting in the formation of kinks in the hydrocarbon
chain.
SFG spectroscopy has established itself as a powerful nano-
characterization method. The technique can probe any inter-
facial medium that is optically accessible, and has been used
to probe many interfacial processes such as corrosion, sur-
face phase transitions, detergency, and the structure of cell
membranes.
Water Absence of CH2 peaks
Methyl
groups Strong CH3 peaks
pointing “up”
1.6
All trans 1.4

alkyl chain
1.2
ISFG (arbitrary units)

Strong
S-Au bonds 0.8
Gold substrate 2800 2850 2900 2950 3000
Wavenumber (cm–1)
(a) (b)
Figure 4.35 (a) A self-assembled dodecane thiol monolayer at the gold-water interface. An all-trans conformation due to strong inter-alkyl chain hydrophobic
interactions creates a close-packed structure with the terminal CH3 groups pointing toward the water phase. The SFG spectrum in (b) shows only features due to the non-
centrosymmetric CH3 groups. The CH2 groups are centrosymmetric and so do not appear in the SFG spectrum.
165
166
2.6
Water 2.4
Carboxyl groups
2.2
randomly oriented
2
1.8
alkyl chain
containing 1.6
“defects”
ISFG (arbitrary units)

1.4
Strong CH2 peaks
1.2
Strong
S-Au bonds 1 Absence of CH3 peaks
Gold substrate
0.8
2800 2850 2900 2950 3000
(a)
Wavenumber (cm–1)
(b)
Figure 4.36 (a) A self-assembled mercaptododecanoic acid monolayer at the gold-water interface. The large anionic headgroup prevents the monolayer from packing
tightly and creates kinks and gauche defects in the chain. This randomized the headgroups. The SFG spectrum in (b) shows only features due to the non-centrosymmetric
CH2 groups.
4.8 X-RAY SPECTROSCOPY

X-ray spectroscopy is a collection of techniques that use x-ray
excitation in order to probe the electronic structure of mol-
ecules in a material. These techniques are particularly useful
methods for determining a material’s composition, probing the
ordering of individual atoms in a crystal, or observing nano-
scale molecular processes. Each technique is based on the
measurement of absorption, emission, or scattering of electro-
magnetic radiation caused by x-rays.
4.8.1 Absorption
When x-rays pass through a thin layer of matter, the inten-
sity of the x-rays is diminished as a result of absorption and
scattering. The effect of scattering can be ignored in wave-
length regions at which significant absorption occurs. Each
element has its own absorption spectrum with well-defined
x-ray absorption peaks that can be used to identify it. In the
x-ray region of light, there is enough energy to cause observ-
able changes in the electronic state of a molecule. When x-ray
energy corresponding to the binding energy of core elec-
trons is absorbed, the core electron is ejected from the atom.
This results in an excited ion. There is a higher probability
of this happening when an x-ray beam with energy equal to
the binding energy of the core electron is used. As the x-ray
beam increases in energy, away from the core electron binding
energy, the probability of the corresponding wavelength being
absorbed diminishes, and so the amount of the x-ray beam
absorbed will decrease. If the x-ray beam is too low in energy,
the corresponding wavelength will not be appreciably present
and will not be able to eject the core electron. This will cause
an abrupt decrease in the amount of the x-ray beam absorbed.
The wavelength of x-rays ranges from about 10 to 10 –6 nm, but
conventional x-ray spectroscopy generally uses only wave-
lengths between 2.5 and 0.001 nm because this range contains
the x-rays with energies corresponding to core electron-bind-
ing energies, which differ between elements.
4.8.2 Fluorescence
The excited ion, which is a result of the x-rays’ ejecting of
a core electron, will fluoresce through transitions of elec-
trons in higher energy levels to the vacancy left by the
ejected core electron. These transitions allow the excited ion
to return to its more stable ground state. This fluorescence
is measurable and is often used in nanoscience techniques
such as fluorescence recovery after photobleaching (FRAP),

fluorescence resonant energy transfer (FRET), and fluores-
cence interference contrast microscopy (FLIC), which are
covered in the next chapter. X-ray fluorescence is also com-
monly used in dyes. These fluorescing dyes can be attached to
target molecules of interest that can be excited and observed
in conditions where dyes in the visual wavelengths of light
are not usable. This is often the case for in vivo studies when
there is a layer between the observer and the dyed molecule
that allows for x-rays to pass through but does not allow vis-
ible light to pass through.
4.8.3 Diffraction
When x-rays pass through matter, the radiation interacts
with electrons in the matter in such a way that the path of
the x-rays can be altered. This scattering effect of matter on
x-rays is known as diffraction. In a crystal, or any ordered
sample, the x-rays scatter in ways that produce higher-intensity
areas and lower-intensity areas, also termed constructive and
destructive interference, respectively. By comparing these
high- and low-intensity areas, it is possible to determine the
architecture and the ordering in matter at the nanoscale. It
is, of course, slightly more complicated in practice, because
when an x-ray beam strikes an ordered crystal, which can be
thought of as multiple layers of atoms, each subsequent layer
of atoms scatters some of the beam and lets the remainder of
the beam through. So that this diffraction can take place in a
way that allows us to relate it to the materials structure, the
space between layers of atoms in the material must be about
the same distance as the wavelength of the radiation used to
probe it, and the atoms comprising the system must be highly
ordered with few defects.
In order to use the high- and low-intensity areas that result
from shining x-rays through an ordered crystal, a way to relate
them to the structure must be known. Fortunately, in 1912, W.
L. Bragg determined this relationship. The Bragg equation for
constructive interference is as follows:
2 d sin θ = n λ (4.36)
In this equation d is the lattice spacing, or the distance

between atoms in the subsequent layers. θ is the incident angle
between the beam of x-rays and the plane of the atom layer. λ
is the wavelength of the x-rays. Lastly, n is an integer related to
the order of the reflection.
4.9 IMAGING NANOSTRUCTURES

One of the more straightforward methods that we have not
yet discussed is taking a “picture” of the nanostructure of
interest. Indeed, so-called nano-imaging methods are often
among the first used to study a specific nanomaterial. There
are a variety of nano-imaging methods that exist. In the next
few sections, we discuss a sampling of these methods, ranging
from sub-diffraction limit optical methods such as near-field
scanning microscopy, to other contrast mechanisms with the
capability of resolving individual atoms such as the scanning
probe microscopies.
4.9.1 Imaging Ellipsometry

4.9.1.1 Imaging Using Conventional Ellipsometry
We have already encountered ellipsometry as a powerful
method for determining the thickness of films as small as a few
angstroms (for a review of ellipsometry, see Section 4.3). This
method involves analyzing the change in the polarization state
of a light beam upon reflection off a planar surface. The ellip-
sometric parameters recorded from monitoring this change
in the polarization state of the light beam can be translated
into thickness values for the surface after using an appropri-
ate mathematical model. It should also be highlighted that a
thickness value obtained in this manner actually represents
the average thickness value within the beam spot on the sur-
face. In order to achieve a better lateral (or horizontal) resolu-
tion, the light beam must be focused on a smaller spot on the
surface. Thus, perhaps the most straightforward approach to
image a surface using ellipsometry would be to maximize the
lateral resolution by using a tightly focused light beam then
determining the film thickness as the beam scans the surface.
We could then obtain a three-dimensional image of the sur-
face topography. However, using conventional ellipsometry to
scan the sample in this way can be time consuming, so quicker
methods called imaging ellipsometry have been developed.
4.9.1.2 Principles of Modern Imaging Ellipsometry

Modern imaging ellipsometry can be thought of as combin-
ing an ellipsometer and a microscope. Rather than using a
tightly focused light beam and scanning it across the sample,
modern imaging ellipsometers generally employ a light beam
with a large diameter (often on the order of millimeters) such
that the entire sample is illuminated. The reflected image of
the entire object is then focused onto a high-resolution CCD
camera using an objective lens. From this image, ellipsometric

data of the sample (including thickness) can be extracted pixel
by pixel, generating a surface topography map of the sample.
Therefore, the ellipsometric data for the entire sample can be
determined quickly, resulting in a 3-D image of the sample.
4.9.1.3 Methods for Extracting Ellipsometric

Data in Imaging Ellipsometry
The “pixel-by-pixel” ellipsometric data from an imaging ellip-
someter setup can be determined using one of several methods.
A popular method is to employ “off-null” mode analysis. We
recall from our discussion of ellipsometry in Section 4.4 that
conventional ellipsometers are often operated in null mode. In
null mode, the change in the polarization state of the reflected
light is determined by changing the polarizer, compensator,
and analyzer so that a null condition is achieved (or so that no
light passes through the analyzer). Depending on the values
of the polarizer, compensator, and analyzer used to achieve
this null condition, different ellipsometric parameters may
be calculated. In “off-null” mode, the imaging ellipsometer is
“zeroed” by determining the null conditions for the bare sub-
strate. These conditions are then kept constant as the entire
sample is imaged. Since the sample is generally not of the same
thickness and refractive index as the substrate, light of greater
and greater intensity passes through the analyzer for thicker
regions of the sample. Therefore, the intensity of the light in
each pixel can be related to the thickness of the sample at that
point. It should be noted, however, that this type of off-null
mode analysis is substrate, sample, and thickness dependent,
so comparisons with a reference material of known thickness
are often recommended. For example, for a pure silicon sub-
strate and a biological sample, the intensity I under an off-null
condition has been reported to be related to the film thickness
d by I = kd2 where k is a proportionality constant, and this
equation is valid with a deviation of approximately 2% up to
d ~ 5nm.
An alternative and perhaps more straightforward method
for obtaining the “pixel-by-pixel” ellipsometric parameters
is to continually adjust the polarizer, compensator, and ana-
lyzer in order to determine the null condition for each pixel
(or group of pixels) and then extract the parameters from those
null conditions as would be done in conventional ellipsometry.
While this approach is slightly more time-consuming, it does
not require the use of any reference samples or the intensity-
thickness assumptions employed in “off-null” mode analysis.
4.9.1.4 Image Focusing

One of the major obstacles in imaging ellipsometry is that only
the center of the image is in focus, due to the large incident
angle of the light beam. To compensate for this limited focus
area, a scanner can be incorporated into imaging ellipsometer
instruments. The image may then be taken at several different
positions, and the computer software combines these images
into one focused image of the entire sample. An alternative
approach has been to use the Scheimpflug method, which
is an optical technique that allows the entire sample to be
placed in focus by tilting the lens to a certain degree, called
the Scheimpflug line.
4.9.1.5 Resolution of an Imaging Ellipsometer

The lateral (x,y) resolution for an imaging ellipsometer as
described above is generally limited by the resolving power of
the CCD camera and is usually on the order of a few microns.
The depth (z) resolution, on the other hand, is comparable
to conventional ellipsometry at a few tenths of a nanome-
ter. Because of its nondestructive nature, compatibility with
a wide variety of sample types, and excellent temporal and
depth resolution, imaging ellipsometry is a powerful tech-
nique to image nanostructures, and is suited for thin film
analysis. Figure 4.37 shows an ellipsometric thickness image
of a single phospholipid bilayer assembled on a glass surface.
6 10
5
5
4
(nm)
0
3
400
2
300 200
200
1 100
(um) 100 (um)
0 0
nm
Figure 4.37 An ellipsometric thickness image showing the height of a phospholipid

bilayer assembled on a solid support. (Image provided by Professor Atul Parikh, University
of California, Davis.)
The bilayer thickness is around 5 nm. The figure shows how

imaging ellipsometry can be used to follow the slow hydration-
induced spreading of the bilayer over the uncovered portion of
the substrate. Thus, imaging ellipsometry provides topograph-
ical data with excellent vertical resolution. However, because
the lateral resolution is on the order of microns, nanostruc-
tures cannot be resolved in the x-y plane. In order to image
nanostructures at atomic-level resolutions in all three dimen-
sions, methods such as the scanning probe microscopies must
be used.
4.9.2 Scanning Probe Methods

The scanning probe microscopies are powerful imaging tech-
niques in which a very sharp tip is scanned across a sur-
face, producing an image with near atomic-level resolution.
Scanning tunneling microscopy (STM) and atomic force
microscopy (AFM) are the most common examples of scanning
probe methods. Both STM and AFM produce three-dimen-
sional images of surfaces that approach atomic or molecular
resolution, making them ideal for the study of nanostructures
on surfaces.
4.9.2.1 Scanning Tunneling Microscopy

STM operates by monitoring the “tunneling” current that is
produced when a sharp tip is brought extremely close to a sur-
face that is able to conduct electricity. In order to bring and
maintain the tip so close to the surface, a piezoelectric trans-
ducer is used. As a refresher, we recall that piezoelectric trans-
ducers possess the ability to physically expand or contract in
response to an applied voltage (for a more complete discussion,
see Section 4.3). Therefore, the sharp metal STM tip, usually
made of platinum, is attached to a piezoelectric scanner that
can cause the tip to move small distances in the x, y, and z
directions. In fact, depending on the type of piezoelectric mate-
rial used to make the scanner (usually some form of ceramic)
and its dimensions, the contraction or expansion of the mate-
rial can be as small as 1 nm per volt applied, allowing for the
metal tip to be brought close to the surface being imaged. In
early models of STMs, the piezoelectric scanners consisted of
transducers arranged in the x, y, and z directions, as shown in
Figure 4.38, but more recent models have made use of a tube-
like piezoelectric transducer to achieve better resolution.
Using a piezoelectric scanner, the STM tip can be brought
close (within 1 nm) to a conducting surface. A small voltage,
generally between 2 mV and 2 V, is then applied between the
Piezoelectric X
transducers
Y
STM tip
Substrate
Figure 4.38 Schematic diagram of the piezoelectric transducers that control the posi-
tion of an STM tip near the sample surface. For many piezoelectric transducers used, a
distance of as little as 1 nm can be affected with a single applied volt. When the STM tip is
positioned sufficiently close to the sample surface and a potential is applied across them
both, a tunneling current will be induced between the tip and the sample. This tunneling
current is the basis of the STM measurement.
conducting substrate and the metal tip, causing electrons to

tunnel between the tip and the surface, creating a current. The
magnitude of this tunneling current depends exponentially on
the distance (h) between the conducting surface and the STM
tip. For a constant applied voltage, the tunneling current, I, is
given approximately by
I ≈ e −2κh (4.37)
Equation 4.37 tells us that the tunneling current decreases

exponentially as we move the tip away from the surface. The
constant κ is called the electronic decay length of the elec-
tron and is a measure of how the probability density of a con-
fined electron decays with distance or, more precisely, it is
the decay length of the electronic wavefunction. Not surpris-
ingly, κ depends on how tightly the electron is bound to the
conducting surface, called the work function of the surface.
The ability of electrons to travel or “tunnel” through a non-
conducting medium between two conducting materials that
are close together is strictly a quantum mechanical effect. The
details of this effect are not covered here. Suffice it to say that
electrons can tunnel between two conducting materials, and
this tunneling ability is exponentially dependent on the dis-
tance between the two materials.
As the STM tip is scanned across a rough surface, the

t unneling current changes as the tip encounters bumps or
dips in the surface (or as h changes). There are two obvious
methods for imaging the surface. In the first method, known
as “constant-height” STM, the tip is held at a constant verti-
cal position and the tunneling current is plotted as a function
of the x-y region scanned. Because the tunneling current is
dependent on the distance between the tip and the surface
(h), the surface can be imaged. The second method, known as
“constant-current” STM, is to maintain a constant tunneling
current as the tip is scanned across the surface by varying
the tip’s vertical position. The piezoelectric transducer moves
the tip up or down as the surface is scanned, ensuring that the
tunneling current is maintained at the predetermined value.
Essentially, h remains constant throughout the scanning pro-
cess, so the instrumentation has to monitor only the changes
in the tip’s vertical position to provide a topographical image
of the surface. In practice, this second method is generally
used to generate a surface image.
As with most of the scanning probe microscopies, STM sur-
face images are typically generated by scanning the tip across
the surface in a raster pattern, or line by line, to create a rectan-
gular grid. In some STM models, the scanning process is con-
trolled by moving the substrate and not the tip, which is moved
only vertically to maintain a constant tunneling current.
Example 4.8 Current Changes in STM
Consider a metal surface with an electronic decay length κ of

10 nm –1. By how much does the tunneling current change when
the height h between the surface and the STM tip increases from
1.0 nm to 1.1 nm?
Solution From Equation 4.37, we have I ≈ exp (−2κh). We can deter-
mine the approximate factor by which the tunneling current changes
by taking the following ratio:
I1.0nm exp(−2 × 10 nm−1 × 1.0 nm)

≈ ≈8
I1.1nm exp(−2 × 10 nm−1 × 1.1 nm)
Thus, we see that the current changes by almost an order of mag-

nitude when the distance h varies by 0.1 nm.
Example 4.8 emphasizes the sensitivity of STM to changes

in height. In fact, if the tunneling current is kept constant then
the height remains constant to within 10 –3 nm, meaning that

atomic-level resolution is achieved.
While such resolutions constitute a major achievement,
STM suffers from its inability to image surfaces that do not
conduct electricity. Non-conducting surfaces are unable to be
imaged unless they are modified or coated with a conducting
substance. Atomic force microscopy, discussed below, does
not suffer from this disadvantage, although STM provides
superior resolution.
For a typical STM, the lateral range of the scanner is gen-
erally from tens of angstroms to approximately 100 µm. The
allowable height is from the sub-angstrom to ~10 µm range.
The tip is generally constructed by manually cutting a plati-
num wire or by electrochemical etching of tungsten metal.
One example of an application of STM is the atomic resolu-
tion STM image of a graphite surface, shown in Figure 4.39.
In this image, the hexagonal arrangement of the carbon atoms
A
B
Cross section:
0.8
0.6
z/Å
0.4
0.2
A B
0
0 5.5 11 16.5 22
x/Å
Figure 4.39 STM image of a graphite surface. Notice the angstrom-level resolution
that is characteristic of STM. (Image from Atamny et al, Phys. Chem. Chem. Phys., 1999, 1,
4113–4118. Reproduced by permission of the PCCP Owner Societies.)
in graphite is clearly discernible, as well as the valleys in

between adjacent carbon atoms, thus demonstrating the util-
ity of STM in studying conducting surfaces.
4.9.2.2 Atomic Force Microscopy

Atomic force microscopy, or AFM, operates on principles simi-
lar to those of STM. As with STM, AFM functions by scanning
a sharp tip across a surface to generate an image. However,
rather than monitoring the tunneling current between the tip
and the surface, AFM monitors the height of the tip as it physi-
cally interacts with the surface at a constant force.
The setup of a typical AFM is shown in Figure 4.40. A sharp
tip, often made of diamond or silicon nitride, is attached to a
cantilever spring and placed in physical contact with the sub-
strate at constant force. The position of the tip is monitored
by reflecting a laser beam off the back of the cantilever and
monitoring the deflection of this beam as the tip is scanned
across the surface. If the tip encounters a bump or a dip, the
reflection of the laser beam is slightly altered and monitored
by the detector. Rather than moving the tip to scan the sur-
face, the substrate is typically mounted on a piezoelectric tube
scanner, allowing for the surface to be moved in a raster pat-
tern under the tip. The cantilever itself is tens of microns in
length, <10 microns in width, and ~1 micron in thickness. The
tip is usually cone or pyramid shaped, with a height of several
microns and a base width of several microns.
Laser
Detector
Cantilever
AFT tip
Substrate
Piezoelectric scanning stage
Figure 4.40 Schematic diagram of an atomic force microscope. The AFM tip is held in
constant force against the sample surface by the cantilever. A laser beam is reflected off the
back of the cantilever to monitor the tip height. The sample is scanned underneath the AFM
tip by a piezoelectric stage.
AFM is able to image both non-conducting and conducting

surfaces, unlike STM. However, in AFM, the physical con-
tact between the tip and the surface can damage the surface,
resulting in a distorted image. This drawback is particularly
problematic if one wishes to image a “soft” surface such as
a biological membrane or a surfactant film. Fortunately, this
problem of surface destruction can be somewhat overcome
by scanning the substrate in tapping mode. In tapping-mode
AFM, the tip is “tapped” on the surface, placing it in contact
with the surface only for a short amount of time. This “tap-
ping” is usually accomplished by oscillating the cantilever
with constant driving force and ensuring that the cantilever is
positioned so that the tip touches the surface only at the bot-
tom of the oscillation. Typical frequencies of oscillation are on
the order of a few hundred kilohertz. By using tapping-mode
AFM, surfaces that would normally be destroyed are able to
be imaged properly.
AFM can be used under water as well as in air, allowing for
imaging at the surface-liquid interface. This ability is particu-
larly important for biological samples, which might be distorted
at the surface-air boundary. One application of AFM in surface
imaging is the detection of different functional groups on a
surface, called chemical force microscopy (CFM). In CFM, the
AFM tip is chemically functionalized, often by coating the tip
with gold then functionalizing it with a self-assembled mono-
layer using thiol-gold chemistry (Figure 4.41). For example, a
gold-coated AFM tip might be functionalized with 11-mercap-
toundecanoic acid, effectively resulting in a –COOH coated
AFM tip [R is a –(CH2)nCOOH group in Figure 4.41]. This type
of chemisorption is discussed in Chapter 5. This chemically
AFM tip
Gold Thiol
RS SR
Deposition (H-SR) SR
RS
SR
RS
RS SR
SR
Figure 4.41 A method for the creation of a CFM tip. An AFM tip is coated with Au and
then is functionalized using thiol-gold chemistry. The functional group R can then be used
to probe the surface.
functionalized AFM tip can then be scanned across a surface

composed of different functional groups. Depending on the
interaction between the surface-bound functional group and
the functional group on the AFM tip, a different “height” is
reported. A –COOH coated tip would be expected to experi-
ence more frictional force from surface-bound –COOH groups
as compared to –CH3 groups, and therefore the –COOH groups
would appear to be “taller” than the –CH3 groups. For exam-
ple, a –COOH-coated AFM tip can be used to scan across a
surface containing both –CH3 and –COOH groups. A different
image is obtained when the same surface is scanned with a
–CH3-coated AFM tip. The results are striking and highlight
the utility of CFM to detect different functional groups on a
surface.
Aside from being used to image surfaces and surface-
bound material, AFM has other applications. One unorthodox
application of AFM in a non-imaging setting is in the recent
development of a cell nanoinjector. In the past, in order to
introduce a substance into a living cell, the cell had to be
permeabilized with an electric current or chemical agent, or
a bulky micropipette had to be used as an injector. This pro-
cess could result in damage to the cell membrane, inducing
undesirable side effects. However, scientists have recently
been able to attach a delivery moiety (such as a carbon nano-
tube) to the end of an AFM tip for use as a “nanoinjector.”
Because of the tip’s small dimensions, the cell membrane is
not perturbed.
The substance (or cargo) to be injected into a cell is chemi-
cally attached to the delivery moiety (through a disulfide link-
age, for instance). A carbon nanotube as a delivery moiety acts
like a “nanoneedle” when it is attached to the AFM tip, which
is then used to inject the carbon nanotube (with the molec-
ular cargo) through the cell membrane. Inside the reducing
environment of the cell interior, the disulfide linkage between
the cargo and the carbon nanotube is severed and the cargo is
released into the cell. The AFM tip then withdraws the nano
needle from the cell interior. Because the nanoneedle is of such
small dimensions, the cell membrane is not perturbed greatly
and the cell remains unharmed. Furthermore, by using the
AFM machinery, excellent spatial precision in the placement
of the cargo can be achieved. Figure 4.42 shows a schematic
diagram of a nanoinjector and Figure 4.43 shows electron
microscope images of the nanoneedle before and after the
attachment of molecular cargo. Although the AFM-operated
nanoinjector is admittedly limited to delivering cargo to one
AFM tip
Delivery
moiety
Molecular
cargo
Cell
membrane
Figure 4.42 Schematic steps of the AFM-operated nanoinjector. The nanoinjector is

linked to the molecular cargo through a disulfide bond and then poised above the cell mem-
brane. The nanoinjector is then inserted through the cell membrane and the molecular cargo
is released inside the cell as the disulfide linkages are reduced. Finally the nanoinjector is
withdrawn from the cell, ideally causing little to no membrane damage.
cell at a time, it still promises to be a useful tool in a wide

variety of biological studies.
4.9.3 Transmission Electron Microscopy

Unlike the scanning probe microscopies, which utilize physi-
cal interactions between a sharp tip and a surface to create an
image, the electron microscopies use electrons rather than a
beam of light to visualize a sample. Because electrons have
much smaller wavelengths than visible light, electron micros-
copies are able to achieve resolutions far greater than optical
microscopies.
Scanning electron microscopy (SEM) is a commonly used
electron microscopy method that functions by monitoring the
electrons that are backscattered after interacting with a sam-
ple and uses those backscattered electrons to reconstruct an
image of the sample. On the other hand, transmission electron
microscopy (TEM) makes use of electrons that are transmitted
through a thin sample to produce an image. For our current dis-
cussion, we focus on TEM as a tool to image nanomaterials.
4.9.3.1 Principles of TEM

The resolution of an image produced by an optical microscope
is limited by the wavelength of the radiation being used. If
180
B C
1 µm
20 nm 20 nm
Figure 4.43 Imaging of the nanoneedle before and after the molecular cargo is attached. (a) shows an SEM image of the carbon nanotube affixed to the AFM tip.
(b) is a TEM image of the carbon nanotube shown in (a). (c) shows a TEM image of the nanoneedle after the molecular cargo has been attached (a quantum dot in this case).
(Image taken from Chen et al, 2007, PNAS, 104, 8218–8222. Copyright 2007 National Academy of Sciences, USA. With permission.)
we use the Abbe equation applied to the resolution of a light

microscope, then the maximum resolution δ is approximately
given as
⎛ λ ⎞
δ ≈ 0.61 ⎜ ⎟⎠ (4.38)
⎝ n sin(β )
where λ is the wavelength of the light being used, n is the

refractive index of the viewing medium, and β is a prop-
erty of the magnifying lens called the semi-angle of collec-
tion. Together, n sin(β) is often called the numerical aperture
(NA) of the objective. To provide a rough estimate of δ, let’s
assume n sin(β) = 1–1.5. We see that the resolution of a light
microscope is approximately 50–60% of the wavelength of the
light being used. Visible light has wavelengths in the range of
~350–750 nm, so conventional optical microscopies are unable
to resolve objects that are smaller than a few hundred nanome-
ters. Since most nanomaterials of interest possess structures
that are much smaller than several hundred nanometers, opti-
cal microscopies are only moderately useful in the imaging
of nanomaterials. In order to visualize these nanomaterials,
something with a much smaller wavelength than visible light
must be used.
Electrons, as all small particles, can be thought of as being
both particles and waves (with characteristic wavelengths).
Indeed, one of the great advances in early quantum mechanics
was the development of the concept of wave-particle duality—
that matter and light can be treated simultaneously as both
particles and waves. De Broglie’s famous equation relates the
wavelength λ of a particle to its momentum p by
h
λ=
p (4.39)
where h is Planck’s constant, equal to 6.626 × 10 –34 J s.

Therefore, in order to calculate the wavelength of an electron
(which gives us an idea of the maximum possible resolution
of an electron microscope), we must be able to calculate its
momentum. This goal can be achieved by understanding the
basic physics of a TEM.
In an electron microscope, an electron produced in the elec-
tron gun is accelerated by an electric potential V toward the
sample to be studied. This acceleration imparts a kinetic energy
K to the electron that is equal to the accelerating potential or

voltage. Therefore, using classical physics, we can write
1
V =K= mev 2 (4.40)
2
where me is the rest mass of the electron (9.11 × 10 –31 kg) and v
is the electron’s velocity. From classical physics we also know
that p = mv, so using Equation 4.40 we can write
⎛1 ⎞
p = mev = 2me ⎜ mev 2 ⎟ = 2meV (4.41)
⎝2 ⎠
This equation, then, shows that we can calculate the momen-

tum of an electron if we know the accelerating voltage used in
the TEM. Plugging this result into de Broglie’s relationship,
we get
h h
λ= = (4.42)
p 2meV
Therefore, we can calculate the wavelength of an elec-

tron from its accelerating voltage in an electron microscope.
Furthermore, we see that as we increase the accelerating volt-
age, we decrease the wavelength of the electron. Ignoring other
effects, an electron microscope can achieve better resolution
by accelerating the electrons to higher energies. However, as
the energy of the electrons increases, so does the likelihood
that it can destroy or damage the sample being studied. This is
one limitation to the maximum resolution that can be achieved
in a TEM.
The magnitude of accelerating voltages typically used
in most TEMs (on the order of hundreds of keV) accelerates
the electron so much (near the speed of light) that relativistic
effects must be accounted for. To account for these relativistic
effects, Equation 4.42 becomes
h
λ= 1
⎡ 2
⎛ V ⎞⎤
⎢2meV ⎜⎝ 1 + 2m c 2 ⎟⎠ ⎥ (4.43)
⎢⎣ e ⎥⎦
where c is the speed of light in a vacuum. Using this equation

we can calculate that an electron that has been accelerated to
100 keV has a relativistic wavelength of ~3.7 pm. For a TEM,
we can approximate the maximum resolution by modifying
Equation 4.43 to
⎛ λ⎞
δ ≈ 0.61 ⎜ ⎟ (4.44)
⎝ β⎠
where β in this case is a property of the electron “lens” used to

focus the electron beam. Therefore, using Equation 4.44 we see
that with a 100-keV electron beam (λ ~ 4 pm), we can achieve a
theoretical maximum resolution (δ ~ several picometers) that
is smaller than the diameter of an atom. While such maximum
resolution is not feasible because we are unable to build per-
fect electron lenses, it highlights the fact that TEM is able to
achieve a high level of resolution.
4.9.3.2 TEM Instrumentation

In a simplified way, a transmission electron microscope oper-
ates like a slide projector. In a slide projector a beam of light
is transmitted through a slide. Some of the light is reflected
or absorbed by the slide, so when the transmitted light is pro-
jected onto a screen, an image is produced. A TEM functions
in essentially the same manner, with the only difference being
that a beam of electrons is transmitted through the sample
rather than a beam of visible light.
As seen in Figure 4.44, an electron gun produces a beam
of electrons, often by heating a metal (most commonly tung-
sten) to such high temperatures that electrons are ejected
from its surface. These ejected electrons are condensed and
then focused by electromagnetic lenses, which are analogous
to optical lenses, but function on different principles. The
focused beam of electrons is shone on the sample, and those
electrons that have been transmitted are gathered by another
lens and then shone onto a fluorescent screen or other detector
to produce an image of the sample. The setup of a TEM can be
modified so that it is able to scan an entire sample, and this
technique is called scanning transmission electron micros-
copy (STEM). As with AFM and STM, STEM typically scans a
given sample in a raster pattern to produce an image.
In order to prevent deflection of the beam of electrons due to
interactions with gas molecules inside the machine, the inte-
rior of a TEM must be operated under high vacuum conditions.
Electron gun
Condenser
Electromagnetic lens
Sample
Electromagnetic lens
Fluorescent viewing screen
Figure 4.44 Schematic diagram of a transmission electron microscope. Electrons are

produced by an electron gun and condensed into an electron beam in the condenser. This
electron beam is focused onto the very thin sample by electromagnetic lenses. The transmit-
ted electrons are collected by another electromagnetic lens and projected onto a fluorescent
screen where they produce a visible image, which can be viewed directly or monitored on a
computer.
This requirement is one of the principal disadvantages of TEM,

although recently there have been rapid developments in the
field of environmental TEM that allow operation under lesser
vacuums. We will focus on the more common, high vacuum
TEM. For this technique, the sample being examined must
be able to withstand the high vacuum; otherwise, the image
won’t be an accurate representation of the sample under nor-
mal conditions. For non-solid samples, two principal methods
of sample preparation have emerged. These methods allow for
the imaging of biological samples or other soft nanomateri-
als that otherwise could not be imaged using TEM. The first
method is to dehydrate the sample, then stain the sample or
coat it with metal to produce the necessary contrast. The sec-
ond is to cryogenically freeze the sample and image the frozen
specimen. One requirement of this second method is that the
sample must be frozen so quickly that it does not have time
to rearrange into a crystalline state. Otherwise, the ordered
ice crystals produce a diffraction pattern of the electron beam
that obscures the image of the sample. If the sample is fro-
zen quickly enough, amorphous ice is produced and one can
obtain an accurate image of the sample in its “natural” state.
The second major difficulty with TEM is the thickness of

the sample. In order to be effectively imaged, the sample must
be thin enough to be transparent to the incoming beam of elec-
trons. As a general rule, the sample should be less than 100 nm
thick, although the exact suitable thickness depends on the
material being examined and the energy of the electrons used.
Thicker samples can be examined if electrons with higher
energies are employed, but at higher energies the electrons
may begin to destroy the sample. In terms of sample mount-
ing, the sample can sometimes be placed on a sample holder
that is very thin and transparent (or mostly transparent) to
the electron beam. For extremely small samples, a sheet of
thin, amorphous carbon films have been used as sample hold-
ers. These thin carbon films are suitable for such applications
because they are electron-transparent nanomaterials that pos-
sess considerable durability even at a thickness of only one
atom.
There are difficulties associated with using TEM to image
nanomaterials, but it is a useful technique for a variety of
applications and boasts impressive resolutions. Normally
TEM is recommended for use in conjunction with other imag-
ing methods to obtain an accurate understanding of the mate-
rial being studied.
Figure 4.43 shows a TEM image of a carbon nanotube that
has been attached to an AFM tip to serve as a nanoinjector
(the operation of a nanoinjector is discussed in the previous
section). The molecular cargo attached to the surface of the
nanotube is clearly visible, thus demonstrating the resolving
power of TEM.
4.9.4 Near-Field Scanning Optical Microscopy

In our introduction to TEM, we discussed the limited utility
of conventional light microscopies in imaging nanostructures.
In general, the resolution of an optical microscope is limited
by the size of the spot to which the light beam can be focused
using magnifying lenses. This limitation is often referred to
as the diffraction limit. The diffraction limit on resolution is
wavelength dependent, as given by the Abbe equation
⎛ λ ⎞
δ ≈ 0.61 ⎜
⎝ NA ⎟⎠ (4.45)
where δ is the resolution of the microscope, λ is the wavelength

of the light, and NA is the numerical aperture of the objective.
For modern objectives examining samples in an aqueous

medium, NA is usually in the range of 1.3 to 1.5. Therefore,
the resolving power of a conventional optical microscope is
approximately half the wavelength of the incident light, typi-
cally ~200 nm for visible light.
Because of the diffraction limit on optical microscopies,
alternative methods have been developed to examine objects
that are smaller than several hundred nanometers. The elec-
tron microscopies (TEM and SEM) and the scanning probe
microscopies (STM and AFM) discussed earlier are the proto-
typical examples. However, these alternative methods gener-
ally suffer from limited applicability to only specific types
of samples. For example, the electron microscopies require
samples that are resistant to very high vacuums and STM
requires samples that are conductive. Furthermore, none
of these alternative methods offer the types of information
that are available to optical methods—such as spectroscopic
information, excellent time resolution, fluorescence detection
capabilities, information about refractive index and reflec-
tance of the sample, and contrasting power using different
staining agents.
Near-field scanning optical microscopy (NSOM or SNOM) is
an optical microscopy that can operate with resolutions below
the diffraction limit, and therefore is able to offer the advan-
tages of optical microscopies with a resolution that is actually
useful for the study of nanomaterials.
4.9.4.1 History and Principles of NSOM

Although the fundamental idea of NSOM is relatively simple,
its practical implementation proved to be rather difficult. The
original idea was developed by Edward Synge and was pub-
lished in a series of papers beginning in 1928. Synge realized
that if the diffraction limit was imposed by the practical limit
to which a beam of light could be focused then the limit might
be overcome by shining light through a very small hole (or
aperture) that was smaller than the wavelength of the light
itself. If this hole were placed close to the sample, the light
would not have time to diffract outward and destroy the reso-
lution of the image. Thus, a sample could be imaged at a reso-
lution below the diffraction limit.
Despite Synge’s development of the theory of NSOM in the
early 20th century, it was not implemented until 1972 when
Ash and Nicholls used an NSOM setup with microwave radia-
tion (λ ~ 3 cm) to image a metal grating sample. They dem-
onstrated that a resolution of 1/60th of the wavelength of the
incident radiation was achievable using their method. Their

results validated Synge’s theory, but practical considerations
prevented the development of an NSOM using the much smaller
wavelengths of visible light until the mid-1980s. During this
period, scientists overcame the technical difficulties of imple-
menting NSOM with visible light. Their practical setup serves
as the basis for modern NSOM instruments.
4.9.4.2 Modern NSOM Instrumentation and

Different NSOM Operating Modes
The major component of any NSOM microscope is the aper-
ture tip or NSOM probe. A variety of NSOM probes exist.
One common NSOM probe is manufactured by heating and
pulling a fiber-optic cable into a very fine point, then coat-
ing the tapered end with reflective metal, except for a very
small aperture at the point. SEM images of this type of NSOM
probe are shown in Figure 4.45. Laser light shone through the
fiber-optic cable emerges from the aperture as a beam with a
diameter that is smaller than the wavelength of light. For this
type of fiber-optic NSOM probe, the fundamental maximum
resolution is ~12 nm, but the practical limit is typically ~50
nm. Another common NSOM probe can be made by using elec-
tron beam lithography to create a nanometer-scaled aperture
through silicon or another metal and then shining the laser
200 nm 200 nm
(a) (b)
Figure 4.45 SEM images of two NSOM probes that have been constructed by tapering
a fiber-optic cable into a very fine point and then coating that tip with aluminum, leaving
only a very small aperture at the point. The ends have also been flattened using a focused
ion beam. In these NSOM probes the apertures are approximately (a) 120 nm and (b) 35 nm.
(Image reprinted with permission from Veerman et al., 1998, Appl. Phys. Lett., 72, 3115–3117
as shown in Dunn, Chem Rev, 1999, 99, 2891–2927. Copyright 1998 American Insitute of
Physics.)
Light source
d<λ
Sample
(b)
Detector
(a)
(c) (d) (e)
Figure 4.46 The different operating modes of NSOM. (a) In Transmission mode,
the light travels from the probe through the sample to a detector on the other side. (b) In
Reflection mode the light from the probe is reflected off the sample surface and captured by
the detector. (c) In Transmission-Collection mode, the sample is illuminated from underneath
and captured by the probe, through which the light travels to a detector. (d) In Reflection-
Collection mode, the sample is illuminated externally to the probe and the reflected light is
captured by the probe and channeled to the detector. (e) In Illumination-Collection mode, the
probe is responsible for both illuminating the sample and collecting the light that reflects
from its surface.
through this nanometric aperture. Both of these probes can be

manufactured with aperture diameters of ~50–100 nm. Recent
developments have also been made in so-called apertureless
NSOM probes. These probes exhibit great promise in achiev-
ing even better sub-diffraction limit resolutions.
An NSOM setup can be operated in at least five different
modes, as listed below (see Figure 4.46).
1. Transmission mode. Light is shone through the NSOM

probe then detected on the other side of the sample.
2. Reflection mode. Light is passed through the NSOM
probe and reflected from the sample. This reflected
light is detected and compiled to produce an image.
3. Transmission-collection mode. Light is shone on the
sample from underneath and the transmitted light is
collected in the NSOM probe and passed to a detector.
4. Reflection-collection mode. Light is shone on the

sample from above and reflected light is collected by
the NSOM probe and sent to a detector.
5. Illumination-collection mode. The NSOM probe is
used as both the light source and the light collector.
Depending on the type of sample examined and the kind of

study undertaken, different NSOM operating modes may be
employed.
Aside from issues related to manufacturing a suitable
NSOM probe, the second major difficulty in the development
of NSOM was probe positioning. In order to function cor-
rectly, the NSOM probe must be placed close to the surface
of the sample being studied, usually less than a few nano-
meters. Such precise positioning is no easy task, especially
when the probe is scanned across a “rough” sample surface.
Furthermore, if the tip comes into forceful contact with the
surface, either the probe or the sample could be damaged.
Therefore, scientists have developed different feedback
mechanisms to ensure that the probe remains at the correct
distance above the sample. The shear-force feedback mecha-
nism and the tapping-mode feedback mechanism are among
the most common feedback mechanisms employed. While the
details of each feedback mechanism are beyond the scope of
this text, the essential idea behind each mechanism is that the
probe is oscillated at a certain frequency and the force that is
placed on the probe as it approaches the surface is monitored.
By using a continuous feedback loop, the probe can be oscil-
lated at the correct height above the sample. Furthermore, if
the tip position is monitored as the sample is scanned in the
x-y plane, then topographical information about the sample
can be obtained (much like with AFM or STM) in addition
to any optical information. Finally, to ensure accurate probe
positioning, the entire NSOM instrument is usually operated
on a vibration isolation table.
A schematic of a typical modern NSOM microscope is
shown in Figure 4.47. The setup depicted is an NSOM micro-
scope with a fiber-optic probe operating in transmission mode.
The laser light is initially passed through a bandpass filter and
a combination of half-wave and quarter-wave plates to con-
trol the wavelength and polarization of the light, respectively.
The light then enters the fiber-optic cable and is transmitted
to the probe tip through which it is shone on the sample. The
transmitted light is collected by the detector and processed
into an image. The height of the probe tip is controlled by the
Optical
components
Laser
NSOM probe
Fiber optic cable/
feedback mechanism Sample on
piezoelectric stage
Detector
Figure 4.47 Schematic of a near field scanning optical microscope operating in

Transmission mode. Laser light is channeled into a fiber-optic cable that ends in an NSOM
probe tip positioned very close to the sample surface. The transmitted light is captured
by the detector to produce an image. The sample is scanned by the piezoelectric scanning
stage, and the NSOM probe height is monitored and controlled by the feedback mechanism.
feedback mechanism, and the probe’s vertical position can be

monitored to produce topographical information. The sample
itself is placed on a piezoelectric stage that can move in the
x-y direction, allowing for the sample to be scanned in a raster
pattern, much like with AFM or STM.
An example of an NSOM image is shown in Figure 4.48,
which shows the topographical and cross-polarized optical
NSOM images of spherical and toroidal (donut-shaped) liquid
crystal droplets suspended in a polymer. Liquid crystals are
materials that have properties between those of a liquid and
those of a solid. There are many types of liquid crystals based
on the orientation of molecules, but for our present purposes
we merely note that the general shape of the droplets is clearly
seen in the topographical images and that the orientation of
the crystals can be seen in the cross-polarized images.
NSOM setups are often used in conjunction with other sur-
face imaging methods to provide complementary information.
For example, NSOM microscopes are commonly incorporated
into conventional fluorescence microscopes, and surface-
enhanced Raman NSOM setups have also been employed.
Overall, NSOM is a useful imaging method that can offer
a wealth of information about a nanomaterial being studied.
While its maximum resolution (often ~50 nm) is not quite as
impressive as some of the other imaging methods we have dis-
cussed, its versatility in being able to image a wide variety
of samples under “normal” conditions and its richness in the
2 µm 2 µm
2 µm 1 µm
2 µm
Figure 4.48 NSOM images of spherical and toroidal liquid crystals suspended in a
polymer. (Image from Mei et al, Langmuir, 1998, 14, 1945–1950 as shown in Dunn, Chem
Rev, 1999, 99, 2891–2927. With permission.)
types of information it provides usually compensate for this

lower resolution.
4.10 LIGHT SCATTERING METHODS

Surface tension measurements can be used to study the for-
mation of a surfactant monolayer at the air-water interface.
Beyond the critical micelle concentration (CMC), surface ten-
sion values from such solutions become constant, indicating
the formation of colloidal particles such as micellar phases or
nanoparticles, which have strong tendencies to scatter radia-
tion, especially x-rays or a beam of neutrons. There are three
ways that particles can scatter light in which there is no energy

transfer (elastic light scattering): Rayleigh scattering, Mie scat-
tering, and geometric scattering. The type of scattering depends
on the size of the particle interacting with electromagnetic
radiation. Consider the unitless parameter, α, defined as,
πD
α= (4.46)
λ
where πD is the circumference of the particle and λ is the wave-

length of the incident light. Rayleigh scattering occurs when
α << 1, or when the particles are small compared to the wave-
length of light. Mie scattering occurs when the particles are the
same size as the wavelength of light (i.e., α ≈ 1) and geometric
scattering occurs when the particles are relatively large (i.e.,
α >> 1).
4.10.1 The Measurement of Scattered Light:

Determining the Aggregation Number of Micelles
The intensity, angular distribution, and polarization of the
scattered light depend on factors such as the shape and size
of the particles, as well as on the interactions between them.
Light scattering experiments can thus provide useful struc-
tural information (particle shape and size) and interparticle
interactions on colloidal systems. The measurements are gen-
erally instantaneous, non-invasive, and allow representative
sampling of polydispersed samples. However, the presence of
small particle impurities, particularly those that have a ten-
dency to scatter the light, can impose serious errors in the
measurement.
In practice, a collimated beam of light of a given wavelength
(λ) and intensity (Io) passes through a solution containing the
dispersed nanoparticles (Figure 4.49). The intensity of the
scattered light is then measured as a function of the angle (ϕ)
between the incident beam and the scattered beam.
We will begin by discussing how light scattering can be used
to determine the aggregation number of a micelle. Micelles
have diameters typically on the order of a few nanometers. The
wavelength of visible light is about two orders of magnitude
greater. Let’s consider a beam of visible light passing through
an aqueous solution containing spherical micelles. The solu-
tions can be described in terms of two refractive indices, the
refractive index of the randomly dispersed nano-spherical
micelles (nmicelle) and the refractive index of the continuous
solvent (nsolvent). These two refractive indices have different
Unpolarized light
Figure 4.49 The scattering of unpolarized light through a sample. The intensity of the
scattered light is measured as a function of the angle (ϕ) between the incident beam and
the scattered beam.
values, and the average refractive index of the solution will

therefore vary with the local concentration of micelles. This
variation will cause light to be scattered. The intensity of the
scattered light depends on the intensity and wavelength of the
incoming light, the solution refractive index increment (i.e.,
how n varies with concentration, dn/dc), and the average num-
ber density of micelles in the solution (N). These parameters
can be determined experimentally and are used to obtain the
optical constant Ko (Equation 4.47).
2
n2 ⎛ dn ⎞
K o = 2π 2
λ 4N ⎜⎝ dc ⎟⎠ (4.47)
In order to understand how the intensity of light varies with

the scattering angle ϕ, we need to describe the incoming unpo-
larized light as being composed of two mutually perpendicu-
lar polarized components (Figure 4.49). For small scattering
angles (ϕ ∼ 0), these components will contribute equally to the
scattered intensity. At very large scattering angles, one of the
two polarized components contributes to a greater degree to
the scattering. In fact, when ϕ = 90° the component polarized
along the direction of the scattered beam has no contribution
to the scattered intensity. By measuring the intensity of scat-
tered light as a function of ϕ, we can determine a quantity
known as the Rayleigh ratio (Equation 4.48).
d2 I
Rφ = 2 (4.48)
1 + cos φ I o
In this equation, d is the distance between the sample and

the detector. The Rayleigh ratio will be different for a micellar
solution compared to the pure solvent. In fact, it can be shown
that the difference between these two Rayleigh ratios is given
by Equation 4.49.
2
2π 2n2 ⎛ dn ⎞ dΨ
ΔRφ = RTc (4.49)
λ 4 ⎜⎝ dc ⎟⎠ dc
The term RTc dψ/dc is known as the concentration fluc-

tuation factor and describes the free energy cost in creating
an inhomogeneity in micelle concentration. For micellar sys-
tems, the concentration fluctuation factor causes a dramatic
change in the scattering intensity, as we go from monomers
to predominantly micelles. Therefore, one would expect that
∆Rϕ would be related to both the CMC of the surfactant and
the concentration of the surfactant monomers (cm). The exact
dependence is shown in Equation 4.50.
K o (cm − CMC )
ΔRφ = 3 (4.50)
10 /M + 2B(cm − CMC )
M represents the molecular weight of the micelle, and B is a

constant known as the second virial coefficient. The sign of
B provides information on the intermolecular interactions
between micelles. A negative value indicates a net attraction
between the micelles, and a positive value indicates a net repul-
sive interaction. A value of zero indicates an “ideal” micellar
solution in which there are intermicellar interactions.
Example 4.9 Determining the Aggregation Number of a Micelle
Consider values of Ko/∆Rϕ recorded as a function of surfactant

monomer concentration. How would you determine the aggregation
number of the micelle?
Equation 4.50 can be rearranged to give
K o (cm − CMC ) −3 1
10 = + 2 × 10−3B (cm − CMC )
ΔRφ M

Thus, a plot of the left-hand side of this equation versus (cm –

CMC) will give a straight line with an intercept equal to 1/M and a
slope equal to 2 × 10 –3 B.
4.10.2 Dynamic Light Scattering

Dynamic light scattering (DLS), also known as photon corre-
lation spectroscopy, is a method used to determine the size
distribution of particles in a solution. This is a powerful tech-
nique that can accurately report the sizes of particles present
across several orders of magnitude. Furthermore, measure-
ments are typically easy to make; this makes DLS ideal for
many applications from measuring the size of micelles and
other nanoparticles as a function of concentration to assaying
a protein solution for the presence of aggregates.
DLS works by shining a laser at a sample solution and
observing the scattered light. If the particles in the solution
are small compared to the wavelength of the light (if the wave-
length is more than 10 or so particle diameters) then this phe-
nomenon will proceed primarily through Rayleigh scattering.
In this process, particles absorb a photon and then reemit a
photon shortly thereafter. However, the emitted photon is sent
in a random direction. Thus, even if the entire incident light
comes from one direction, the emitted light is radiated in all
directions (scattered). This is an elastic process, meaning that
if light of a particular wavelength is absorbed, the scattered
light will have exactly the same wavelength. With this in
mind, if one conducted an experiment shining a 632-nm HeNe
laser at a solution of interest, one might naively expect the scat-
tered light to look like that shown in Figure 4.50(a). This figure
depicts all scattered light being detected at exactly 632 nm.
In actuality, however, scattered light is observed in a distribu-
tion centered on the expected wavelength [Figure 4.50(b)]. The
reason for this is that the particles in solution are undergoing
Brownian motion. This means that at any given moment, some
particles are traveling toward the detector and have their emit-
ted light blue-shifted relative to the incident light. Similarly,
some particles are moving away from the detector and have
their emitted light red-shifted. On average, however, the par-
ticles are not moving relative to the detector and thus the distri-
bution’s center lies on the wavelength of the incident light.
The observed wavelength of scattered light is not the only
thing affected by Brownian motion. The total observed inten-
sity also fluctuates in time. This is because Brownian motion
causes the distance between any two adjacent atoms to vary
in time. The light emitted from these atoms experiences inter-
ference; the distance between the atoms determines whether
this interference is constructive or destructive. Thus, as this
distance varies for a certain pair of atoms, the net light emitted
Intensity 632
Wavelength (nm)
(a)
Intensity
632
Wavelength (nm)
(b)
Figure 4.50 (a) In a simplistic model all scattered light is detected at one wavelength
(the wavelength of the incident light). (b) The actual observed scattered light; the wave-
lengths are distributed around the expected wavelength due to Doppler shifting of the light
from Brownian motion of the particles.
from them will fluctuate in intensity. Since there are so many

particles in solution, the net intensity of scattered light approx-
imately evens out and is never very far from the average inten-
sity. However, it does not perfectly average out, and the net
intensity of signal in solution can be seen to fluctuate.
Critically, the intensity at any given point in time depends
on the location of all of the particles in solution. Thus, if one
takes a measurement at time zero and another measurement
very quickly after that, the intensity of the second measure-
ment will be very similar to that of the first measurement so
long as the particles have not had sufficient time to diffuse far
from their original positions. If, on the other hand, the parti-
cles have had enough time to move significantly, the intensity
will be essentially random. This concept can be expressed as
a correlation. That is, if the second measurement is likely to
be more similar than average to the first measurement, that is
a positive correlation. If, however, the second measurement is
A nanopartical
(e.g., SiO2, 95 nm diameter)
Correlation
A protein of 5 nm diameter
Time
Figure 4.51 A plot of correlation versus time delay between subsequent measurements
for two different particles. As expected, the larger particles stay correlated for a longer
period of time, indicating that they diffuse more slowly.
likely to be less similar than average to the first measurement,

that is called negative correlation. A correlation of zero means
there is no relation between the measurements and the first
intensity gives the experimenter no information about the sec-
ond intensity. In a DLS experiment, as the time delay between
the two intensity measurements increases from zero to infin-
ity, the correlation decreases from unity (perfect correlation)
to zero (no correlation).
We can take this data and plot correlation versus time delay
for solutions of different particles (Figure 4.51). This graph
gives a time scale for how fast a given particle diffuses. The
faster it diffuses, the more quickly it will reach zero correla-
tion as the particles need less time to move to different posi-
tions from where they started. In fact, these correlation graphs
can be used to calculate the rate of diffusion, D, which can be
employed in further calculations.
The aforementioned rate at which a particle diffuses can be
entered into Equation 4.51 to calculate the particle’s hydrody-
namic radius. In this equation dH is the hydrodynamic diam-
eter, kB is the Boltzmann constant, T is temperature, η is the
solution viscosity, and D represents the rate of diffusion (called
the translational diffusion coefficient).
k BT
dH = (4.51)
3π ηD
As one might expect, the larger the particle, the slower it
diffuses, as shown in the Figure 4.51; this is why the diffusion
rate and diameter have an inverse relationship. The number
of particles observed at each diameter can be plotted versus
diameter to obtain a graph such as that in Figure 4.52. This

figure demonstrates how a complex correlation curve with
two distinct exponential regions can be transformed into an
easy-to-interpret graph with two peaks at the corresponding
particle sizes.
There are several things that can affect diffusion speed
besides particle size that need to be controlled for (or at least
considered). The first is the size of the hydration shell around
each particle. Since it is the hydrodynamic radius of each par-
ticle that is measured, if two particles have equal size, but one
is more strongly coupled to the surrounding media, that parti-
cle will diffuse more slowly and register as larger. Second, the
ionic strength of a solution affects the coupling of the media
to the particles, so ionic strength must typically be set at a
standard level. Third, the surface morphology of a particle
can affect diffusion speed. If a particle has long, comb-like
appendages on its surface, they will slow its diffusion and make
it appear larger than a smooth sphere of equal size. Finally,
changing particle geometry can affect diffusion speed. If a
particle shifts from a more compact spherical formation to
a more extended conformation, it will be slowed down and
again appear larger despite not changing mass at all. This last
effect is of particular interest in protein analysis because it
can sometimes detect protein morphology shifts between dif-
ferent forms.
One important limitation to DLS is that the signal inten-
sity received from a given particle is proportional to the sixth
power of the particle’s diameter. Thus, if one has a solution
containing a 1:1 mixture by molar concentration of 10-nm par-
ticles and 100-nm particles, the peak area corresponding to
the larger particles will be a million-fold larger than that for
the smaller particles despite the solution containing an equal
number of each. This problem only worsens as the size differ-
ence increases and serves to limit the range of particles that
can be accurately analyzed in a single solution.
Dynamic light scattering is an excellent method for deter-
mining the sizes of particles present in a solution. It is able to
accurately size particles from the nanometer scale to above
the micron scale so long as any one particular solution does
not contain widely varying particle sizes. This technique has
found utility in applications ranging from colloidal science to
proteomics, and has become an integral part of many nanosci-
ence laboratories. Few other methods of particle sizing present
such an attractive combination of ease of use and wide region
of sensitivity.
3.5 nm
15 nm 400 nm
Intensity
Correlation
0
Time Log diameter (nm)
Figure 4.52 A plot of signal intensity versus particle size along with the correlation plot from which such a graph is derived.
199

• Evans, D. F. and Wennerström, H. 1999. The Colloidal
Domain, 2nd ed. New York: Wiley-VCH. Chapter 2 pro-
vides a light read on AFM and STM.
• Tompkins, H. G. 2006. A User’s Guide to Ellipsometry.
Dover Publications, Mineola, NY. This is a graduate-
level book that covers the theory of ellipsometry in
detail and provides many interesting applications of
the technique.
• Vickerman, J. C. 2003. Surface Analysis—The Principal
Techniques. John Wiley & Sons, Chichester, West Sussex,
UK. This book provides an excellent coverage of surface
science techniques, many of which are not covered in
this chapter. The book is recommended for students
who wish to gain a better understanding of vibrational
spectroscopy at surfaces and scanning probe methods.
• Boyd, R. W. 1992. Nonlinear Optics. Academic Press,
San Diego, CA. This book is recommended for the
advanced student interested in nonlinear methods
such as second-harmonic generation and sum-fre-
quency generation. The book provides an excellent
fundamental treatment of nonlinear optics.
• Bonnel, D. 2000. Scanning Probe Microscopy and
Spectroscopy: Theory, Techniques, and Applications,
2nd ed. Wiley-VCH, New York. This book focuses
mainly on STM and is only recommended for the stu-
dent who is seriously interested in this method.
• Berne, B. J. 2000. Dynamic Light Scattering: With
Applications to Chemistry, Biology, and Physics. Dover
Publications, Mineola, NY. This book explains DLS
and how Maxwell’s equations lead to the intensity of
the scattered radiation. This is an advanced graduate-
level textbook.

1. Equation 4.1 can be used to determine the surface ten-
sion of a liquid by measuring how far the liquid travels
up a narrow capillary tube. How far would water rise
up a narrow capillary of radius 1 mm? Assume that
water makes a zero contact angle against the surface of
the tube.
2. A paper plate of width 10 mm and thickness 1 mm is
withdrawn from a surfactant solution. A force of 0.77
mN is measured just before the plate detaches from the

surface of the solution. Estimate the surface tension of
the surfactant solution.
3. In a QCM experiment, the following values of fre-
quency (at the third overtone, n = 3) were measured
in real time as a nanofilm of mammalian DNA was
adsorbed on a quartz crystal with Sauerbrey constant
C = 17.7 ng Hz–1 cm–2.
Time 0 5 10 15 20 25 30 35 40 45 50 55 60 65
(min)
∆F 0 0 –2 –5 –7 –15 –25 –42 –55 –60 –63 –64 –65 –65
(Hz)
Use the Saurbrey equation to generate a plot of time

versus adsorbed mass. Assume the adsorption follows
first-order kinetics, i.e., [Adsorbed DNA mass] = (1 –
Ae –kt) where A is a constant and k is the first-order rate
constant. Use this equation to obtain a value for k.
4. Consider solutions of bovine serum albumin (BSA), a
common blood protein, made at pH 3, 5, and 7 flowed
over a silicon waveguide in a DPI experiment. pH 5
is close to the isoelectric point of BSA (or the pH at
which it has no net charge), where it is expected that
each BSA protein experiences little electrostatic repul-
sion with neighboring proteins at pH 5.
(a) Predict at what pH you would expect the maxi-
mum adsorption of BSA. Explain your answer.
(b) In bulk solution, BSA adopts an ellipsoid shape
with dimensions of 140 Å × 40 Å and it has a
molecular weight of 66.43 kDa. Using this infor-
mation in conjunction with the DPI mass values,
how would you calculate the area occupied by
each protein molecule on the waveguide surface
at each pH?
(c) If we were to assume that the BSA proteins
adsorbed parallel to the waveguide surface to
form a saturated monolayer and that they were
not distorted upon adsorption, then we could
use the bulk phase dimensions of BSA to calculate
the nanofilm thickness (it’s ~40 Å and each mol-
ecule occupies ~5600 Å2 on the surface at pH 5).
The nanofilm layer thickness values are much
less for the BSA proteins at pH 3 and 7, and the
area-per-molecule values are much greater than
predicted. Thus, it can be inferred that the BSA
molecules at pH 3 and 7 are spread much more

thinly on the waveguide surface and that they
assume a much more distorted form. The relative
rates of adsorption of BSA at different pH to the
silicon surface are pH 3 > pH 5 > pH 7. What forces
are responsible for the adsorption of BSA to the
waveguide surface at each pH? What explanation
can be offered for the observed order of the rela-
tive rates of adsorption of BSA at differing pH?
5. The following table shows the refractive index of poly-
ethylenimine (PEI) as a function of concentration. The
refractive index of a film of PEI on the surface of a DPI
waveguide is 1.55. Assuming this film is under a pure
water solution, estimate the density of the film. In a
separate experiment, ellipsometry determined this
film to have a thickness of 1 nm. What is the mass of
the film?
Concentration (mM) 1.0 5.0 10 15 20

Refractive Index 1.44 1.45 1.46 1.47 1.48
6. Common IRE materials used in ATR-FTIR are zinc

selenide (ZnSe) and germanium (Ge). Using tabulated
values in the scientific literature, we can find the index
of refraction values for ZnSe and Ge at a variety of inci-
dent light wavelengths. Using other known experimen-
tal parameters allows us to calculate the penetration
depth of the evanescent wave into the sample being
studied. For example, at IR frequencies of ~1700 cm–1
(the typical stretching frequency of a C=O bond), ZnSe
has an index of refraction of ~2.35, whereas the RI of
Ge is ~4.0. Assuming that the angle of incidence used
in an IRE of an ATR-FTIR is 45°, compare the evanes-
cent wave penetration depths in an aqueous sample
(RI ~1.5) for an IRE made of either ZnSe or Ge when IR
radiation of 1700 cm–1 is used.
Hint: See the Appendix for a useful equation to calcu-
late the penetration depths.
7. Consider a nanofilm comprised of discrete layers of a
chromophore, such as a dye molecule. Ellipsometry
can measure the thickness of such a film. If the molar
absorptivity of the molecule is known, show how
Equation 4.23 can be used to determine the mass of
the film.
8. The Sauerbrey equation underestimates the mass

adsorbed to the crystal surface in liquids because it
was originally developed for oscillation in air and does
not account for the decrease in resonant frequency due
to the change in viscosity. It is possible to amend the
Sauerbrey equation for an accurate measurement of
mass in liquid by correcting for these viscosity effects
according to the following equation:
⎛ ηρ ⎞
Δf = − fo3/2 ⎜ l l ⎟
⎝ πρq μ q ⎠

where f0 is the resonant frequency. ρl is the density of

the liquid. ηl is the viscosity of the liquid. ρq is the
density of quartz (2.648 g/cm3). μq is the shear modulus
of quartz for AT-cut crystal (2.947 × 1011 g/cm s2). In
which liquid, water or deuterium oxide, is the detec-
tion limit of a typical QCM-D instrument at its fun-
damental resonant frequency more sensitive? What is
the ratio of their sensitivities? (The interested student
should see the following reference for more details: K.
K. Kanazawa and J. G. Gordon, II, “The oscillation fre-
quency of a quartz resonator in contact with a liquid,”
Anal. Chimica Acta, vol. 175, pp. 99–105, 1985).
9. (a) The text mentions that many techniques cannot
by themselves provide absolute measurements of
properties such as refractive index, thickness of
a film, or density. Choose two non-spectroscopic
techniques discussed in the chapter and describe
how their combined use provides more informa-
tion than either would alone. Outline a set of
experiments that you would undertake to learn as
much as possible about a particular film-forming
substance such as SDS (or provide a compound).
(b) Choose a spectroscopic method, such as IR, and
describe how this technique could add to your
understanding of the material and its adsorption
properties.
10. Describe how you would use information given about
molecular fluorescence spectroscopy to calculate Φf
for a set of data. (Hint: Use an appropriate equation
and integrate it to find gx and thus calculate Φf.)
11. Scanning tunneling microscopy is used to image

a metal surface with an electronic decay length of
5 nm–1. If the tunneling current increases by a factor
of 5, then how much has the height changed? Describe
what this height refers to in your own words. (Hint:
See Example 4.8.)
12. Rank from tallest to shortest the height of an AFM tip
that has been functionally coated with –COOH as it
traverses a surface of the following makeup:
COOH CH3 OH NH4+
CH2 CH2 CH2 CH2
A B C D
13. I’m using NSOM to detect the incorporation of malar-

ia-causing Plasmodium proteins in the membrane
of malaria-infected human erythrocytes. I place my
sample 300 nm from the NSOM probe, but see that the
resulting pictures are not resolved. I then realize that
the wavelength of the light source from my laser is 150
nm. What must I do to increase the resolution?
14. What is the concentration fluctuation factor of a sur-
factant that forms micelles with very large molecular
weight? Express your answer in terms of the second
virial coefficient.
15. (a) Explain in words why the distribution of Rayleigh
scattered light is a broad peak instead of a sharp
signal.
(b) If a laser emitting light at 700 nm is sent through
a dispersion and the Rayleigh scattered light is
centered around 640 nm, what can be deduced
about the direction of motion of the particles in
solution?
(c) Does the shift mentioned in part (b) correspond to
a red shift or blue shift?
16. (a) If the hydrodynamic radius of a particle doubles,
what happens to its diffusion speed?
(b) If the radius is halved, the temperature is qua-
drupled, and the viscosity of solution increases
by half, by what factor does the diffusion speed
change?
(c) In a solution of hexane, would one expect to ob-

serve multiple signals? Explain why or why not.
17. Explain why the maximum feature resolution in con-
tact lithography is approximately λ/2. (See Section 5.9.)
18. This question concerns SFG spectroscopy. Consider
the overlap of two light beams (one visible of wave-
length 523 nm and the other IR of wavelength 2.2 µm)
at a solid surface. The angle of incidence of the visible
beam is 30° and that of the IR beam is 40°.
(a) Calculate the angle of the emitted SFG beam.
(b) What is the wavelength of the emitted SFG
beam?
(c) Describe qualitatively the difference between the
SFG sprectra of SDS, CTAB, hexanol, and decanol
in terms of the CH2 and CH3 bands. Assume that
monolayers of these molecules are present between
a hydrophobic surface and an aqueous phase.
19. In a study conducted by the Tokyo Institute of
Technology, QCM was used to directly monitor the
reactions of the Klenow fragment of Escherichia coli
DNA polymerase I. The reactions take place on DNA
oligonucleotides, containing either a (TTTTC)3 or
(TTTTC)10 template and the primer necessary for poly-
merase binding, which have been immobilized on the
QCM. This enzyme acts as a catalyst for complemen-
tary base pairing of dATP and dGTP monomers from
solution into a second strand. The results of two trials
with different initial conditions for the (TTTTC)3 tem-
plate are shown below:
Polymerase
0 0
(b)
∆m/ng cm–2
–100
∆F/Hz
dATP + dGTP
(a)
100
–200 Step 1 Step 3
Step 2
–300
200
0 10 20 30 31 32 33
Time/min.

In (a) polymerase was added at first, and then

excess monomers were added after 30 minutes. In
(b) polymerase was added in the presence of excess
monomers.
(a) What is physically occurring at steps 1, 2, and
3 of (a) in terms of the biological components of
the reaction? What is physically occurring in (b)
that causes a different curve? Why does it make
sense that the final changes in frequencies are the
same?
Hint: The frequency change in step 1 of (a) is of
roughly equivalent magnitude to the frequency
change in step 3 of (a), and the frequency change
at step 2 of (a) is roughly equivalent to the total
frequency change in (b).
(b) What can be said about the relative magnitudes
of the rate constants for each of the three steps in
the DNA polymerase reaction mechanism based
off the data above?
(c) What relative changes in frequency, coupled mass,
and rate constants would be expected if using
the (TTTTC)10 template instead of the (TTTTC)3
template?
Five
Types and Uses of

Some Nanomaterials
CHAPTER OVERVIEW
This final chapter is written more like a scientific review of
nanomaterials, with the goal of helping the student to transition
from the fundamental textbook-style material to the primary
scientific literature. The previous chapters should have given
the student the necessary scientific background and terminol-
ogy to understand the material presented in this chapter. This
chapter can be thought of as two separate sections: tuning opti-
cal and energetic properties of materials using nanomaterials
(Sections 5.1 to 5.4) and functionalizing surfaces using nanoma-
terials (Sections 5.5 to 5.9), brought together as they both define
useful ways to employ nanotechnology. There are many other
potential applications of nanomaterials, but herein an overview
is given of how to functionally use various technologies, each
of which is constantly evolving the state of nanoscience.
5.1 SUPRAMOLECULAR MACHINES

In every self-assembly process highly selective interactions
occur that control the transition from molecular scale aggre-
gates to nanoscale entities. As the assembly and character-
ization of molecular-level components has progressed, the
potential for nanoscale systems to be functionalized to per-
form simple tasks is beginning to be realized. Supramolecular
chemistry is concerned with the study of the basic features of
these interactions and with their implementation in specially
designed non-natural systems. The broad category of all such
nanoassemblies has been dubbed supramolecular machines.
These machines are comprised of organized systems of
cooperating molecules, and are deliberately and specifically
engineered with the ability to function as sensors, processors,
and custom catalysts. One area of supramolecular fabrication
207
that is generating considerable interest is the specific pro-

gramming of energy transfer through an assembly, which is
accomplished by a system of interacting dye molecules. This
section addresses several systems of interacting dye molecules
in nanofilm assemblies and discusses several applications
where the careful orientation of donor and acceptor fluores-
cence partners is exploited for specific film functions. This
section was inspired by the excellent coverage of this mate-
rial in Kuhn and Försterling’s textbook, Principles of Physical
Chemistry: Understanding Molecules, Molecular Assemblies,
Supramolecular Machines. Students are encouraged to read
Chapters 22 and 23 of this book.
5.1.1 Model Dye System

In order for an assembly of molecules to become a useful sys-
tem for performing work, it is necessary to fabricate an effective
means of energy transfer to and/or across the machine. Photon
energy is a highly favorable way to deliver such an energy
source for a variety of reasons, including the ability to tune the
wavelength (color) of the source, minimal invasiveness, and
the ability to selectively excite various sites of an assembly.
Processes that harness light energy for later use as chemical
energy are abundant in nature, and thus make the understand-
ing and fabrication of such devices an important task.
The most basic method by which photon energy can be
harnessed by molecular systems is through the exploitation
of the specific absorption-fluorescence characteristics of dye
molecules. Let us consider a model system of two interact-
ing dye molecules: donor (D) and acceptor (A) dyes, as shown
in Figure 5.1. Donor D can be irradiated with light of wave-
length 450 nm (blue light), resulting in excitation and fluo-
rescence at 550 nm (green light). A different dye, acceptor A,
behaves similarly upon exposure to 550-nm wavelength light
by absorbing it and fluorescing red light at 620 nm. We can
exploit the overlap of D’s fluorescence and A’s absorbance in
Blue in Red out
450 nm D 550 nm A 620 nm
Figure 5.1 Excitation of donor molecule (D) by blue light (450 nm) and its subsequent
fluorescence of green light (550 nm). The emitted green light is absorbed by acceptor mol-
ecule (A) which finally emits red light (620 nm).
Types and Uses of Some Nanomaterials 209
Acceptor molecule (A)
Distance d
Donor molecule (D)
Fatty acid CH3(CH2)5COOH
Glass substrate
Figure 5.2 A layer-by-layer assembly in which a monolayer of donor molecules is sepa-

rated from a monolayer of acceptor molecules by “inert” fatty acid multilayers. The dis-
tance d is determined by the number of discrete fatty acid monolayers between the A and
D monolayer.
order to fabricate a simple molecular machine. Specifically,

we can monitor the intensity I of fluorescence from dye mono-
layers on fatty acid precursors and spacers to determine their
distance of separation, d. These films can be assembled easily
via the hydrophobic/hydrophilic interactions in a Langmuir-
Blodgett (discussed later) or electrostatic self-assembly (ESA)
deposition process. As shown in Figure 5.2, the hydrophilic
fatty acid headgroups associate with a glass substrate, expos-
ing the molecule’s hydrophobic tail. Subsequent depositions
of dye molecules D, a certain number of fatty acid spacers, and
molecules of A result in a simple molecular machine. Using
photoquenching equations derived by Drexhege et al. (1963)
and later applied to this system by Mobius (1969), the intensity
I of the green (D) and red fluorescence (A) can be calculated
at any separation distance d and verified by monitoring I in
spectroscopic experiments.
Figure 5.3(a) expresses the quantum yield as a ratio of green
fluorescence intensity from D when A is located at a dis-
tance d (Id.D) to the intensity of green fluorescence when d is
infinitely large or A is absent (I∞.D). This value is a function of
(a)
Id/I∞
(b)
Id/I∞
do d (nm)
Figure 5.3 Quantum yield profiles illustrating energy transfer from D to A. (a) Green
fluorescence yield from D as a function of distance d. (b) Red fluorescence yield from A as
a function of distance d. As d increases, energy transfer to A is reduced resulting in more
fluorescence from D and less from A.
the separation distance of the two dyes, where d0 is the dis-

tance where half of the fluorescence from D is quenched by A.
Conversely, Figure 5.3(b) expresses the ratio of red fluorescence
intensity from A when precursor D is located at a distance d
(Id.A) to the intensity of red fluorescence when d is infinitely
large (I∞.A) as a function of d. On the left, Id/I∞ for donor fluo-
rescence is plotted versus d. As expected, green fluorescence
increases with d since more green fluorescence is transmitted
as acceptor dye A becomes less accessible by d increasing. In
Figure 5.3(b), Id/I0 for acceptor fluorescence is plotted against
d. As expected, red fluorescence decreases with d since A is
excited with less efficiency as it becomes more distant from
the excitons (the excited regions of the assembly) released
from D, which are needed to create fluorescence in A.
These results themselves do not furnish us with much
information beyond a confirmation that our theoretical model
for energy transfer between the dyes is valid. However, Lehn
(in Supramolecular Chemistry, VCH Weinheim, 1995) presents
a compelling application for such behavior in dye assemblies.
Consider two antibodies that both bind specifically and at a
different site to an antigen. By coupling dyes A and D to the
two different antibodies, the dyes become sufficiently close to
allow energy transfer. By introducing an excess of the modified
dyes, spectroscopic analysis could be used to perform a simple
immunoassay. After accounting for background fluorescence
from modified antibodies unbound to the antigen interacting
with one another directly, fluorescence from acceptor A could

be monitored and would indicate the specific concentration of
the antigen in question.
5.1.2 Photorelaxation
The functionality of the aforementioned dye assemblies
hinges on the differences in absorbance and emission spectra,
or Stokes shift, of the two entities (not to be confused with
Stokes shifts in Raman spectra). As a molecule is subjected to
energy, in this case energy in the form of photons, it enters an
excited, quantized electronic state. This excited state cannot
be sustained—in fact, exciton lifetime is approximately 10
nanoseconds—and the molecule quickly reverts to its ground
state. Upon returning to ground electronic state, the mole-
cule is left with excess energy from the initial excitation, and
must thus emit that energy in several forms. In dyes, energy is
dissipated via the emission of a photon. Visible color is emit-
ted from photorelaxing dyes in very specific cases when the
energy band gap, or energy difference between ground and
excited states, lies between the energies available to photons
with a visible wavelength. Since the band gap and emission
of a photon from an excited molecule can be quantified by
hc
E BandGap = E Photon =
λ (5.1)
where h = 6.62606896 × 10−34 J s and c = 2.998 × 108 m/s,

we can determine that the energy band gap must exceed
2.64866 × 10−28 J and be less than 5.22762 × 10−28 J (or 1.65316
× 1019 – 3.26282 × 1019 eV) in order to allow for the dye to emit
a quantum of visible light. The difference between the excita-
tion photon wavelength necessary to induce an electronically
excited state and the wavelength of the subsequently emitted
photon is known as the Stokes shift. The Stokes shift is most
often positive; in other words, the energy of the exciting pho-
ton is partially diminished via other non-radiation energy loss
processes (see Figure 5.4) during excitation, and thus the emit-
ted photon is of lower energy (higher wavelength).
Example 5.1 Stokes Shift
Estimate the Stokes shift in kJ/mol for the spectra shown in

Figure 5.4.
Solution ∆λ is around 100 nm. According to Equation 5.1,
Δλ
Fluorescence
Absorption
300 400 500
Wavelength (nm)
Figure 5.4 A typical excitation (absorption) and emission (fluorescence) spectrum of a

dye molecule. The difference ∆λ is a measure of the Stokes shift.
E=
hc
=
( )(
6.62 × 10−34 Js 2.998 × 108 m/s
≈ 2 × 10−18 J
)
−9
λ 100 × 10 m
For one mole we multiply E by the Avogadro number (6.022 ×

1023 mol–1) and then divide by 1000. This gives 1200 kJ/mol.
In certain materials, however, anti-Stokes shifts occur and

are interesting phenomena. At first glance, an anti-Stokes shift
seems to break fundamental laws of nature: a lower energy
photon is transformed into a higher energy one. Where is this
energy coming from? The answer to this question lies in the
molecule’s ability to remove heat from itself. In materials exhib-
iting anti-Stokes shift behavior, photon energy is absorbed;
however, before reverting to ground state, the exciton also
absorbs a quantized mode of vibration. This mode is modeled
as a particle, called a phonon. This extra energy absorption
creates a higher energy photon emission; however, loss of this
vibration energy creates a loss of heat in the molecule. In this
way, anti-Stokes shift materials cool as they photorelax.
5.1.3 Formation and Properties of the Exciton

When we examine dye molecules 1 and 2 in interaction with
one another with 2 deposited on top of 1, we find that the absorp-
tion spectrum is different from the sum of their individual
spectra. Specifically, dye 1, absorbing at lower wavelengths, is
shifted to even lower wavelengths (higher energy), and dye 2,
absorbing light at higher wavelengths, is shifted to higher
wavelengths (lower energy). In order to explain the cause of
this difference, we can approximate the behavior of the dye
molecules by oscillating charges exposed to the oscillating
forces of the incident light. When considering our system

as a whole, we must account for the interactions among the
electrons to be excited. As incident light reaches the system,
each π-electron system is either in phase or out of phase with
its neighbors. In-phase oscillation shows a strong absorption
band because both oscillating charges are accelerated by the
electric field of the excitation. In out-of-phase oscillation, a
weak absorption band is realized because one charge is accel-
erated while its neighbor feels an opposing electric field and
is decelerated.
However, we must consider our aggregate system of dyes D
and A. In an aggregate layer, the oscillator that we use to rep-
resent a dye molecule will be strongly coupled to all the other
dye molecules (oscillators) in the layer due to the tight packing
in the system. Namely, the π-electrons of each molecule to be
excited are in close contact with all the other electrons, and
their oscillations are in phase. This means that as incident
light is exposed to the dye interface, an in-phase oscillation
among all the layers’ oscillators occurs and an excited domain
forms, having absorbed the energy from the incident photons.
We refer to this excited domain as the exciton.
We can quantify the approximate lifetime of the exciton in
our model system. It has been shown that the fluorescence life-
time of a monomer τ0 in terms of our oscillator model is
3mε 0c03
τ0 = (5.2)
2Q 2πν20n
where m, Q, v0 is the mass, charge, and frequency of the oscilla-

tor, ε0 is the permittivity of the vacuum, c0 is the speed of light in
vacuum, and n is the refractive index. In the aggregate, N oscilla-
tors are oscillating in phase, contributing to the excited domain.
Therefore, we should replace m by Nm and Q by NQ, obtaining
3(Nm)ε 0c03 1
τagg = = τ0 (5.3)
2(NQ )2 πν20n N
We can exploit the fact that the excited domain must exist
through a balance between attractive forces of the in-phase
oscillating dipoles (excited portion) and the thermal motion
tending to disrupt such a domain in order to find the num-
ber N of molecules contributing to the excited domain. It
turns out that the binding energy (–∆E/N) in a pair of oscil-
lators decreases the energy of the excited state, making it
more favorable. Thermal energy (kT) works against the for-

mer quantity and makes it less favorable, thereby pushing the
oscillator out of phase. Therefore,
−ΔE
= kT (5.4)
N
It has been found that ∆E = –0.24 eV. Restating Equation 5.4,
ΔE 0.24eV 1
N =− = ⋅ (5.5)
kT k T
and for room temperature,
3000K
N= = 10 (5.6)
300K
We can now use the number of molecules N contributing to
the excited domain, which we solved for in Equation 5.6, in order
to approximate the fluorescence lifetime from Equation 5.3.
T
τagg = τ0 (5.7)
3000K
Solving Equation 5.2 for m = me, Q = e, n = 1.5, and v0 = 0.75

× 1015s–1, we find that τ = 5 × 10 –9 s. This yields
300K
τagg = × 5 × 10−9 s = 5 × 10−10 (5.8)
3000K
We can also approximate the speed υ of an exciton through
a monolayer. In order to do so, we must consider the area A
covered by the exciton of width L during τagg:
A = Lexciton ⋅ v ⋅ τagg
(5.9)
Let us envision a situation in which a donor photon has a

50% probability of reaching an acceptor molecule. Under this
condition, A must be equal to the area in which just one accep-
tor molecule is present. Thus, we redefine A as Za where a is
the area covered by a donor molecule and Z is the number of
donors per acceptors and
A Za
v= =
Lexciton ⋅ τagg Lexciton ⋅ τagg (5.10)
Plugging in typical variable values when the donor is in

excess such that Z = 10,000, a = 1500 × 400 pm2, L exciton = 5
nm (at 300 K), and τagg = 5 × 10 –10 s we find that v = 2.4 km s–1.
For more information see Chapter 23 in Kuhn and Försterling’s
textbook, Principles of Physical Chemistry: Understanding
Molecules, Molecular Assemblies, Supramolecular Machines.
Certain experimental observations support a calculation of
such a fast-moving exciton. Most importantly, this theoreti-
cal number helps explain how the excited domain fluoresces
or excites another dye molecule. In this oscillation model,
charges in the excited domain change simultaneously and the
nearest-neighbor charges are always opposite; thus an attrac-
tion of nearest-neighbor charges occurs, leading to a compres-
sion of the excited domain. This compression creates a wave
that carries the exciton, which does not change shape or lose
energy until its destruction through absorption by an acceptor
or through fluorescence. This is one possible explanation of
how an exciton, or energy, is able to travel through a medium.
5.2 NANOWIRES
A nanowire is a nanostructure that has a diameter on the
scale of a nanometer and an unrestricted length. An example
of a nanowire is shown in Figure 5.5. Nanowires have been
synthesized to be as long as one millimeter (10 –3 m), but are
more commonly made to be about one micrometer (10 –6 m).
Regardless, the typical nanowire has a width-to-length ratio
Figure 5.5 A vertical silicon nanowire array on a silicon substrate. (Image provided by
Professor Hari Reehal, London South Bank University.)
of 1:1000 or more, allowing the structure to be viewed as one-

dimensional (1-D). Combined with the comparatively small
size of nanowires, the 1-D structure is responsible for the
unique properties that grossly distinguish nanowires from the
average (three-dimensional) wire.
The potential applications for nanowires are immense. As
technology continues to advance, the trend in industry is to
generate smaller but more powerful electronic devices than
previously sold on the market. As a result, computer com-
ponents such as resistors, capacitors, and circuits continue
to shrink. To link these components, increasingly small
wires are desired. There is also great interest in nanowires
from the battery industry, which believes that the increased
surface area of condensed nanowires relative to bulk phase
metals is the future of anode and cathode design. Industries
also see the potential for nanowires in such applications as
chemical composites, field emitters, and biomolecular nano-
sensor leads. Moreover, nanowires’ conductance can be con-
trolled by synthesizing the wire from any one of a variety of
substrates, allowing for nanowires to have more specialized
roles. As a result, nanowires are categorized as (in order of
increasing conductance ability) insulators (SiO2, TiO2), semi-
conductors (Si, GaN, InP), or metallics (Au, Cu, Pt, Ni, etc.).
With such major applications, research relating to the gen-
eral properties and synthetic strategies of nanowires is in
high demand.
5.2.1 Basic Quantum Mechanics of Nanowires

It is well understood that energy is quantized. That is to say,
substances are limited to having a discreet amount of energy
at any given time. Since the difference in energy between
allowed energy states is typically small, large substances can
neglect the effect of quantum mechanics. However, quantum
mechanics plays an increasing role as devices decrease in size.
Nanowires’ small sizes result in quantum mechanics having
a significant effect on their behavior and properties. This is
because quantum mechanics results in substances only being
able to absorb or release specific values of energy such that
they can maintain their allowed value of total energy. In semi-
conductors and insulators, the differences in energy levels are
often significant and can inhibit electron flow and lower con-
ductance. Metals are different from insulators and semicon-
ductors because for metals the differences in energy states are
often negligible. This property of metals results in nearly free
electron flow from metal molecule to metal molecule.
Since most applications of nanowires are related to elec-

tronics, the quantum effect on nanowires’ conductance is of
particular interest. Quantum mechanics needs to relate the
effect of the discrete energy levels intrinsic of the nanowire
molecules to the nanowire’s overall conductance. The equa-
tion that is used to relate conductance with discrete energy
levels is the von Klitzing equation (Klitzing et al., 1980):
h
R=
ne 2 (5.11)
where R is resistance and is measured in ohms (Ω), h is Planck’s

constant, e is the charge of an electron, and n is an integer
(n = 1, 2, 3, …). Conductance is simply equal to the inverse of
resistance. Therefore, a nanowire has, at n = 1, a conductance
of 0.0387/kΩ.
5.2.2 Conductivity
As already mentioned, quantum mechanics has great control
over conductivity. Yet conductivity varies significantly
depending on the composition and the physical properties of
the nanowire (width and length). A nanowire’s extremely large
length-to-width ratio results in high levels of resistance. This
resistance is compared to that of a bulk phase wire, composed
of the same substances, whose resistance is much lower due to
the much smaller length-to-width ratio. The nanowire’s nar-
row width contributes in a number of ways to increasing resis-
tance. For one, decreased width increases the relative effect
of the wire’s surface molecules. The molecules at the surface
of the wire are much less tightly bound than molecules found
in the interior of the wire. Also, surface molecules are in con-
tact with far fewer of the wire’s molecules than those found in
the interior of the wire. The overall consequence is that mol-
ecules found at the surface of a wire have greater difficulty
propagating electrons to other wire molecules. Therefore, high
surface-area-to-volume ratios tend to increase the resistance.
The basic concept of width effects on resistance is related to
what is known as the mean-free path.
Mean-free path relates to the distance an electron travels
between subsequent collisions with other moving particles.
Collisions are not favorable for electrons in nanowires since
the electron can be deflected in any direction (Figure 5.6) and
impact the wire’s forward flow of electrons. Collisions occur
most frequently when the width of the nanowire is smaller
than the mean-free path, resulting in frequent collisions and
w<λ
(a)
w>λ
(b)
Figure 5.6 A schematic demonstrating the effects of mean-free path. The open circle
indicates the electron and the solid black circles represent larger atoms. w is the width of
the wire, and λ is the mean free path of the conduction electron. (a) Diffusive transport
occurs when the width of the wire is smaller than the mean-free path of the electron, and
(b) ballistic transport occurs when the width of the wire is greater than the mean-free path.
impaired wire conductivity. In the opposing case, a nanowire

is said to be a “ballistic transporter” when the mean-free path
is less than the width. The result is electrons always moving
in a forward direction. Ballistic transporters (typically metal
nanowires) have very high conductance values compared to
those whose wire widths fail to exceed the mean-free path
(Takayanagi et al., 2001).
5.2.3 Nanowire Synthesis

Two general concepts exist for nanowire synthesis, each
with their own unique synthetic strategies. The “top-down”
approach is one such concept. In this approach, wires are
synthesized starting from a bulk substrate and chiseled down
until only a field of nanowires remain. Nanolithography and
electrophoresis are examples of the top-down techniques
employed to make nanowires. Although the theory of the
top-down technique is sound, our current resources limit the
approach. It is difficult using current instruments to synthe-
size ideal one-dimensional nanowires. The instruments and
techniques are currently limited by the length and width
of the nanowires they can synthesize as well as the general
accuracy they exhibit when working on a substrate. For these
reasons, the top-down approach is presently less commonly
employed than its alternative.
The “bottom-up” approach synthesizes nanowires by con-
tinuously extending a thread of bound molecules. Various tech-
niques have developed using the bottom-up approach including
suspension-based techniques, where a nanowire is elongated
in a vacuum using chemical etching or ion-bombardment
techniques, and vapor-liquid-solid (VLS) techniques, where a
nanowire is synthesized via a catalyst converting a vapor into

a solid below the catalyst, continuously elongating the chain
of solid molecules (Crossland et al., 2007). Solution-based bot-
tom-up techniques hold particular promise since they allow
scientists to synthesize nanowires in much larger numbers
than alternative methods. One particularly intriguing solution-
based technique uses copolymers to form templates inside
which nanowires can be synthesized (Figure 5.7). Certain copo-
lymers are ideal for template formation because, at the correct
temperature and electric field strength, they form evenly spaced
hexagonal pores. Once the substrate is added, these pores serve
as the locations for nanowire synthesis. Mild reagents can then
be used to degrade the template and isolate the nanowires.
5.2.4 Summary
Nanowires have great potential. As the field grows, we will
begin to see new synthetic techniques capable of producing
A B
500 nm 50 nm
C D
10 nm 100 nm
Figure 5.7 Gold nanowires were synthesized using a solution-based bottom-up tech-
nique utilizing a copolymer template. Images were taken using a scanning electron micro-
scope. (Reprinted with permission from Chen, Jingy, Benjamin J. Wiley, and Younan Xia.
“One-Dimensional Nanostructures of Metals: Large-Scale Synthesis and Some Potential
Applications.” Langmuir 8, 2007: 4120–4129. © 2007 American Chemical Society.)
highly elongated nanowires in large quantities. These tech-

niques will be exceedingly precise, allowing for wire shape
and function to be maintained during the entirety of the
nanowire synthesis. As the study of nanowires continues to
progress, new imaging devices will be forced to develop in
parallel to nanowire synthetic techniques. These new devices
will allow nanowires to be seen in higher magnification and
clarity, allowing for more accurate analysis of experimental
results. Studies will continue to be done on synthesized nano-
wires to help define the nanoparticle’s unique properties so
it can be better applied to products in the industry including
batteries, computers, and chemical composites.
5.3 CARBON NANOTUBES

Carbon nanotubes are constructs of carbon that have diam-
eters on the order of a few nanometers and lengths that can
be significantly longer. They are extremely strong; the tensile
strength of a single-walled carbon nanotube is many times
stronger than steel. They also have unique electrical and opti-
cal properties and are efficient conductors of heat. These prop-
erties have made the study of carbon nanotubes surge in recent
years (Dai et al., 2002).
5.3.1 Carbon Nanotube Structure

Carbon nanotubes are a member of the fullerene family of car-
bon structures; they may contain hexagonal, pentagonal, and
sometimes heptagonal rings of pure carbon. It may be help-
ful to imagine a flat “sheet” of such rings that can be rolled
up to make a single-walled nanotube. Regardless of the num-
ber of carbons within a ring, each shares a bond with three
adjacent carbons. This makes every carbon within the tube
sp2 hybridized. Carbon nanotubes are often capped with a car-
bon hemisphere that is also composed of sp2 hybridized car-
bons in rings. Carbon nanotubes can come in single-walled or
multi-walled form. The multi-walled nanotube structure can
be thought of as concentric single-walled nanotubes encom-
passing one another.
Because differing carbon nanotubes may consist of similar sp2
hybridized carbons, one might assume they all will have simi-
lar properties. This is an incorrect assumption. Differences in
how the carbon atoms are structurally organized play a signifi-
cant role in determining the properties of the carbon nanotube
as a whole. For example, one way of organizing the carbons can
make the nanotube metallic and conducting, whereas another
Zig-zag
Arm-chair
Ch
ma2
θ
a2
a1
Figure 5.8 Schematic diagram showing how a hexagonal sheet of graphite can be rolled
to form a carbon nanotube. The properties of the nanotube will depend on the chiral vector
and the chiral angle. The unit vectors are shown in the lower left-hand side of the diagram.
organization scheme will make it a semiconductor. Because of

this, the atomic structure of a nanotube is often described in
terms of its chirality, which can be broken down into a chiral-
ity vector, C h, and a chiral angle, θ (Figure 5.8).
The length of the chirality vector is related to the diameter
of the nanotube it describes and can be easily determined if the
interatomic spacing of the carbon atoms is known. The inter-
atomic spacing of carbon atoms is often a known parameter
because it is the same for all nanotubes with a similar lattice, or
unrolled “sheet” of rings. An equation for describing the chiral
vector is given in Equation 5.12 (Thostenson et al., 2001).
C h = na1 + ma2 (5.12)
In this equation n and m are the number of steps in the direc-

tion of the unit vectors, a1 and a2, in their respective direc-
tions across the sheet of unrolled carbon rings. Figure 5.8 is a
graphical depiction of this along a graphene sheet, composed
of only hexagonally ordered carbon atoms. The chiral angle,
θ, is a measurement of the resulting twist in the “rolled up”
nanotube and is defined as the angle between the chiral vector
and the unit vector a1. This angle can be determined by relat-
ing the number of steps n and m as shown in Equation 5.13
(Saito et al., 1998).
cosθ =
(2n + m)
(
⎡2 n2 + m2 + nm ⎤
⎣ ⎦ ) (5.13)
Nanotubes with chiral vector steps n = m are known as “arm-

chair” nanotubes and have a chiral angle of 30°. Nanotubes
with one of the step integers equaling zero are referred to
as “zig-zag” nanotubes and have a chiral angle of 0°. Both of
these are achiral nanotubes, because their mirror images are
identical to the original structure. In all other cases, where
n ≠ m, the nanotubes are chiral, with a chiral angle between
0° and 30°.
5.3.2 Some Properties of Nanotubes

Interestingly, the relationship of steps, n and m, can be used to
determine if the nanotube will be conducting or semiconduct-
ing. If n – m is a positive multiple of three, or zero, the nano-
tube has no band gap, and therefore is a conductor. If n – m is
not a positive multiple of three, or zero, then the nanotube will
have a small band gap and therefore is a semiconductor.
The ability of carbon nanotubes to conduct heat is depen-
dent on temperature and is comparable to that of diamond or
a graphite sheet. The peak thermal conductivity for a single-
walled nanotube is reached at around 100 K and is around
37,000 Wm–1K–1. From there the thermal conductivity drops as
the temperature increases to around 3000 Wm–1K–1 at a tem-
perature of 400 K.
The precise values for mechanical and elastic properties of
carbon nanotubes are currently disputed. They are thought to
be the strongest and stiffest materials known to man in terms
of tensile strength and elastic modulus. It has been shown
that they are many times stronger than steel and that they can
return to their original state after being bent. Nanotubes have
also been found to be stable at very high temperatures, nearly
3000 K in a vacuum and 700 K in air.
5.3.3 Methods for Growing Nanotubes

There are three primary methods for nanotube production:
arc discharge, laser ablation, and chemical vapor deposition
(Baddour et al., 2005). Arc discharge uses a high-voltage arc to
vaporize a carbon source. This frees carbon, which then accu-
mulates into nanotubes in the growth chamber. Following
accumulation the nanotubes are collected and purified. Laser
ablation is similar to the arc discharge method, but it makes
use of a high-energy laser, instead of a current, to vaporize the
carbon source. Chemical vapor deposition relies on the decom-
position of a carbon-containing gas in the presence of a metal
catalyst, such as iron, cobalt, or nickel, which helps facilitate
carbon accumulation and nanotube growth. Each method has
many variations with their own strengths and weaknesses

that should be considered when choosing which to use.
5.3.3.1 Arc Discharge

Two carbon rods are set up as an anode and cathode and are
placed in a closed growth chamber. The chamber is filled with
an inert buffer gas such as helium, argon, nitrogen, or hydro-
gen and kept at relatively low pressures. Interestingly, the
inert gas in the chamber has been found to affect the diameter
of grown nanotubes while the pressure in the growth chamber
does not. Keider and Waas (2004) found that when mixtures
of helium and argon are used to fill the chamber, the molar
fraction of argon determines the nanotube diameter. Once the
chamber is filled with gas, a direct current is applied until a
stable electric arc is formed between the carbon rods. Once the
anode reaches a critical temperature, it vaporizes into a gas-
eous carbon source. The gas in the chamber causes the carbon
to deposit. The structures found in these deposits are nano-
tubes, which need to be purified after collection. This method
has been used for the creation of both single-walled and multi-
walled nanotubes. In order to produce substantial amounts of
single-walled carbon nanotubes a metal catalyst needs to be
added to the electrodes.
Unfortunately, once collected, the nanotubes must be
manipulated onto substrates for further use, which can be dif-
ficult for large quantities of such small structures. Arc dis-
charge also requires a large amount of power and is operated
at a high temperature, making it an unfavorable method for the
production of large quantities of nanotubes.
5.3.3.2 Laser Ablation

Laser ablation uses a high-energy laser beam to vaporize a car-
bon source. The carbon source is sealed off in a quartz tube and
heated to temperatures ranging from a few hundred degrees to
just over a thousand degrees Celsius. The tube is filled with an
inert gas that flows toward a cooled collector at its end. As the
laser vaporizes the carbon source, the flow of inert gas carries
it down the tube, where it deposits. The nanotubes are then
collected from the deposits and purified.
In both methods the vaporized carbon is released as low
molecular weight carbon, which coalesces to form larger
molecular weight structures. The carbon begins to condense
very quickly after vaporization. Using metal catalyst parti-
cles doped into the carbon source prevents carbon from clos-
ing into a cage structure and decreases the amount of carbon
deposited in non-nanotube form by aggregating with the car-

bon quickly. This results in more single-walled nanotubes
being formed when a metal catalyst is used. Nanotube length
can be controlled by changing the time the carbon spends in
the high temperature area of the growth chamber, which can
be controlled by using heavier or lighter gases and altering the
flow rate. Research is undergoing into changing growth condi-
tions, which may make laser ablation a viable option for the
large-scale production of nanotubes.
5.3.3.3 Chemical Vapor Deposition

Chemical vapor deposition relies on the decomposition of a
carbon-containing gas, such as methane, ethane, benzene, acet-
ylene, or ethylene, which then aggregates on a catalyst or sub-
strate to form nanotubes. Studies have shown that gases with
saturated carbon bonds produce structures with more walls
than those grown from unsaturated gases. Therefore, methane
and ethane are preferred for single-walled carbon nanotube
growth, and ethylene, benzene, and acetylene are preferred for
multi-walled carbon nanotube growth. Because high tempera-
tures create more high-quality crystals by reducing defects and
increasing crystallization, the source gas must also be one that
does not thermally decompose into amorphous carbon and
reduce the purity of nanotubes. The amorphous carbon can
accumulate on the aggregation sites and inactivate them from
forming nanotubes. The aggregation sites are typically metal
catalysts on a substrate, but can be a dust or gas if raw material
not on a substrate is desired. The size of the catalyst on the sub-
strate determines the diameter of the grown nanotubes.
The nanotubes are similar to those produced by other meth-
ods; however, because the carbon is in the gas phase already it
does not have such a limited carbon source size as laser abla-
tion or arc discharge methods. There is also more control over
the aggregation sites where the carbon nanotubes are grown,
meaning that they can be grown as a raw material or on a
substrate. Unfortunately, the lower temperatures of chemical
vapor deposition result in nanotubes that are not as high qual-
ity crystalline as those produced in the laser ablation and arc
discharge methods, although techniques have been designed
to help overcome this drawback and make this method viable
for the large-scale production of high-quality nanotubes.
5.3.4 Catalyst-Induced Growth Mechanism

Nanotube growth does not start immediately after carbon
gas comes into contact with the metal catalyst. Carbon first
dissolves into the metal catalyst particle. Once the particle is

saturated, the carbon atoms assemble into an sp2 hybridized
structure at a less reactive surface. The carbon then moves out
of the metal, adding more carbon atoms to the nanotube grow-
ing on the less reactive surface at the edges of the particle.
This is why the diameter of the metal particle determines the
diameter of the nanotube being grown.
There are two suggested growth mechanisms for nanotube
growth from a metal catalyst. The first method is root growth
in which the nanotube grows upward from the metal catalyst
particle, which stays attached to the substrate. The second
suggested growth mechanism is tip growth in which the nan-
otube grows from the catalyst downward into the substrate.
This pushes the catalyst up and detaches it from the substrate
(Sinnot et al., 1999). The two mechanisms are illustrated in
Figure 5.9.
5.4 QUANTUM DOTS

Quantum dots are inorganic semiconductor nanoparticles,
typically 2–10 nm in size. The excitons (electron-hole pairs)
of quantum dots are confined to three dimensions. Quantum
wires are confined to two dimensions, and quantum wells
Metal catalyst
Solid support Solid support Solid support
(a)
Solid support Solid support Solid support
(b)
Figure 5.9 The two suggested models of carbon nanotube growth through chemical
vapor deposition onto a metal catalyst. In both models, the carbon grows around the edges
of the metal catalyst.
Inorganic core
Shell material
Organic capping agent
2–10 nm
Figure 5.10 Structure of a quantum dot. The size and nature of the inorganic core and
the shell material can be varied. Quantum dots are often capped with an organic thin film.
are confined to one dimension. As the confining dimensions

decrease toward a limit, the energy gap starts to change with
it. This tends to happen at the nanoscale, thus the small size of
quantum dots gives them optical and electronic properties in
between those of bulk semiconductors and discrete molecules
(Brus, 2007). Each dot is a mesoscopic entity with individual
properties, differentiating it from a colloidal particle.
Quantum dots typically consist of a core, shell, and final
coat, as shown in Figure 5.10. They are characterized by the
nature of each of these layers, their size and aspect ratio, their
quantum efficiency in optical materials, and their coercivity
in magnetic materials. Properties of an ensemble of quantum
dots are additionally determined by particle size distribution
and differences in morphology within the ensemble.
5.4.1 Optical Properties

Due to quantum confinement, quantum dots have character-
istic optical properties that can be used to optimize signal-to-
background ratios. High molar extinction coefficients (of the
order of 0.5 – 5 × 106 M–1 cm–1) make quantum dots particularly
bright probes when light intensities are severely attenuated by
scattering and absorption—their absorption rates are faster
than those of organic dyes at the same excitation photon flux.
Due to the increased rate of light emission and the large absorp-
tion cross-section of quantum dots, they have been found to be
10–20 times brighter than organic dyes.
Additionally, the broad absorption profiles of quantum dots
allow simultaneous excitation of multiple wavelengths; thus
molecular and cell targets can be tagged with different col-
ors. The fluorescence emission wavelengths can be continu-
ously tuned from 400 to 2000 nm by varying particle size
(c)
Intensity (a.u.)
(b)
1/e
(a)
0 t1 t2 50
Time (ns)
Figure 5.11 Comparison of the excited state decay curves (monoexponential models)
between quantum dots and common organic dyes. Line (a) indicates excitation. Lines (b) and
(c) indicate the decay curves for the organic dye and quantum dot, respectively. A measure of
the fluorescence lifetime is indicated by time t1 (for the dye) and t2 (for the quantum dot).
and chemical composition. For example, if one compares the

emission spectra of rodent skin and that of quantum dots
obtained under the same excitation conditions, one finds that
the quantum dot signals can be shifted to a spectral region
where autofluorescence is reduced in order to improve signal-
to-background ratios. Moreover, quantum dots are remarkably
resistant to photobleaching and are thus well suited for con-
tinuous tracking studies over a long period of time. Signal-to-
background ratios can be further improved by time-domain
imaging, which is used to separate quantum dot fluorescence
from background fluorescence due to the longer excited states
of quantum dots (Figure 5.11).
Quantum dots exhibit Stokes shift—the absorption of light at
one-wavelength leads to emission at a slightly lower energy, as
seen in Figure 5.3. This difference between the excitation and
emission peaks can be as large as 300–400 nm, making the
quantum dot signal easily recognizable above the background.
5.4.2 Synthesis of Quantum Dots

5.4.2.1 Precipitative Methods
This method is similar to colloidal synthesis in that it involves
supersaturation of a solution followed by crystallization.
Like traditional chemical processes, colloidal semiconduc-
tor nanocrystals are synthesized from precursor compounds
dissolved in solutions. The precursors are heated to tempera-

tures high enough to convert them to monomers. Once the
monomers reach a high enough supersaturation level, the
nanocrystal growth starts with a nucleation process. The tem-
perature is a critical factor in determining optimal conditions
for crystal growth—it must be high enough to allow for the
rearrangement and annealing of atoms during the synthesis
process while being low enough to promote crystal growth.
Furthermore, the size of the crystal depends on temperature,
i.e., it is larger at higher temperatures. Another critical factor
is the monomer concentration. At high monomer concentra-
tions, the critical size of the crystal is relatively small, result-
ing in the growth of nearly all particles and thus facilitating
monodispersity. When the monomer concentration is depleted
during growth, the critical size becomes larger than the aver-
age size present.
The preparation of CdS quantum dots requires the controlled
nucleation of CdS in dilute aqueous solutions of cadmium
sulfate and ammonium sulfide. The dynamic equilibrium
between solvated ions and solid CdS in acetonitrile as a solvent
in the presence of a styrene or maleic anhydride copolymer
allows the preparation of stable CdS nanoparticles (average
size 3.4–4.3 nm). To obtain highly monodispersed nanopar-
ticles, post-preparative separation techniques like size exclu-
sion chromatography or gel electrophoresis are used.
Weller et al. (1985) injected phosphine (PH3) into solutions
containing metal salts to make particles of Zn3P2 and Cd3P2.
The particle size can be controlled by varying temperature and
phosphine concentration. Even though this method is cheap,
it has several disadvantages. It lacks reaction control, which is
problematic in large-scale syntheses. Furthermore, important
semiconductors like GaAs or InSb are not available for use in
this method as they are air and/or moisture sensitive.
5.4.2.2 Reactive Methods in High-

Boiling-Point Solvents
Solutions of dimethylcadmium, (CH3)2Cd (in tri-n-octylphos-
phine, TOP), and tri-n-octylphosphine selenide (TOPSe) in
hot tri-n-octylphosphine oxide (TOPO) at 120–300°C produce
TOPO-capped nanocrystals of CdSe. Hydrophobic organic
molecules like TOPO serve as a reaction medium and also
coordinate with unsaturated metal atoms on the QD surface to
prevent the formation of bulk semiconductors.
The TOPO method is one of the most widely used meth-
ods of quantum dot synthesis as it has several advantages
over other methods. The size of the dots can be controlled

by temperature, i.e., they are larger at higher temperatures,
and monodispersity can easily be achieved. Additionally, the
yields of the reactions are usually high, producing hundreds
of milligrams in a single experiment. Moreover, this method is
readily adapted to the production of core-shell structures with
materials having high quantum efficiencies.
5.4.2.3 Gas-Phase Synthesis of

Semiconductor Nanoparticles
This method involves the atmospheric or low-pressure evapora-
tion of either powders of the pre-formed semiconductor, or co-
evaporation of two elemental components (like Zn metal and
sulfur). It is not a very favorable method due to the large size
distribution of the particles (10–200 nm) and the tendency of par-
ticles to aggregate due to the absence of a surface-capping agent.
5.4.2.4 Synthesis in a Structured Medium

A matrix is used to define a reaction space—it provides a meso-
scopic reaction chamber in which the crystal can grow only to a
certain size. The properties of the system control the properties
as well as the size of nanoparticles. Substances typically used
as matrices are layered solids, micelles/microemulsions, gels,
polymers, and glasses. This technique is often used in biologi-
cally related processes. An example of this is the use of empty
polypeptide cages found in the iron storage protein ferritin in
the synthesis of bio-inorganic nanocomposites of CdS-ferritin.
5.4.3 In Vivo Molecular and Cell Imaging

Due to their optical properties as discussed earlier, quantum
dots are better than organic dyes for the purpose of in vivo
molecular and cell imaging. Researchers have achieved real-
time visualization of single molecule movement in single living
cells using quantum dots. Their high electron density allows
correlated optical and electron microscopy studies of cellular
structures. Quantum dot bioconjugation can be achieved by
passive absorption, multivalent chelation, or covalent bond
formation as shown in Figures 5.12 and 5.13 (Gao et al., 2005).
Streptavidin-coated quantum dots preferentially bind to
biotin with a K D of approximately 10 –14 M, one of the stron-
gest covalent bonds known in nature. Such bioconjugation is
widely used for molecular and cellular imagining. An appli-
cation of this is seen in an experiment conducted to identify
β-sheet aggregate-binding ligands using one-bead one-com-
pound screening. The ligands under study were each attached
C NH
OH
H
C O
QD capping
OH
OH
O
C
O
ligand TOPO
O
C
NH
C
O
NH
O
C
C
O
NH
C O OH
O=P
C
O=P
P
O
O=
O
P
C OH
C OH
=
O
O
=P
P
PEG
O
O=
O O
C NH Quantum C NH (–CH2–CH2–O–)n
O=P
O=P
dot
O
O=
C O P O=P O
H O C CH
=
O
P
O
=
O=
O=
O=P
O
P
C
C
P
P
OH
OH O
O
C
C
Amphiphilic
NH O
NH
C
O OH polymer coating
O
C
O
OH
C OH
C
O
O
C
OH
HN
Affinity ligands: antibody, peptide,

small-molecule drug, inhibitor, etc
Figure 5.12 The structure of a multifunctional quantum dot probe. Schematic illustra-
tion showing the capping ligand TOPO, an encapsulating copolymer layer, tumor-targeting
ligands (such as peptides, antibodies, or small-molecule inhibitors), and polyethylene glycol.
(Reprinted from Gao, X.; Yang, L.; Petros, J. A.; Marshall, F. F.; Simons, J. W.; Nie, S. In vivo
Molecular and Cellular Imaging with Quantum Dots. Current Opinion in Biotechnology 2005,
16, 63–72. With permission from Elsevier.)
to a bead. When treated with the biotinylated peptide, only

one of these three ligands binds to it. The streptavidin-coated
quantum dots preferentially bind to this peptide-ligand com-
bination, and thus the ligand is easily detected due to the flu-
orescence exhibited by the quantum dots. Analysis by mass
spectroscopy reveals the identity of the ligand.
5.5 LANGMUIR-BLODGETT FILMS

First demonstrated by Katharine Blodgett in 1934 (Blodgett,
1935), the Langmuir-Blodgett (LB) technique has been by far
the oldest and most extensively studied organic nanofilm for-
mation method. Essentially, a monolayer at the air-water inter-
face (the Langmuir film) is mechanically transferred to a solid
substrate generating the LB film. Although interest waned
until the 1970s, the potential application of LB films in the
field of optics and material science has revived LB research.
(a)
COOH
2
NH
H
CO
O
CO
O H
N
Quantum H EDAC Quantum
COOH COOH + H2N 2
dot dot
NH2
Antibody
CO
O H
O
CO
COOH
(b)
NH2
OH
CO
CO
OH
NH2 Quantum NH2 SH SMCC Quantum

HS
dot + SH HS dot
CO
OH
OH
NH2
CO
Antibody fragments
(c)
OH
CO
CO
OH
Quantum
Quantum Adaptor dot
+
dot protein
Antibody
CO
OH
OH
CO
(d)
COOH
Binding site
OH
CO
OH
CO
O O
O
COOH Quantum O Quantum
CONH +
dot N N
2– dot
His6
O
His-tagged
CO
OH
O
OH
COOH
CO
Ni-NTA peptide
Figure 5.13 Methods for conjugating QDs to biomolecules. (a) Traditional covalent
cross-linking chemistry using EDAC (ethyl-3-dimethyl amino propyl carbodiimide) as a
catalyst. (b) Conjugation of antibody fragments to QDs via reduced sulfhydryl-amine cou-
pling. SMCC, succinimidyl-4-N-maleimidomethyl-cyclohexane carboxylate. (c) Conjugation
of antibodies to QDs via an adaptor protein. (d) Conjugation of histidine-tagged peptides
and proteins to Ni-NTA-modified QDs, with potential control of the attachment site and
QD:ligand molar ratios. (Reprinted from Gao, X.; Yang, L.; Petros, J. A.; Marshall, F. F.; Simons,
J. W.; Nie, S. In vivo Molecular and Cellular Imaging with Quantum Dots. Current Opinion in
Biotechnology 2005, 16, 63–72. With permission from Elsevier.)
This was fueled by the results that showed an LB film com-

prised of a dye to be the first multilayer film to produce inter-
esting nonlinear optical (NLO) effects. Nowadays, research
toward the synthesis of tailor-made molecules with intrinsic
nonlinear susceptibilities and their incorporation into LB
films is flourishing.
5.5.1 Langmuir Films

The first step of LB multilayer assembly is to form a Langmuir
film by dissolving surface-active organic molecules, such as
surfactants or amphiphiles, in a nonpolar, volatile solvent and
then spreading the solution onto a polar liquid surface, usu-
ally water. Once the volatile solvent evaporates, the remaining
amphiphilic molecules are oriented at the air-water interface
so that the hydrophilic headgroups are buried in the bulk
aqueous phase while the hydrophobic tailgroups are directed
upward into the air. However, for this orientation to occur, two
conditions must be met: (1) the hydrophobic tail must be long
enough to prevent its dissolution in the aqueous media, and
(2) the hydrophilicity of the headgroup must be strong enough
to prevent the formation of thicker multilayer films at the
interface or evaporation of the surfactant molecules. Moving
barriers on either side sweep the water surface and force the
amphiphiles to pack, forming an ordered, compressed mono-
layer of the Langmuir film (Figure 5.14).
Air
Water
Π Π
Air
Water
Π Π
Air
Water
Π Π
Air
Water
Figure 5.14 A Langmuir film. Amphiphilic molecules in chloroform solvent are depos-
ited on water surface. The Langmuir film is formed by confining the amphiphiles at the air-
water interface with movable barriers.
The packing density of a Langmuir film is crucial in deter-

mining the final structure of an LB film. As we have already
seen, surface pressure is a measure of the change in the sur-
face tension of a pure liquid due to the presence of a surfac-
tant. Basically, it is defined as the difference between the
surface tension of the pure water and the surface tension of
the aqueous solution containing amphiphilic solutes. This
phenomenon is especially important in Langmuir films. If the
number of molecules comprising the Langmuir film is known,
then surface pressure as a function of the area occupied by
each molecule may be studied. The resulting plot is known as
a pressure-area (Π-A) isotherm, and its shape is unique to the
molecule used to form the film. Figure 5.15 illustrates a typical
isotherm for a long-chain carboxylic acid [CH3(CH2)nCOOH] at
the air-water interface. At constant temperature, the Π-A iso-
therm is given by the changes in surface tension as the mono-
layer is compressed.
The Π-A isotherm may consist of several regions, although
molecules usually exhibit only a few of them. When the mol-
ecules are first deposited onto the film, the monolayer behaves
most like a 2-D gas (G), shown by Equation 5.14, since no exter-
nal pressure is applied.
ΠA = k BT
(5.14)
Π (mN/m)
LC
LE
G
0.1 0.5
A (nm2/molecule)
Figure 5.15 A simplified Π-A isotherm for a simple long-chain fatty acid. Extrapolating
the slope of the S phase to zero pressure enables one to obtain the molecular area per mol-
ecule at zero pressure (see Example 5.2).
In this equation, Π is the surface pressure, A is the molecular

area, kB is the Boltzmann constant, and T is the temperature.
In the G phase, all interactions with the water are attractive, so
even the hydrophobic tail is in intimate contact with the water.
Transition to the liquid-expanded phase (LE) occurs as mono-
layer compression begins, and the tail starts to lift away from
the surface. The transition states form a plateau because the
existence of two phases gives the monolayer only one degree
of freedom, and constant temperature gives constant pressure,
Π (Knobler 1990).
The liquid-condensed phase (LC) forms with further com-
pression. The hydrocarbon chains are highly aligned and
a high number of trans conformations appear, indicating
the achievement of long-range order due to van der Waal’s
interactions. Finally, application of great external pres-
sure leads to the ordered solid (S) phase, and the area per
molecule is comparable to the closely packed chains of the
amphiphile’s 3-D crystal. Further compression leads to the
collapse pressure, Πc, where the molecules are ejected ran-
domly throughout the film.
Molecular dimensions can be obtained through the Π-A
isotherm by extrapolating the slope of the S phase to zero
pressure to get the molecular area per molecule at zero pres-
sure, A0. A0 can be compared to theoretical values of the cross-
sectional area of hydrocarbon chains, and a close correlation
indicates that a closely packed monolayer has been formed
with the hydrophobic tails oriented normal to the surface.
Furthermore, the Π-A isotherm gives information on the sta-
bility of the monolayer at high pressures.
Example 5.2 Determining the Limiting

Surface Area per Molecule
Use Figure 5.15 to determine the area per molecule for the amphiphile
described by the isotherm.
Solution By extrapolating the slope of the S phase to zero pressure (the

dashed line in Figure 5.15), one obtains an area of 0.2 nm2 for each
molecule.
5.5.2 Langmuir-Blodgett Films
The LB film is subsequently formed by the immersion of a
substrate into the water to break the Langmuir film and
transfer the monolayer onto the substrate. Additional immer-
sions result in the fabrication of the multilayer film, whose
Air
Substrate
Water
(b)
Air
Substrate
Air
Water
Water
(a) (c)
Figure 5.16 (a) The LB method is used to transfer a film at the air-water interface to a
solid substrate. (b) The downstroke results in the monolayer being transferred to the hydro-
phobic substrate. (c) Withdrawing the substrate generates a bilayer as a Y-type film.
thickness depends largely on the molecular chain length and

number of dippings. The LB deposition technique is illus-
trated in Figure 5.16.
Deposition of tightly packed molecular monolayers with
unilateral molecular directionality normal to the substrate
can be induced through annealing. Annealing involves
repeated compression and then expansion of the Langmuir
film. Studies have found that the calculated area per mole-
cule obtained through the Π-A isotherm actually decreases
under such a procedure, leading to a closely packed and well-
aligned monolayer.
Film deposition onto the hydrophilic or hydrophobic sub-
strate occurs only when the direction of the motion coincides
with the meniscus curvature at the solid-liquid interface.
If a hydrophobic substrate is used, the meniscus will curve
downward upon immersion into the subphase, resulting in
deposition on the downstroke and adhesion of the tails to
the hydrophobic surface. If this is the first immersion, subse-
quent monolayer deposition behavior determines the LB film
type. If withdrawal of the substrate results in subsequent layer
formation in a head-to-head fashion, then a Y-type LB film is
formed. On the other hand, if monolayer transfer occurs only
upon insertion of the hydrophobic substrate, then an X-type
LB film is formed. A third type of LB film, Z-type, is formed
on a hydrophilic substrate where the amphiphilic headgroup
adheres to the substrate and transfer occurs only upon with-
drawal of the substrate.
The Y-type film is the most common type of LB film formed.

Since monolayer transfer occurs on both the down- and
upstrokes, multilayer films may also deposit on hydrophilic
substrates with the exception that no monolayer adsorbs in
the first insertion. Due to the head-to-head and tail-to-tail con-
formation, the film formed, if composed of only one kind of
amphiphilic molecule, is completely centrosymmetric. Such
a film possesses no dipole moment, rendering the assembly
useless for NLO applications. However, if fatty acid spacers,
such as stearic or arachidic acid, are incorporated between the
other amphiphilic species, a net dipole moment may result.
This procedure requires two separate LB compartments for
each type of molecule. The substrate is inserted through the
amphiphile of interest, moved to the other compartment under
water, and withdrawn out of the fatty acid spacer Langmuir
film. Research has shown a relatively large SHG signal for
such a polar-ordered arrangement. Conversely, X- and Z-type
films are inherently arranged in an acentric manner with a
net dipole moment due to the head-to-tail molecular confor-
mation. In fact, studies by some researchers have found that
Z-type LB films generate a strong SHG signal.
In spite of the virtues of the LB technique, it has its
shortcomings. One limitation is the unstable nature of the
adsorbed films. In practice, although the transfer process
occurs very slowly, the layers are not always transferred
onto the substrate as would be desired. After deposition, the
molecules have been shown to rearrange to more stable con-
formations, which could lead to loss of order and NLO activ-
ity. Another disadvantage is the occurrence of chromophore
randomization. It has recently been shown that intermix-
ing between LB layers is present. This produces structural
defects and chromophore tilting that becomes amplified
when the number of LB layers is increased. Thus, it is not sur-
prising that experimental results show that the SHG signal
resulting from the head-to-head, polar-ordered (Y-type) LB
films decreased with thickness (Johal et al., 1999). Moreover
the fatty acid spacers used for acentricity could lead to even
more disorder in the LB film. Stearic acid has been found
to form two-dimensional crystalline structures that result
in a heterogeneous monolayer filled with defects. Finally,
staggered molecular structures may form due to repulsive
interaction between strong dipoles. The effect of molecular
interactions on the order of the LB film are still sparsely
understood, and current research addresses the effects of
dipolar interactions on film formation. With the limitations
of the LB technique and other aforementioned methods, the

introduction of the electrostatic self-assembly method was
readily welcomed.
5.6 POLYELECTROLYTES
Polyelectrolytes are polymers, or chains of molecules, that
contain free ions that make them electrically conductive.
Soluble in water, polyelectrolytes become charged when in
solution, and are often countered by a salt ion of opposite
charge. Figure 5.17 shows some common polycations and
polyanions. The amount of charge on a polyelectrolyte deter-
mines whether it is classified as strong or weak. Strong poly-
electrolytes are fully soluble, whereas weak polyelectrolytes,
with fractional charge, are only partially soluble. Typically,
n
HN
SO2
N
N
n
HO O
COO – N n
Na + CH2 CH
OH n SO3– Na+ H
(a) (b) (c) (d)
CH2 CH2
n
n
+ Cl–
N
NH3+ Cl– H3C CH3
(e) (f)
Figure 5.17 Some common polyelectrolytes. Polyanions are (a) PAZO (poly{1-[4-
(3-carboxy-4-hydroxyphenylazo)benzenesulfonamido}-1,2-ethanediyl, sodium salt]), (b)
PAA [poly(acrylic acid)], and (c) PSS [poly(styrenesulfonate)]. Polycations are (d) PEI
[poly(ethylenimine)], (e) PAH [poly(allylamine hydrochloride)], and (f) PDDA [poly(diallyldimethyl
ammonium chloride)].
polyelectrolyte conformation is “flat” because the charges on

a linear polymer repel each other due to Coulombic repulsion.
With the addition of salt ions, however, this conformation can
be altered to a more coiled or collapsed state. This ability to
change polyelectrolyte conformation can change their func-
tion and is very useful in materials chemistry.
Synthetic organic chemists have been able to control the
characteristics of polyelectrolyte assemblies in many different
manners: most simply by altering monomeric units, electro-
lyte group, and polymer length. As a result, polyelectrolytes
make a very tunable building block. They have been used in
a number of applications ranging from drug delivery agents,
to components in conducting films, to facilitating colloidal
suspensions, and to creating mimics of biological molecules.
Recently, polyelectrolyte microcapsules have found use as
microreactors for catalyst activity, precipitation reactions,
crystallization reactions, and polymerization reactions.
5.6.1 Electrostatic Self-Assembly

One benefit of using polyelectrolytes is that many of them can
interact with each other in a process known as electrostatic
self-assembly (ESA). As a simple example, Figure 5.18 illus-
trates the complexation of two oppositely charged polyelec-
trolytes in the bulk phase. This basic idea can be applied to
charged surfaces using polyelectrolytes resulting in the for-
mation of multilayer nanofilms. Because ESA relies on attrac-
tive interactions between materials of alternating charge, a
number of different kinds of molecules can be used, such as
Spontaneous
self-assembly
Polycation Polyanion Polyelectrolyte

complexation
Figure 5.18 The complexation reaction between oppositely charged polyelectrolytes.

The ion-pair formation (or ion exchange reactions) occurs due to a large entropic gain when
counterions are expelled into solution. Filled circles represent anionic sites and open circles
cationic sites.
small inorganic molecules, proteins, dendrimers, and poly-

electrolytes. Our exploration of ESA films will focus on poly-
electrolytes. The flexibility of the polyelectrolytes used for
ESA, as well as the ease of the self-assembly process, gives
it advantages over previous methods of building films. Prior
to ESA, the LB technique was the most common for creating
films. However, the LB technique requires far more instru-
mentation, time, and cost to create films that are limited in
long-term stability.
Films formed through ESA techniques can be composed of
many adjacent layers and are generally well ordered on the supra-
molecular scale, depending on the reaction conditions in which
they are made. Both the amount of polyelectrolyte as well as the
order in which they are deposited can be easily controlled. In
the formation of the first layer, a polyion adsorbs to an oppositely
charged substrate surface, because of favorable electrostatic
attraction. In the case of ESA, polyelectrolytes in aqueous solu-
tion adsorb to the substrate surface, since they are removed from
solution and irreversibly bound to the substrate by electrostatic
interactions. By altering the strength of this attraction, it is pos-
sible to alter the amount of polyelectrolyte deposited.
A commonly used system is a positively charged polyelec-
trolyte on an oxidized silica or silanol surface. The silica
substrate surface is negatively charged due to deprotonated
silanol groups. When exposed to a cationic solution, the nega-
tive charge on the substrate quickly attracts the cations to the
surface. ESA methods typically expose a solid substrate sur-
face to the desired adsorbant by “dipping” the substrate into
an aqueous solution of the adsorbant for about 10 minutes. In
order to build multilayered assemblies, this “dipping” proce-
dure is repeated with alternating cation and anion aqueous
solutions, with a rinse step with either water or slightly ionic
solution in between that increases the ability of the next layer
to be well ordered. The rinse step removes weakly adsorbed
molecules that are not a part of the film and primes the surface
for the oppositely charged adsorbant. This approach is gener-
ally known as the layer-by-layer (LbL) technique.
Repetition of this process can be used to make many lay-
ered films. To better understand the overall formation of a
multilayer assembly we must break down the mechanisms of
ESA further, into adsorption of the polyelectrolyte layer to the
substrate film, interpenetration of polyelectrolyte materials
within the film, and the complexation of molecules that occurs
within the film. Figures 5.19 and 5.20 illustrate how ESA can
be used to construct polyelectrolyte multilayer assemblies.
240
Bound counterions
– – – – – – + – –
Negatively charged + – + ++
––––––––––––––––– – + + +
+
+ + + + + + Electrostatic
Glass substrate –––––––––––––––––
attraction
– –
– –
– – –
Process –
repeated
–
-
– – – – – – –
- -- -
- - –
+ + + +
Multilayer
+ +
+
+ + + + + +
– –– –– –– – –– – – –– – – –
Entropy gain
Bilayer
Figure 5.19 Polyelectrolyte self-assembly. The polycation adsorbs from aqueous solution to the negatively charged substrate. This is followed by exposure of this film
to an aqueous solution of the polyanion. Adsorption of the polycation to the polyanionic film results in a bilayer film. The process can be repeated many times to produce a
multilayer nanofilm.
Polyelectrolyte multilayers n
Hn
SO2
Fuzzy nanoassemblies
N
N
8 nm
(80 Å)
CO2–
OH
Polyanion PAZO
0.8 nm
(8 Å)
Polycation N+ n PEI
H H
Figure 5.20 Dimensions of a typical multilayer constructed using the polycation PEI
and the polyanion PAZO. The thickness of a bilayer as measured by ellipsometry is around 1
nm. A 10-bilayer film has a thickness between 100 and 150 nm. Interlayer interpenetration
is illustrated as the “fuzzy” nanoassembly on the right. The degree of interpenetration is
drastically underexaggerated.
5.6.2 Charge Reversal and Interpenetration

Within the multilayer film, there are two main regions: the
surface, at which charge overcompensation or reversal occurs,
and the bulk area, in which interpenetration occurs. The sur-
face of the film is repeatedly exposed to polymer solutions of
alternating charge. The bulk is not exposed to the aqueous
environment. This greatly impacts the differences in chemis-
try that occurs in these two main regions.
At the surface, charge reversal, or charge overcompensation,
helps drive the formation of layers within the film. Typically,
polyelectrolytes are assumed to take a structure with 1:1
charge stoichiometry. However, in order to drive the forma-
tion of multilayers, an overall surface charge must be exposed
prior to the next deposition step in an oppositely charged poly-
electrolyte solution. Charge reversal of the surface enables the
electrostatic interactions that drive the self-assembly process.
The charge overcompensated region of the terminal layer is
matched by oppositely charged salt counterions in solution.
Upon exposure to the next polyionic solution, the counteri-
ons are displaced by the polyion. This displacement is driven
primarily by entropic gain from displacement of the counte-
rions. Entropic gain combined with electrostatic interactions
results in a net negative free energy of adsorption that allows
the polyelectrolyte to remain adsorbed to the surface. Further
adsorption of the polyelectrolyte will stop when repulsive
interactions from material already adsorbed prevail. As a
result, approximately reproducible depositions occur in the

formation of each subsequent layer. Once a polyelectrolyte
has fully adsorbed to the surface, charge overcompensation
occurs again at the terminal layer surface, yielding a reversal
in surface charge, compared to the film prior to deposition. As
happened for the layer before, the charged surface is balanced
by positive counterions from the polyelectrolyte solution. An
assembly with one polycationic and one polyanionic layer
comprises a single bilayer.
The presence of counterions at the charged surface can
be better explained through the use of image forces. Image
forces result from the differences in the dielectric constants
of water and the film interface. The dielectric constant dif-
ference causes a repulsion of water charges from the surface,
allowing a polyelectrolyte counterion to bind at the surface
instead. These counterions will subsequently be displaced by
a polyelectrolyte with the same charge, opposite to the sur-
face—a process driven by gains in entropy because more salt
molecules are displaced from the surface than the number of
polyelectrolyte molecules that adsorb.
To maintain the overall neutral charge of the film, charge
compensation—either intrinsic or extrinsic—must be pres-
ent. Extrinsic compensation refers to polymer charges that
are balanced by extrinsic materials, such as the salt counte-
rions. It has been found that there is actually very little salt
concentration observed within the film. Instead, there is a 1:1
stoichiometric ratio of polycation to polyanion charge. As a
result, the overall charge balance within the multilayers must
be attributed primarily to intrinsic compensation. However,
some extrinsic compensation has been found to occur at the
substrate surface, where charge overcompensation is neutral-
ized by salt counterions.
Within the bulk of the film, we might expect clearly defined
layers of alternating charge. This is not what occurs. Upon
adsorption the polyelectrolyte does not simply “lie” on the
surface. It has been commonly observed that after adsorption,
the outer polyelectrolyte then diffuses slightly into the inner
polyelectrolyte layer. This effect results in blurred or “fuzzy”
distinctions from layer to layer (Figure 5.20). The effects of
interpenetration underscore the lack of organized structure
observed within the bulk film.
The degree to which interpenetration occurs, while not
completely understood, is linked to ionic concentration
and charge density of the polyelectrolyte. In general, ionic
concentration is correlated with interfacial overlap. The salt
ions cause conformational changes of the polyelectrolytes by

shielding the charges on the polyelectrolytes, enabling a poly-
mer to “extend.” As a result, though it becomes more difficult
for the polyelectrolyte to diffuse deeper into the assembly,
the presence of salt ions opens up more polymer segments to
competitive ion exchange with counterions. Polyelectrolyte
charge—whether a polyelectrolyte is strong or weak—also
contributes to the extent of overlap observed. With a weaker
polyelectrolyte, or a polyelectrolyte with lower charge den-
sity, a flat/extended conformation of the polymer cannot be
achieved. As a result, the polyelectrolyte cannot penetrate the
outer film layer to as great a degree. Overall, ionic concentra-
tion plays a significant role (in addition to the charge density
of the polyelectrolyte) in contributing to the interpenetration
of the polyelectrolyte within the film.
Interpenetration results in the bulk of the film appearing to
be more homogeneous, while the surface of the terminal layer
has an overall charge. Interfacial overlap also results in a lack
of crystalline structuring. This limits the applications of ESA
films in producing materials that require ordering for func-
tionality purposes, such as chromophore ordering for nonlin-
ear optical responses. However, the organizational structure
of ESA films has been found to be useful for the construction
of conductive polymeric materials. The overlap from inter-
penetration allows an electron to flow easily through the film,
which would not be the case if the layers were stratified into
organized layers.
5.6.3 Multilayer Formation

Charge reversal, interpenetration, and complexation all con-
tribute to the structure and formation of a multilayer. Broader
factors to consider for successful multilayer formation are film
stability and film thickness. The first, film stability, requires
that a completed multilayer assembly be stable in structure
when removed from solution, so that the film does not change or
degrade following assembly. The second, film thickness, can be
tailored based on ESA conditions and number of multilayers.
Film stability requires that the polyelectrolyte assembly be
irreversible. As a result, spontaneous desorption of material
from the solid substrate film cannot occur in a well-formed
assembly. Desorption due to small ion competitors or desorp-
tion due to exposure to polyelectrolyte solution all cannot occur
in order to maintain the irreversibility of the film. It has been
determined that desorption of material occurs at such a low
kinetic rate that the ESA film can be considered irreversible.
Control of multilayer thickness is determined by a variety

of factors, such as polymer chain length, polymer charge, and
strength of ionic solution. The most obvious way to control
film thickness is to control the number of multilayers formed
on the substrate surface. A greater number of multilayers will
result in a thicker overall film, whereas fewer multilayers
will decrease the film thickness. In addition to the number
of multilayers, layer thickness is highly dependent on ionic
concentration, with an almost linear relationship. Nearly lin-
ear multilayer growth has been observed, up to 100 bilayers
(Reveda and Petkanchin, 1997). Low salt concentrations will
allow the polymer to assume a flat, extended conformation.
Higher salt concentrations result in greater shielding of the
charges on the polyelectrolyte, allowing the charges to be in
closer proximity and resulting in a more coiled structure. Low
ionic concentration screens charges on the polyelectrolyte and
the surface charges, allowing for the flat conformation. The
looped and coiled conformation that results from increased
ionic concentration creates thicker layers and thus an overall
thicker film, though it will not necessarily change the amount
of polymer adsorbed to the surface. Charge density on the
polymer can also affect film thickness. A lower charge density
correlates with thicker layer formation. Similar to the results
at high ionic concentration, a polymer with lower charge den-
sity will assume a more coiled conformation due to fewer
electrostatic interactions between the substrate surface and
the weakly charged polyelectrolyte. Perhaps contrary to intu-
ition, it is important to note that polymer molecular weight
and polymer branching do not cause significant changes in
multilayer thickness.
5.7 MODEL PHOSPHOLIPID BILAYER

FORMATION AND CHARACTERIZATION
Biological membranes are complex structures that contain
a large amount of proteins and are difficult to study in vivo.
By using model phospholipid bilayers, we can study how spe-
cific membrane components interact by incorporating only
these specific components into the membrane (Castellana et al.,
2006). Furthermore, phospholipid bilayers closely resemble cell
membranes in many ways. They retain two-dimensional fluid-
ity and can be great environments for presenting membrane
proteins. Therefore, model membranes have historically been
used to study membrane properties in vitro. They have also
been used for the investigation of biological processes involving
membranes that occur at the cellular level, such as ligand-re-

ceptor interactions, viral attack, and cellular signaling events.
A host of techniques are used to study these membrane sys-
tems. In this book, discussion will be limited to widely used
fluorescence techniques such as fluorescence recovery after
photobleaching (FRAP), fluorescence resonance energy trans-
fer (FRET), and fluorescence interference contrast microscopy
(FLIC), as well as a few surface-sensitive techniques such as
atomic force microscopy (AFM) and quartz crystal microbal-
ance (QCM). There are a number of other important methods,
not covered in this section, for studying model bilayers such
as electron microscopy, which allows for the direct visualiza-
tion of membranes in vivo, impedence spectroscopy, which
allows for electrical measurements of membranes that are
important for understanding neurons, as well as neutron and
x-ray scattering techniques that are used to probe the struc-
ture and periodicity of membranes.
5.7.1 Black Lipid Membranes

Several methods for producing black lipid membranes exist.
All involve the formation of a membrane over a small aperture
usually less than 1 mm in diameter. The hole is formed in a
hydrophobic material such as Teflon and is usually part of a wall
separating two compartments that can be filled with aqueous
solution [see Figure 5.21 (a)]. The two most popular methods for
black lipid formation involve the painting of the lipid solution
over the aperture or the formation of a folded bilayer. Painting
is carried out with a small artist’s paintbrush. The formation of
folded lipid bilayers requires a container with two compartments
separated by a small aperture, and the solution levels in each
compartment must be controlled independently [Figure 5.21
(b)]. The desired solution is filled into each compartment and a
monolayer of phospholipid material is spread on top of one of the
compartments. The solution level containing the monolayer can
be raised and lowered over the aperture to deposit the bilayer.
Black lipid membranes have been used to study various
biophysical processes. Black lipid membranes are suspended
in solution, where there are no unwanted interferences of the
membrane with an underlying support. The absence of a support
allows for transmembrane proteins to be incorporated within
the phospholipid bilayer such that they remain fully mobile
and active. The ability to insert single protein pores into black
lipid membranes allows for the creation of potential nanode-
vices. This has been accomplished by Gu et al. (1999) through
the use of genetically modified α-hemolysin. α-hemolysin
100μ-1 nm pinhole
Hydrophobic
support
(a)
(b)
Figure 5.21 (a) Illustration of a black lipid membrane. (b) The formation of a folded
lipid bilayer. (Reprinted from Castellana, E. T. and Cremer, P. S. Solid Supported Lipid Bilayers:
From Biophysical Studies to Sensor Design. Surface Science Reports 2006, 61.10: 429–444.
Copyright 2006, with permission from Elsevier.)
mutants that can non-covalently capture cyclodextrin mol-

ecules within their pores were created. A current change is
measured due to the restriction of the pore by cyclodextrin.
The binding and unbinding of small organic molecules at the
single molecule level can be measured using this process.
However, the lack of support limits the lifetime of the bilayer
due to poor stability. Furthermore, the methods of detection
and characterization of black lipid membranes are also lim-
ited. Because there is no support, surface-sensitive techniques
cannot be used for characterization.
5.7.2 Solid Supported Lipid Bilayers (Richter et al., 2006)

Supported bilayers, as illustrated in Figure 5.22, are more
robust and stable than black lipid membranes and can also
Solid support
Figure 5.22 Schematic diagram of a supported phospholipid bilayer. (Reprinted from

Castellana, E. T. and Cremer, P. S. Solid Supported Lipid Bilayers: From Biophysical Studies
to Sensor Design. Surface Science Reports 2006, 61.10: 429–444. Copyright 2006, with per-
mission from Elsevier.)
be analyzed by surface-specific analytical techniques such as

AFM, QCM, DPI, and SPR. In these systems fluidity is main-
tained by a 10–20 Å layer of trapped water between the sub-
strate and the bilayer (Groves and Boxer, 2002).
In order to support a high quality membrane the sur-
face should be hydrophilic, smooth, and clean (Tamm and
McConnell, 1985). A few commonly used substrates for bilayer
formation are silica, mica, gold, and titanium oxide. There are
three general methods for the formation of supported phospho-
lipid bilayers on planar supports. The first involves the deposi-
tion of a lower leaflet of lipids from the air-water interface by the
Langmuir-Blodgett technique [Figure 5.23(a)]. This is followed
by the transfer of an upper leaflet by the Langmuir-Schaefer
procedure, which involves horizontally dipping the substrate
to create the second layer. A second method of supported
bilayer formation is the adsorption and fusion of vesicles from
an aqueous suspension to the substrate surface [Figure 5.23(b)].
The third method, a combination of the first two methods, can
be employed by first depositing a monolayer via the Langmuir-
Blodgett technique followed by vesicle fusion to form the
upper layer [Figure 5.23(c)]. Each of the three methods contains
particular advantages and disadvantages. The first and third
methods are useful for the formation of asymmetric bilayers.
Pull
Push
(a) (b) (c)
Figure 5.23 Common methods for forming supported bilayers. (a) Langmuir-Blodgett
technique to deposit monolayer followed by pushing the substrate horizontally through
another lipid monolayer. (b) Vesicles in solution adsorb and fuse to the surface to form a
bilayer. (c) A combination of (a) and (b). (Reprinted from Castellana, E. T. and Cremer, P. S.
Solid Supported Lipid Bilayers: From Biophysical Studies to Sensor Design. Surface Science
Reports 2006, 61.10: 429–444. Copyright 2006, with permission from Elsevier.)
The adsorption and fusion of small unilamellar vesicles

(SUVs) is one of the easiest and most versatile means for form-
ing supported bilayers. Small unilamellar vesicles can be
prepared many different ways. Lipids are often stored in chlo-
roform to prevent degredation. After a known quantity of lipids
is withdrawn from the chloroform suspension, they should be
placed into a small glass container with a large surface area.
The chloroform can then be evaporated with gaseous nitrogen.
This makes a thin coating of lipids in the container, which
should be dried in a vacuum manifold to ensure that there is
no residual chloroform. The dried lipids can be rehydrated to
multilamellar vesicles by exposing the film to whatever sol-
vent is desired to be trapped within the vesicle, then sonicat-
ing them. Extrusion of these multilamellar vesicles through
porous polycarbonate membranes at high pressure will result
in SUVs with a size distribution dependent on the size of the
polycarbonate membrane’s pores.
Parts of the mechanism of vesicle adsorption and fusion
to form a bilayer have been elucidated by QCM studies
(Figure 5.24). The process begins with the adsorption of SUVs
from the bulk solution onto the substrate. In the early stages
after adsorption, SUVs may fuse with one another to form
larger unilamellar vesicles. The vesicles then rupture, form-
ing a supported bilayer in a process that depends upon the
destabilizing interactions affecting the vesicle, such as the
10
0
–10
∆f3 (Hz)
–20
–30
–40 Rinse
Rinse
–50
–60 Rinse
∆D3 (10–6)
2.5
2
1.5
1
0.5
0
0 5 10 15 0 50 100 150 0 5 10 15 0 5 10 15 20
Time (min)
A–DOPC/DOPS (1:2) B–DOPC/DOPS (1:1) C–DOPC/DOPS (4:1) D–DOTAP
Figure 5.24 Lipid deposition pathways measured by QCM-D on silica. (a) Vesicles do
not adsorb. (b) Vesicles adsorb and remain intact, forming a supported vesicular layer (SVL).
(c) Vesicles adsorb and remain initially intact. At high vesicular coverage an SLB is formed.
(d) Vesicles adsorb and rupture instantaneously to form an SLB. (Reprinted with permission
from Richter, R. P., Berat, R., and Brisson, A. R. Formation of Solid-Supported Lipid Bilayers:
An Integrated View. Langmuir 2006, 22.8 : 3497–3505. © American Chemical Society.)
osmotic pressure due to a salt concentration within the vesicle

that differs from the salt concentration outside the vesicle, or
the attraction between the individual lipids and substrate that
causes vesicles to deform from their more stable spherical
shape into a more oblong shape. Figure 5.24 shows four ways
that lipid vesicles interact with the solid support as followed
by quartz crystal microbalance with dissipation monitoring
(QCM-D). The QCM-D technique is a valuable tool for screening
the overall properties of the deposited lipids. The dissipation
parameter allows for distinguishing between intact, adsorbed
vesicles (high dissipation) and bilayer patches (low dissipa-
tion). As shown in Figure 5.24, vesicles either do not adsorb
[Figure 5.24(a)]; adsorb and remain intact, giving rise to a sup-
ported vesicular layer [Figure 5.24(b)]; or form a supported
bilayer [Figure 5.24(c) and (d)]. As observed by QCM, supported
bilayer formation can occur via two scenarios with distinct
kinetics. In one case the vesicles rupture quickly upon interac-
tion with the solid support [Figure 5.24(d)]. In another case, a
large amount of intact vesicles are adsorbed at an intermediate
stage of the process [Figure 5.24(c)].
The main disadvantage of supported bilayers is that the
supported bilayer is not truly decoupled from the underly-
ing substrate. The layer of hydration at the bilayer-support
interface allows for membrane fluidity, but this layer is also
too thin to prevent transmembrane proteins from interacting
unfavorably with the underlying substrate. Such interactions

can denature the protein or cause the membrane to lose its
fluidity.
5.7.3 Polymer Cushioned Phospholipid Bilayers

Figure 5.25 also shows the same system in the presence of a
lipopolymer support that prevents denaturation. The addition
of a polymer layer effectively decouples the membrane from
the surface and still allows for investigation by an array of
surface science techniques. A well-designed polymer cushion
should behave much like a cytoskeleton, present in eukary-
otic cells. In physisorbed systems, weak interactions between
the phospholipid bilayer and the polymer support can result
in an unstable system. This can be overcome by covalently
attaching the polymer to the substrate. Next, anchor lipids or
alkyl side chains capable of inserting into the phospholipid
bilayer are employed in order to further increase stability.
These effectively tether the membrane to the underlying sup-
port and allow for the incorporation of membrane proteins
without denaturing them. Unfortunately, this ability to study
membrane proteins comes at the cost of membrane fluidity.
In general, it is desirable for the polymer cushion to be soft,
hydrophilic, and not too strongly charged.
Two classes of polymers, polyelectrolytes and lipopoly-
mers, are emerging as popular choices for cushion mate-
rial. Polyelectrolyte cushions can be directly adsorbed from
solution to a variety of substrates by means of layer-by-layer
deposition. Lipopolymers consist of a soft hydrophilic poly-
mer layer presenting lipid-like molecules at their surface that
can insert into a phospholipid membrane and tether it to the
polymer spacing.
5.7.4 Fluorescence Recovery after Photobleaching

FRAP is an optical technique capable of quantifying the two-
dimensional lateral diffusion of fluorescently labeled lipids
in a bilayer. This technique is one of the most common tech-
niques performed in order to characterize lipid bilayers. The
technique begins by saving a background image of the fluo-
rescing bilayer. Next, a light source is focused onto a small
patch of the viewable area either by switching to a higher
magnification objective or shrinking the field of view via a
pinhole. The fluorophores in this region are quickly photo-
bleached, which is the photochemical destruction of a fluoro-
phore, and an image reveals a noticeable dark spot surrounded
by undamaged fluorophores in the bilayer [Figure 5.26(b)]. As
Polymer
Denatured
support
protein
Types and Uses of Some Nanomaterials
Solid support Solid support
Figure 5.25 Peripheral domains of transmembrane proteins can become immobilized and denatured on a solid support. A polymer cushion can help shield the protein
from the substrate. (Reprinted from Castellana, E. T. and Cremer, P. S. Solid Supported Lipid Bilayers: From Biophysical Studies to Sensor Design. Surface Science Reports
2006, 61.10: 429–444. Copyright 2006, with permission from Elsevier.)
251
Light
(a) (b)
(d)
(c)
Figure 5.26 Principle of FRAP. (a) The bilayer is uniformly labeled with a fluorescent
tag. (b) The label is selectively photobleached. (c) The intensity within the photobleached
region is monitored as a function of time. (d) Eventually uniform intensity is restored.
Brownian motion proceeds, the still-fluorescent lipids will

diffuse throughout the sample [Figure 5.26(c)] and replace the
destroyed lipids in the bleached region. This diffusion pro-
ceeds in an ordered fashion, which can be modeled by the
diffusion equation (Smith et al., 2008), which shows how the
florescence intensity f(t) recovers (increases) over time, t:
⎡ ⎛ 2τ ⎞ ⎛ 2τ ⎞ ⎤
()
f t = e −2τD /t ⎢ I 0 ⎜ D ⎟ + I 1 ⎜ D ⎟ ⎥ (5.15)
⎣ ⎝ t ⎠ ⎝ t ⎠⎦
where I0 and I1 are known as modified Bessel functions, and

τD = r 2/4D where r is the radius of the bleached area at t = 0 and
D is the fission coefficient. The diffusion constant D can be
simply calculated from a fit of this equation. Diffusion con-
stant values for lipids tend to be in the range of 0.5–5 μm2/s.
Figure 5.27 shows what a typical FRAP profile looks like (a
plot of fluorescence intensity versus time). Fitting this curve
to Equation 5.15 allows the determination of D.
5.7.5 Fluorescence Resonant Energy Transfer

FRET is a useful tool for quantifying molecular dynamics in
nanoassemblies of biophysical and biochemical importance.
For monitoring the locations of two molecules relative to one
another, one of them is labeled with a donor and the other with
an acceptor fluorophore. When they are not near each other,
only the donor emission is detected upon donor excitation.
However, when the donor and acceptor are in proximity (1–10
nm) due to the interaction of the two molecules, the acceptor
f (t)
Figure 5.27 Illustration of a typical FRAP recover curve with corresponding images.
The solid line represents a fit to the data using a 2D diffusion model equation.
emission is predominantly observed because of the intermo-

lecular transfer of energy from the donor to the acceptor.
FRET has been used by Lei and Macdonald (2003) in order
to monitor the fusion of two lipid vesicles. Video images of
two vesicles not fusing and two vesicles fusing are shown
in Figure 5.28. As the vesicles come into contact, the contact
area becomes a different color (depending on the fluorophores
used—red and green in this case), corresponding to simultane-
ous emission of both red and green from the same area (shown
in greyscale in the figure). Thereafter there are dramatic color
and intensity changes [Figure 5.28(j)], which are interpreted as
hemifusion; the color of the contact area changes, the contact
area becomes slightly fatter, and then red color can be seen
beginning to diffuse from the contact area over the surface of
the positive vesicle, reducing the intensity of the green color
in the process. This donor-acceptor relationship was exam-
ined in more detail in Section 5.1.
5.7.6 Fluorescence Interference Contrast Microscopy

In FLIC, the proximity of fluorescent probes to a reflective plane
leads to the modulation of the fluorescence intensity, analysis
of which can provide nanometer-scale topographic informa-
tion. FLIC can have resolution of a few to hundreds of nanome-
ters. The principle underlying FLIC is that in the presence of a
reflective surface, light can travel to a fluorophore along a direct
path or a reflected path; the difference in optical path length
between these two routes leads to interference (Figure 5.29).
The resulting intensity is a function of this path-length dif-
ference, which is a function of the height of the fluorophore
above the reflective plane. Interference occurs independently
254
A B C D E
F G H I J
10 µm
Figure 5.28 FRAP images showing the fusion of two vesicles. The red and green color is shown in grayscale. (Reprinted from Lei, G. and MacDonald, R. Biophysical
Journal, 2003, 85, 1585–1599, Biophysical Society, Elsevier. Copyright 2003, with permission from Elsevier. See Lei and MacDonald for the color image.)
Excitation Emission
Assembly Fluorophore
SiO2 layer
Silicon substrate
Figure 5.29 Basic principles of a FLIC experiment. The resulting intensity is a function
of the path length difference between the direct and reflected light paths, which is a func-
tion of the height of the fluorophore above the reflective plane.
for the excitation and the emission light, each of which must be
considered separately. The typical setup makes use of a silicon
wafer as the reflector. A transparent oxide (SiO2) layer acts as a
spacer. The system of interest (i.e., a lipid bilayer with fluores-
cent probes) is placed on top of the oxide layer.
FLIC can also be used to look at the topography of a second-
ary membrane. A secondary lipid bilayer can be assembled atop
a solid supported bilayer via the rupture of giant lipid vesicles
with diameters that approach tens of microns. The secondary
membrane is separated from the primary membrane by a con-
fined layer of water and is therefore free to exhibit nanometer-
scale height fluctuations.
5.8 SELF-ASSEMBLED MONOLAYERS

A self-assembled monolayer is formed when a particular group
on a molecule has a strong affinity for a specific surface. Thiols
and silanes are examples of such molecules, and these are
discussed in the next two sections. Spontaneous chemisorp-
tions of these adsorbate molecules may occur from either the
vapor or liquid phase and often leads to the formation of an
organized monolayer. The adsorbate, being a self-assembling
molecular building block, often has a relatively simple struc-
ture with the reacting group attached to an alkyl chain and
terminated by some other inert group. The chemisorption step
is usually rapid, lowering the surface energy of the substrate.
This is followed by slow two-dimensional organization of the
alkyl chains. This organization step is dominated by inter-
chain hydrophobic interactions.
Surface organization itself typically occurs in stages, where
each stage represents a distinct phase. First, a low-density
phase is formed comprised of adsorbate molecules randomly

dispersed on the surface of the substrate. This is followed by
a transition to a second intermediate phase with all of these
randomly dispersed molecules lying flat on the substrate. The
final transition to a high-density phase involves the molecules
orienting themselves normal to the substrate surface. Exactly
how these phase transitions proceed depends on temperature.
If lateral interactions are ignored, chemisorption follows the
Langmuir adsorption isotherm. The kinetics of adsorption is
approximately described by Equation 5.16:
dθ/dt = k (1 – θ) (5.16)
where θ is proportional to the amount of area occupied and k

is the rate constant.
5.8.1 Thiols on Gold

Thiols are compounds having the structure RSH (where R ≠ H).
These molecules are also referred to as mercaptans. Some com-
mercially available thiols are shown in Figure 5.30. The sulfur
atom of the thiol is extremely reactive to gold and other noble
metals resulting in the formation of a strong metal-sulfur cova-
lent bond having a bond strength on the order of 100 kJ mol–1.
O SH
HSCH2(CH2)gCH2 OH Fe
O
SH
HS O CH3
HS
X X X X X
S S S S S S
Au Au
Figure 5.30 Upper: Some examples of thiols used to functionalize a gold surface.
Lower: Thiol bound to a gold substrate via a strong Au-S bond. The grey tube is typically a
hydrocarbon chain of specified length. X represents some surface functional group.
The resulting monolayer on the metal surface is thermally sta-

ble and is resistant to various solvents and electrolytes.
Gold is usually the preferred metal for thiol deposition,
mainly because it is inert, biocompatible, and can withstand
harsh chemical cleaning treatments. Gold is also easy to pat-
tern using lithographic methods. Low molecular weight thiols
can be deposited on gold surfaces by vapor deposition. Most
other thiols can be deposited using a simpler method. First the
thiols are dissolved in an organic solvent such as chloroform
or ethanol. The gold is then immersed in this dilute solution
(~1 mM) for 12 to 72 hours at room temperature. The metal is
then removed, washed with appropriate solvents, and dried
in a stream of nitrogen. It should be stressed that the gold sur-
face needs to be free of impurities before being immersed in
the thiol solution. The surface is usually cleaned in appropri-
ate solvents and then exposed to intense UV light in the pres-
ence of oxygen for about 30 minutes. The production of oxygen
atoms and ozone during the UV exposure causes organic
impurities on the gold surface to be oxidized.
The final structure of SAMs constructed using thiols
depends on the curvature of the substrate. Planar substrates
have no curvature. However, SAMs on nanoparticles such
as colloidal particles and nanocrystals tend to stabilize the
reactive surface of the particle and provide specific organic
functional groups at the particle-solvent interface. This sur-
face functionalization is particularly useful for applications,
such as immunoassays, that are dependent on the chemical
composition of the surface.
5.8.2 Silanes on Glass

Silanes are generally used to construct SAMs on nonmetal-
lic oxide surfaces, such as those found on silicon and glass
surfaces (SiO2). The most commonly used silane to produce
a SAM is octadecyltrichlorosilane (also known as OTS). The
structure of OTS is shown in Figure 5.31. It is an organometal-
lic compound in which a long octadecyl chain is connected
to the reactive trichorosilane group (R-SiCl3). OTS reacts vio-
lently with water and is sensitive to air. Like thiols, the sub-
strates (glass) are immersed in a dilute solution of OTS, where
the solvent is organic.
OTS reacts with the Si-OH groups found on the surface of
clean glass or silicon. The reaction is illustrated in Figure 5.31.
In order for the reaction to proceed efficiently, the glass sub-
strates must be cleaned thoroughly. This is usually achieved
by first immersing the glass slides into a hot mixture of
Si
Cl
Cl Cl
Si
O O O OH
OH OH OH OH SiO2
SiO2
Figure 5.31 Octadecyltrichlorosilane (or OTS) chemisorbed on a clean glass substrate

resulting in a close-packed SAM.
concentrated sulfuric acid and 30% hydrogen peroxide for

about 30 minutes. This mixture, known as piranha etch solu-
tion, is a powerful and extremely hazardous oxidizing agent
that rapidly removes organic impurities from the glass sur-
face. Silicon is usually cleaned by exposing the surface to
high intensity UV light for about 30 minutes. As mentioned
previously, the production of oxygen atoms and ozone causes
organic impurities on the silicon to be oxidized.
OTS and other silanes such as dodecyltrichlorosilane are
used in the semiconductor industry to form nanofilm SAMs on
silicon dioxide substrates. More specifically, these molecules
act as thin insulating gates in metal-insulator semiconductors.
OTS is also used in conjunction with conducting polymers in
organic-substrate LCD displays.
An interesting extension of the silane chemistry discussed
above is the sol-gel processing method used to prepare highly
porous silica networks. As an example, let’s consider ethox-
ysilane (Figure 5.32), which can easily be hydrolyzed to the
(R-O)3Si-OH form. Since silanes are highly reactive to OH
groups, (R-O)3Si-OH molecules can self-react in a condensation
process leading to the silica network containing nanopores.
5.9 PATTERNING
As the field of nanotechnology expands, the demand for sur-
face manufacturing techniques that are cheaper, more flexible,
O O
–OH or H+/H O
O 2
Si OH
O Si EtOH O
O O
Hydrolysis
O O
O O
O Si OH
OH OH Si
Si + Si O
O HO –H2O
O O
O O
Condensation
HO
HO HO HO O
O O
O O Si Si
O Si Si Si *
* Si
O
O O O
O O O O
Si
Si Si O O
O O O OH
OH O O *
O O Si Si
O Si Si Si
Si
* O
O O O
O O HO HO
Si
Si *
* *
*
Silica Network
Figure 5.32 The sol-gel process. Silanes are highly reactive to OH groups and
(R-O)3Si-OH molecules, and can self-react in a condensation process leading to the silica
network containing nanopores.
and allow greater feature resolution will continue to increase.

The word “nanolithography” can refer to a wide variety of
nanoscale surface manufacturing techniques for creating pat-
terned surfaces, such as those found in semiconductor cir-
cuits. There are many different ways that such patterns can be
fabricated. Here we present a survey of these methods, their
advantages, their disadvantages, and possible applications.
5.9.1 Optical Lithography

One important method is optical lithography (or photolithog-
raphy), which is a method capable of producing patterns less
than 100 nm using very short wavelength (~190 nm) lasers.
The wavelength of light determines the resolution of the
nanopattern. For example, x-ray lithography can reach a reso-

lution of ~15 nm by using light of wavelength ~1 nm.
Contact lithography is a common form of photolithography.
A photomask is placed over the region to be patterned. The
region is typically a flat substrate coated with an imaging pho-
toresist layer, as shown in Figure 5.33. A photoresist is essen-
tially a light-sensitive material that will typically degrade or
“burn away” after being exposed to light. The general con-
cept consists of aiming electromagnetic radiation at a surface
through openings in a mask template that protects the surface
so that only areas not covered by the mask are exposed. Then
a chemical that functions as a developing solution is used to
remove either the exposed or unexposed areas, leaving either a
“negative” or “positive” imprint of the mask. In this way pho-
tolithography selectively removes parts of a thin film or the
bulk of a substrate. The procedure is usually followed by a
chemical treatment to either remove the degraded material or
to engrave the exposure pattern into the material beneath the
photoresist. The technique excels at creating regular, patterned
surfaces, and the process produces surfaces fairly quickly, but
the fabrication of the mask can be difficult.
As mentioned above, the resolution of the pattern is deter-
mined by the wavelength of light. Usually deep UV light is
used to create the pattern down to about 50 nm. Equation
5.17 shows how the minimum feature size (α) is related to the
wavelength (λ) of light used.
λ
α=k
φ (5.17)
The constant k in Equation 5.17 is usually around 0.4, and ϕ

is the numerical aperture of the lens seen from the substrate.
This equation tells us that the resolution can be improved by
using a smaller wavelength and a larger numerical aperture.
Maximum feature resolution, d, as a function of wavelength is
λ/2. The lower limit for d is around 100 nm. However, resolu-
tions under 100 nm have been achieved by using the principles
of interference combined with photolithographic methods.
Various types of radiation can be used depending on one’s
purpose. Electron beams in particular can create surface
structures with good resolution, down to 116 nm in height
or depth and 105 nm in width. Additionally, electron beams
have been used to create these structures on a silicon surface
without intermediate developing stages; the beams them-
selves etch the surface. Electron-beam lithography suffers the
UV irradiation
Mask
Substrate
Photoresist
Figure 5.33 Patterning a substrate using UV light. The light passes through a photomask and degrades exposed regions of the photoresist.
261
drawback of not being very up-scalable, and it is also time and

money intensive.
5.9.2 Soft Lithography

Another common patterning technique is soft lithography. This
method refers to a family of lithographic techniques for pat-
terning using molds and elastomeric stamps. The stamps are
usually pieces of polydimethylsiloxane (PDMS) that have been
patterned usually against a master to form a relief pattern.
As an example, let’s see how a pattern of octadecane
thiol (ODT) molecules can be transferred to a gold substrate
(Figure 5.34). First the stamp is placed in a solution containing
ODT. ODT molecules adsorb onto the PDMS surface forming
the “ink” component of the stamp. The solvent is removed and
the PDMS stamp is put in contact with a gold substrate, where
the ODT molecules spontaneously chemisorb onto the gold
surface. Thus the pattern from the stamp is transferred for the
gold via the ODT ink. Further details on the chemisorption of
thiols onto gold can be found earlier in this chapter.
5.9.3 Nanosphere Lithography

Nanosphere or colloidal lithography uses a packed array
of nanospheres as the mask for the surface to be patterned.
Chemicals can be deposited in the exposed areas between the
spheres, or the surface in these exposed areas can be etched
by using ion beams. The pattern produced can be manipulated
in several ways. For example, chemical deposition can be done
ODT in ethanol
PDMS
(a)
Gold Gold Gold

(b)
Figure 5.34 (a) A solution of ODT is placed over the PDMS stamp. The solvent is
removed and ODT molecules assemble on the stamp. (b) The PDMS stamp with the ODT “ink”
is placed on a gold substrate. When the stamp is removed, the ODT molecules in contact with
the gold surface chemisorb to the substrate, transferring the pattern from the stamp to the
gold via the ODT ink.
at different angles, or heating the packed array of nanospheres

at different temperatures can cause it to deform in desired
ways. It is also possible to create multiple layers of sphere
arrays, where the upper layers can act as a mask for the lower
layers, leaving a complex three-dimensional structure. It is
worth noting that the resolution is dependent on the size of
the nanospheres used in the mask. By using sufficiently small
spheres, one can achieve resolutions below 100 nm.
This technique has the benefit of being simple and inex-
pensive compared to other nanolithographic techniques.
Nanosphere lithography can also produce structures that are
difficult or impossible to create with other nanolithographic
techniques. Unfortunately, it is difficult to produce a flawless
surface with the method; errors occur as the result of difficult-
to-avoid disorder in the nanosphere mask, which may lead to
small differences between masks that are formed in similar
conditions. Furthermore, the space between features and their
size cannot be changed independently, as both are directly
determined by the size of the nanospheres used. The number
of structures that can be created by this method is also limited,
because structures created by nanosphere lithography must be
based on the unoccupied interstices of a packed sphere array.
For example, it is hard to imagine a field of square nanosphere
pillars being produced by a mask of patterned spheres.
5.9.4 Patterning Using AFM

In addition to being used as an imaging tool, the atomic force
microscope (AFM) can be used to manipulate surfaces on a
very small scale. In standard use, the AFM images a surface in
three dimensions by measuring the deflection of a small, very
sharp tip as it moves along a surface. For patterning, however,
the tip can be used to interact with a surface and alter it in
order to produce very small structures.
Some nanolithographic techniques that employ AFM may
be termed “destructive”—they etch or otherwise damage a
surface to create the intended surface structures. The tip can
be used to break apart surfaces on a very small scale, or to
move very small objects around on a surface. The AFM tip can
also be modified to irreversibly alter the surface in various
ways. For example, a catalyst can be added to the tip to initiate
reactions. Another common method for surface alteration is
the initiation of very localized redox reactions by running a
current through a conductive AFM tip. UV light can be used to
control this localized redox process because it affects electron-
hole pairs in a way that changes their resistance to current
flow. This means that a low-voltage current running through

the tip will be able to oxidize a surface only when UV light
is shined on the surface at the same time. This method can
produce smooth surface features with a height of about 40 nm
and a lateral length of about 1 μm. Surface-destructive AFM is
thus fairly flexible, but its use can be limited by expense and
low production speed.
Other nanolithographic techniques that use AFM may be
termed “constructive” because they build up the intended
structure on the surface. As with destructive AFM nanolithog-
raphy, there are several different possible implementations.
A common constructive method is the use of AFM as a “dip
pen.” The AFM tip is covered in some molecule to be used as
“ink,” which is then deposited on the surface at the meniscus
between the tip and the surface. A resolution below 50 nm is
currently possible with dip-pen methods.
Metals can be transferred to the tip by applying pressure to
the tip on a metal surface, heating it, and allowing it to cool.
The substance can be transferred from the tip to a surface by
reheating the tip as it is dragged over the surface. Using this
technique with indium has been successful in the creation of
conductive nanoscale wires, with widths as low as 50 nm.
There are many different molecules that can be used as ink
for this method, giving a great deal of flexibility in its uses for
surface patterning. For example, a surface can be patterned
with chemicals that will protect it against a later etching treat-
ment, allowing for an interesting combination of constructive
and destructive lithographic techniques to make more com-
plex patterns. Another possible use of this method is the pat-
terned deposition of particles that can act as “seeds” for new
pattern creation, either by catalyzing reactions or by binding
to other molecules with interesting shapes that form the basis
for a new pattern.
An advantage of dip-pen nanolithography is that the tech-
nique can also be scaled up by using an array of several AFM
tips instead of just one, allowing something analogous to
desktop printer functionality for nanofabrication. Using an
array of tips would allow for the creation of patterns at greater
speeds and over larger areas. This has shown great potential
for the use of AFM in the creation of nanoscale electronic
components such as transistors.
Dip-pen techniques also have drawbacks, the most unfor-
tunate of which is that the created patterns are not perfectly
reproducible. A number of different factors affect molecule
deposition: temperature, humidity, the type of molecule used
as ink, the characteristics of the AFM tip, the write speed, and
the pressure used can all affect the pattern ultimately produced
on the surface. For example, altering the humidity can change
the size and structure of the meniscus where ink is deposited,
resulting in a larger or smaller area of the surface onto which
the same amount of ink should have been deposited.
“Fountain-pen” methods, in which the tip is placed within
a nanoscale tube that delivers a constant flow of “ink” mol-
ecules to the tip, are a related surface fabrication method with
similar advantages and drawbacks. The resolution here is
worse than dip-pen methods, as the fountain-pen methods can
produce only patterns with feature widths above 200 nm.
AFM tip hammering nanolithography is yet another varia-
tion on AFM nanolithography. In AFM tip hammering nano-
lithography, the AFM tip is used to imprint the surface. One
paper by Wang et al. (2009) gives an example of the technique
being used on a polystyrene-block-poly(ethylene/butylenes)-
block-polystyrene (SEBS) copolymer. The AFM tip deforms
the polystyrene sphere component of the copolymer, result-
ing in an imprinted area. Embossing as well as imprinting is
possible; one simply imprints the area around the area to be
embossed to get a raised structure.
This technique has several advantages. For example, the
surface imprint is reversible. Depending on the copolymer
used as the surface, heating it at high temperatures that allow
for reorganization can erase the imprint. This does not mean
that the imprint is unstable at all temperatures. At room tem-
perature, below the threshold energy required for reorganiza-
tion of many molecules, the patterns are fairly stable. For the
SEBS copolymer, it was found that the imprint is erased in
about 5 minutes at 70°C and that around 50% of the original
surface contrast remains after 70 days at 25°C. More stable
copolymers would require a higher temperature to reverse
the patterning and clear the imprint, but they would likely
also be more stable for longer at room temperature. Also,
the AFM tip hammering nanolithography has a very good
feature resolution and is currently able to produce patterns
with features as small as 13 nm for imprinting and 18 nm
for embossing. The combination of feature resolution, surface
stability, and reversibility makes the method attractive for
high-density data storage. The method also has drawbacks,
the primary one being surface specificity. The surface chosen
must be imprintable with a reasonable level of force, but also
stable enough that the imprinting will survive a desirable
amount of time.
5.9.5 Summary
Nanolithographic methods can be used individually or com-
bined with other fabrication methods to amplify the range of
possible surface structures and their uses. For example, pho-
tolithography can be used to create a pattern on a surface on
which nanowires can be grown. The result is multiple levels
of roughness, a surface property with practical applications in
superhydrophobicity. Nanolithography is thus a quickly devel-
oping field encompassing many techniques for the fabrication
of very useful nanoscale patterned surfaces.
5.10 DNA AND LIPID MICROARRAYS

Over the past two decades, advances made in our understand-
ing of genetics have led to the creation of DNA microarrays.
DNA microarrays are surfaces that have been coated with
specific oligonucleotide sequences and have been particularly
useful in gene expression studies. DNA microarrays vary in
the number of oligonucleotides as is determined by the nature
of the study; diagnostic DNA microarrays generally use tens
of oligonucleotides, whereas those for research and screen-
ing can have hundreds of thousands of oligonucleotides on a
single microarray. This section will address the basic proto-
col of a DNA microarray experiment, the various methods of
array fabrication, optimization of an array, and their applica-
tions. The student is encouraged to review basic biochemistry
of DNA before reading this section.
5.10.1 Using a DNA Microarray

Each DNA microarray is unique to a certain purpose, such as
diagnosing cardiovascular disease or monitoring the entire
genome of a strain of Escherichia coli; as such, a microar-
ray must be fabricated or purchased for each experiment.
Depending on the size of the microarray and the availability of
equipment, it is often more feasible to purchase prefabricated
microarrays. Fabrication techniques will be described in more
detail later in this section.
Each spot on a DNA microarray contains thousands of
oligonucleotide molecules called probes that complement
their corresponding mRNA. A sample is prepared for DNA
microarray analysis by extracting the cells’ total RNA. The
RNA can be purified by treating it with poly(A) polymerase
followed by oligo(dT) chromatography. The purified RNA is
then labeled for detection; in the past, 33P has been used,
although radioactive labeling has mostly been replaced by
the use of fluorescent markers Cy3 and Cy5. The labeled

RNA “targets” are then spotted on the DNA microarray
probes to which the RNA binds in a process called “hybrid-
ization.” Before measuring fluorescence, the DNA microar-
ray is washed to remove any excess or unhybridized targets.
Fluorescence can be quantified by confocal laser scanning
microscopy or other fluorimeters. Using internal standards,
fluorescence can be used to calculate the number of targets
that hybridize to the probes, which can then be used to deter-
mine the change in gene expression in a normal state com-
pared to an experimental state. Running a DNA microarray
experiment is relatively quick and easy; the biggest factor
detracting from more widespread use is the lengthy fabrica-
tion process.
5.10.2 Array Fabrication

There are three primary methods for DNA microarray fabri-
cation: in situ synthesis, contact printing, and non-contact
printing (Dufva et al., 2005). In situ synthesis creates the oligo-
nucleotides directly on the microarray and is capable of produc-
ing the highest quality microarrays, although these high-quality
arrays require expensive robotics systems and the use of clean
room techniques. All methods of in situ synthesis use nucleo-
sides that have already been modified with a protecting group to
ensure that the correct sequences of oligonucleotides are pres-
ent upon completion. In the first method [Figure 5.35(a)], nucle-
osides are modified with the photolabile reagent 2-nitrophenyl
propoxycarbonyl (NPPOC). These groups are removed using
light and can then react with available nucleosides. Synthesis
is controlled by the use of various masks that are placed on
top of the microarray, allowing only specific regions to react.
After subjecting a mask to light, the appropriate nucleoside is
added, the microarray is washed, and then the process repeats.
The downside to this method is that it requires a different mask
for each step of the synthesis. Another method [Figure 5.35(b)]
involves similar chemistry, yet instead of masks, a series of mir-
rors is used to direct light to the spots to be deprotected. Yields
of 77% for 25 base pair oligonucleotides have been achieved
using NPPOC as the protecting group.
Because they are cheap and still feasible, printing meth-
ods are generally the preferred methods of DNA microarray
fabrication. Contact printing uses a robotic system of high
definition pins to dispense quantities of probe solutions at
given coordinates. Non-contact printing differs in that instead
of pins, inkjets or similar technologies are used to dispense
(a) G
G G
G
A T C A T C A T C
T G T T G T T G T
G A G G A G G A G
C C A C C A C C A
(b)
G
G G
G
A T C A T C A T C
T G T T G T T G T
G A G G A G G A G
C C A C C A C C A
(c)
G
G
A T C A T C A T C
T G T T G T T G T
G A G G A G G A G
C C A C C A C C A
(d)
A B C
O O O O O O
LI LI LI LI LI LI
G G G G G G
O O O O O O
A A B A B C
Step 1 Step 2 Step 3
Figure 5.35 Methods of fabrication. (a) In situ with masks. (b) In situ with mirrors.
(c) In situ printing. (d) Printing pre-synthesized oligonucleotides. (Reprinted from Dufva,
M. Fabrication of High Quality Microarrays. Biomolecular Engineering 2005, 22, 173–184.
Copyright 2006, with permission from Elsevier.)
picoliter quantities of solution. Non-contact printing is pre-

ferred over contact printing as non-contact printing creates
smaller spots, and thus more spots can be fit on a microarray.
Non-contact printing has been used in one method of in situ
synthesis [Figure 5.35(c)]. In this method, all bases are depro-
tected followed by site-specific dispensing of phosphoramidite-
protected nucleosides; cycles of deprotection and the addition
of nucleosides are repeated to complete oligonucleotide synthe-
sis. Perhaps the most practical methods for fabrication of DNA
microarrays are the printing methods in which all oligonucle-
otides are synthesized beforehand. Via contact or non-contact
(a) Quartz Template
Chrome Layer
Aqueous Phase (PBS)
Supported Lipid Bilayer
Hydration Layer
Silica Substrate
100X
UV (184–275 nm) Zoom-in
(b) (d)
PBS 2 um
100 um
180 (e)
160
140
120
100
80
(c) 60
PBS
Pixel Intensity (a.u.)

40
20
PBS 0
0 60 120 180 240 320 360 420
100 um Sample Position (pixel)
Figure 5.36 Direct patterning of void arrays within bilayer membranes using UV photolithography. (a) A schematic diagram of the key process steps. (b) Bright-field image,
where bright squares are quartz and dark regions reveal the chrome background. (c) Epifluorescence images revealing resultant fluorescence patterns. (d) A high-magnification
fluorescence image of the sharp boundary between the UV-exposed and UV-protected regions. (e) Fluorescence intensity profile across an arbitrarily chosen line spanning four
alternating UV-exposed and unilluminated bilayer regions. (Reprinted with permission from Yee, C. K., Amweg, M. L., and Parikh, A. N. Direct Photochemical Patterning and
269
Refunctionalization of Supported Phospholipid Bilayers. Journal of the American Chemical Society 2004, 126.43:13962–13972. © American Chemical Society.)
printing, known quantities of each individual oligonucleotide

are dispensed at their desired position [Figure 5.35(d)].
5.10.3 Optimization
A great deal of optimization is required before a DNA microar-
ray can be used effectively. The quality of a DNA microarray
is greatly influenced by the method in which it was fabricated.
For example, spot density, the number of spots that can be
placed in a given area of a microarray, varies greatly among in
situ, non-contact printing and contact printing methods. Using
in situ methods, spot sizes with diameters less than 10 μm
have been created, whereas printing methods can create spot
sizes between 20 and 30 μm. Also affecting spot density is the
array geometry. It is important for spots to be arranged in a
geometry that maximizes spot density while preventing any
overlap; additionally, precise array geometry is significant for
future data analysis. Currently, our ability to create high spot
density arrays has surpassed our ability to accurately interpret
fluorescence. Spot sizes of approximately 30 μm are the small-
est that result in distinct fluorescence from separate spots. The
creation of higher resolution fluorimeters will be required for
further miniaturization of DNA microarrays.
Spot morphology is another parameter that must be moni-
tored. A homogeneous spot is preferred for data analysis. Poor
spot morphology is less of an issue with light-directed in
situ syntheses. Spot morphology of printing methods can be
improved by controlling humidity and temperature, as well as
by adjusting the spotting buffer.
Perhaps the most important factors to be controlled for
optimization are probe density and hybridized density. These
two factors are closely related, as probe density is the num-
ber of probes in a given spot, and hybridized density is the
fraction of these probes that hybridize when saturated with
target molecules. In order to get the greatest fluorescent signal,
hybridized density must be maximized; however, a number
of issues arise. If probe density is too high, probe molecules
may interact with other probe molecules on the surface and
become unable to hybridize. Additionally, high probe density
leads to high steric hindrance, which prevents targets from
hybridizing. A number of troubleshooting methods have been
developed to optimize these parameters. In many cases, the
microarray surface and the probe molecules can be modified
to maximize functional probe density.
Another method to optimize hybridized density uses linker
molecules to move the probes farther from the surface to
minimize or prevent any unwanted interactions. Innovation of

the linker has resulted in a dendrimeric linker, which has been
shown to greatly increase probe density and hybridized density.
5.10.4 Applications
DNA microarrays are excellent tools for studying gene expres-
sion and have valuable uses in both research and industry.
One of the first uses of DNA microarrays was in the study
of E. coli. Samples of stationary-phase E. coli and log-phase
E. coli were analyzed by DNA microarray. The data found from
microarray analysis led to the discovery of multiple growth-
regulating genes and an overall improved understanding of
this significant microorganism (Ye et al., 2001).
Industrially, DNA microarrays are of great significance in
optimizing biocatalysis. Particularly, fermentation processes are
often optimized using DNA microarrays. As biological hydrogen
fuel cells gain more attention, it is likely that DNA microarrays
will be of great importance in optimizing these systems.
DNA microarrays have great potential to impact healthcare.
Intelligent drug design has benefited from the use of DNA
microarrays as it allows drug developers to anticipate adverse
effects or, alternatively, to discover new drug targets. Other
kinds of microarrays are also effective in diagnosing diseases.
DNA microarrays have been designed to screen for inherited
genetic diseases; also, changes in the expression of certain
genes can be attributed to cancer and other chronic illnesses,
allowing for earlier detection. Perhaps the most exciting pos-
sibility with DNA microarrays is the development of personal-
ized medicine. Using a DNA microarray to better understand
an individual’s genotype will allow doctors to more effectively
treat each person, ultimately resulting in a healthier country.
5.10.5 Arrays of Supported Bilayers and

Microfluidic Platforms
Spatially addressed microarrays have found use as extremely
rapid and powerful means of data collection. Unlike DNA-,
protein-, or peptide-based arrays, phospholipid bilayer sys-
tems must remain hydrated at all times in order to retain the
desired supramolecular structure (Figure 5.36). This require-
ment creates a significant challenge for creating arrays of sup-
ported bilayers.
The first method for patterning surfaces with solid sup-
ported phospholipid bilayers was created by Groves et al.
(1997). A typical formation procedure involved the patterning
of photoresist on fused quartz wafers by means of standard
photolithographic techniques. SUVs were then fused onto the

substrate between the barriers, creating a lithographically pat-
terned array of identical planar-supported membranes.
Arrays of supported membranes can also be fabricated
by selectively destroying regions of a continuous supported
bilayer. This is achieved by high-intensity UV illumination
through a photomask under aqueous conditions (Yee et al.,
2004). The UV radiation generates both ozone and singlet oxy-
gen in highly localized regions. These species react with and
degrade the lipids to form water-soluble components. These
patterns of holes or of corralled bilayers display long-term sta-
bility, retaining their geometric shapes, sizes, and distribution
as well as their relative position on the substrate surface.
As shown in Figure 5.37, sharp corners of geometrical fea-
tures on the mask always result in smoothly curved or rounded
edges. This observation is attributed to steric crowding and line-
tension effects of the phospholipids. This approach is applica-
ble for producing void patterns of arbitrary shapes, sizes, and
densities at predetermined regions within the bilayer.
A neat feature of this patterning method is that the voids
can be backfilled by subsequent exposures to the same or a dif-
ferent vesicle solution. This makes it possible to manipulate
membrane compositions and dynamically probe lipid-lipid
diffusive processes. POPC bilayer samples patterned using the
photochemical method shown in Figure 5.37, when exposed to
small unilamellar vesicles, causes the nonfluorescent voids to
Vesicle Microfluidic Flow Patterning
solution
injectors
E
Normalized Intensity Normalized Intensity
1.00 (40 V/cm) 50 µm

0.75
Texas Red
0.50
0.25
Flow 0.00
outlet 1 2 3 4 5 67 8
100 µm
1.00
0.75
0.50
NBD
0.25
0.00
1 2 3 4 5 67 8
Corral (from left) Field-induced separation of
oppositely charged dye-lipids:
Texas Red (–) and DiD (+)
Captured Membrane micro-array
Figure 5.37 Addressing by laminar flow in a microfluidic channel. Diffusive mixing

in a microchannel under laminar flow conditions provides a concentration gradient of dif-
ferent dye-labeled vesicles. (Reprinted with permission from Groves, J. T. and Boxer, S. G.
Micropattern Formation in Supported Lipid Membranes. Acc. Chem. Res. 2002, 35: 149–157.
© American Chemical Society.)
be filled with lipids from the secondary vesicles. Depending on

the composition of the secondary vesicles, the pattern can be
erased or retained. If the lipids in the secondary vesicles are the
same as the initial patterned bilayer, the lipids will diffuse and
homogenize quickly, thereby erasing the pattern. If the lipids in
the secondary vesicles have significantly different translational
mobility, they will retain the backfilled pattern longer. This
suggests the possibility of creating a fluid bilayer background
with patterned microdomains at specific locations.
The use of laminar flow inside microfluidic channels is also
an effective means of producing composition arrays of supported
phospholipids bilayers in which two distinct chemical compo-
nents can be varied simultaneously along a one-dimensional
gradient. This allows for the addressing of patterned substrates
by the flow of concentration gradients of SUVs formed by the
diffusion mixing of two different SUV solutions. Figure 5.38
demonstrates the process of forming a one- or two-component
composition array by laminar flow in microfluidic channels.
A drawback to this method is the limited number of distinct
components that can be simultaneously addressed as well as
the lack of control over the positioning of the bilayers.
More recently, Smith et al. (2008) used a 3-D continuous flow
microspotter (CFM) system for the preparation of multiana-
lyte lipid bilayer arrays. This method is capable of producing
higher density multicomponent arrays compared to traditional
2-D microfluidics. The poly(dimethylsiloxane), or PDMS,
microspotter consists of a series of inlet and outlet wells con-
nected by pairs of microfluidic channels embedded within the
CFM Cross- PDMS CFM Print head

section vesicle solution
Microarray
substrate
Flow
channels
Silica substrate
Figure 5.38 Schematic of the CFM apparatus and closeup of the CFM print head in
contact with a silica substrate used for bilayer formation. (Reprinted with permission from
Smith, K. A., Gale, B. K., and Conboy, J. C. Micropatterned Fluid Lipid Bilayer Arrays Created
Using a Continuous Flow Microspotter. Analytical Chemistry 2008, 80.21: 7980–7987.
© American Chemical Society.)
polymer. When the PDMS print head contacts the substrate,

one continuous channel is formed between the inlet and out-
let pairs, resulting in the continuous flow of solution over the
substrate. Each channel is individually addressable, allowing
for the production of 2-D bilayer arrays. A prepatterned sub-
strate is not necessary because the bilayers are effectively cor-
ralled into discrete micrometer-sized domains by the residual
PDMS deposited on the silica substrate from the PDMS print
head, which prevents the lipids from spreading. The packing of
the lipids within a 400 × 400-μm2 area with spacing of 400 μm
between areas resulted in well-behaving bilayers.
5.10.6 Summary
The DNA and lipid microarrays are fairly new technologies
whose potential utility has yet to be reached. Widespread
use is currently limited by the length of time involved in
oligonucleotide synthesis and array fabrication techniques,
which make the cost of a microarray experiment rise signifi-
cantly. As technology advances, microarrays are likely to be
an essential tool in any biochemist’s or molecular biologist’s
arsenal.
CITED REFERENCES
Baddour, C. E.; Breins, C. Carbon Nanotube Synthesis: A Review.
Int. J. Chem. React. Eng. 2005, 3, 1–20.
Blodgett, K. B. Films Built by Depositing Successive Unimolecular
Layers on a Solid Surface. J. Am. Chem. SOC. 1935, 57,
1007–1022.
Castellana, E. T.; Cremer, P. S. Solid Supported Lipid Bilayers:
From Biophysical Studies to Sensor Design. Surf. Sci. Rep.
2006, 61(10), 429–444.
Crossland, E. J.; Ludwigs, S.; Hillmyer, M. A.; Steiner, U.
Freestanding Nanowire Arrays from Soft-Etch Block
Copolymer Templates. Soft Matter. 2007, 3, 94–98.
Dai, H. Carbon Nanotubes: Opportunities and Challenges. Surf.
Sci. 2002, 500, 218–241.
Drexhage, K. H.; Zwick, M. M.; Kuhn, H. Sensitized Fluorescence
after Light-Released Energy Transfer Through Thin Layers.
Ber. Bunsen-Ges. Phys. Chem. 1963, 67, 6267.
Dufva, M. Fabrication of High Quality Microarrays. Biomol.
Eng. 2005, 22, 173–184.
Gao, X.; Yang, L.; Petros, J. A.; Marshall, F. F.; Simons, J. W.; Nie,
S. In Vivo Molecular and Cellular Imaging with Quantum
Dots. Curr. Opin. Biotechnol. 2005, 16, 63–72.
Groves, J. T.; Boxer, S. G. Micropattern Formation in Supported

Lipid Membranes. Acc. Chem. Res. 2002, 35, 149–157.
Groves, J. T.; Ulman, N.; Boxer, S. G. Micropatterning Fluid
Lipid Bilayers on Solid Supports. Science. 1997, 275(5300),
651–653.
Gu, L. Q., et al. Stochastic Sensing of Organic Analytes by a
Pore-Forming Protein Containing a Molecular Adapter.
Nature. 1999, 398(6729), 686–690.
Johal, M. S.; Parikh, A. N.; Lee, Y.; Casson, J. L.; Foster, L.;
Swanson, B. I.; McBranch, D. W.; Li, D. Q.; Robinson,
J. M. Study of the Conformational Structure and Cluster
Formation in a Langmuir-Blodgett Film Using Second
Harmonic Microscopy and FTIR Spectroscopy. Langmuir
1999, 15, 1275.
Keidar, M.; Waas, A. M. On the Conditions of Carbon Nanotube
Growth in the Arc Discharge. Nanotechnology 2004, 15,
1571.
Klitzing, K. V.; Dorda, G.; Pepper, M. New Method for High-
Accuracy Determination of the Fine-Structure Constant
Based on Quantized Hall Resistance. Phys. Rev. Online
Arch. 1980, 45, 494–497.
Knobler, C. M. Science 1990, 249, 870.
Lei, G.; MacDonald, R. Lipid Bilayer Vesicle Fusion: Intermediates
Captured by High-Speed Microfluorescence Spectroscopy.
Biophysical Journal, 2003, 85, 1585–1599, Biophysical
Society, Elsevier.
Möbius, D. Z. Naturforsch. 1969, 24a, 251.
Radeva, T.; Petkanchin, I. Electrical Properties and Conformation
of Polyethylenimine at the Hematite-Aqueous Interface.
J. Colloid Interface Sci. 1997, 196, 87.
Richter, R. P.; Berat, R.; Brisson, A. R. Formation of Solid-
Supported Lipid Bilayers: An Integrated View. Langmuir.
2006, 22(8), 3497–3505.
Saito, R.; Dresselhaus, G.; Dresselhaus, M. S. Physical Properties
of Carbon Nanotubes. Imperial College Press: London,
1998.
Sinnot, S. B.; Andrews, R.; Qian, D.; Rao, A. M.; Mao, Z.;
Dickey, E. C.; Derbyshire, F. Model of Carbon Nanotube
Growth through Chemical Vapor Deposition. Chem. Phys.
Lett. 1999, 315, 25–30.
Smith, K. A.; Gale, B. K.; Conboy, J. C. Micropatterned Fluid
Lipid Bilayer Arrays Created using a Continuous Flow
Microspotter. Anal. Chem. 2008, 80(21), 7980–7987.
Takayanagi, K. Suspended Gold Nanowires: Ballistic Transport
of Electrons. JSAP Int. 2001, 3, 3–8.
Tamm, L. K.; McConnell, H. M. Supported Phospholipid-

Bilayers. Biophys. J. 1985, 47(1), 105–113.
Thostenson, E. T.; Resn, Z.; Chou, T. Advances in the Science and
Technology of Carbon Nanotubes and Their Composites:
A Review. Compos. Sci. Technol. 2001, 61, 1899–1912.
Wang, et al. AFM Tip Hammering Nanolithography, Small
2009, 5(4), 477–483.
Ye, R.; Wang, T.; Bedzyk, L.; Croker, K. Applications of DNA
Microarrays in Microbial Systems. J. Microbiol. Meth.
2001, 47, 257–272.
Yee, C. K.; Amweg, M. L.; Parikh, A. N. Direct Photochemi
cal Patterning and Refunctionalization of Supported
Phospholipid Bilayers. J. Am. Chem. Soc. 2004, 126(43),
13962–13972.

• Hamley, I. W. 2007. Introduction to Soft Matter,
Revised Edition. John Wiley & Sons, Chichester, West
Sussex, UK. Chapter 6 provides an excellent read on
lipid membranes, proteins, and other macromolecu-
lar assemblies. Since material on liquid crystals is not
covered in detail, it is recommended that the student
read Chapter 5 in this book. This book also has some
excellent problems.
• Kuhn, H., and Försterling, H.-D. 2000. Principles
of Physical Chemistry: Understanding Molecules,
Molecular Assemblies, Supramolecular Machines. John
Wiley & Sons, Chichester, West Sussex, UK. Chapters
22 (“Organized Molecules Assemblies”) and 23
(“Supramolecular Machines”) are crucial reads for any-
one interested in the physical chemistry of self-assem-
bly and supramolecular processes. These chapters are
very well written and accessible to undergraduates.
• Hanson, G. W. 2008. Fundamentals of Nanoelectronics.
Prentice-Hall, Upper Saddle River, NJ. This book
gives a nice quantum mechanical treatment of nano-
electronics. Chapter 9, “Nanowires and Nanotubes”
is particularly useful. This book assumes a strong
background in physics and is recommended for stu-
dents interested in free and confined electrons in
nanomaterials.
• Lehn, J. M. 1995. Supramolecular Chemistry. VCH
Weinheim. An essential reference to supramolecular
chemistry written by the Nobel Laureate who coined

the term.
• Prasad, P. N. 2003. Introduction to Biophotonics. John
Wiley & Sons, Hoboken, NJ. This book, while not focus-
ing on nanomaterials, provides some very interesting
examples of biomaterials that are used in nanofilms.
These include materials for biosensors and microarray
technology for genomics and proteomics. There is also
an interesting chapter on bionanophotonics.
• Prasad, P. N. 2004. Nanophotonics. John Wiley &
Sons, Hoboken, NJ. This is a highly recommended
read. There is excellent coverage on nanolithog-
raphy and nanophotonics for biotechnology and
nanomedicine.
• Decher, G., and Schlenoff J. B. (Eds.). 2003. Multilayer
Thin Films: Sequential Assembly of Nanocomposite
Materials. John Wiley & Sons, Weinheim, Germany.
This is the best source of information on electrostatic
self-assembly of polyelectrolytes. This book is a crucial
read for those interested in the layer-by-layer assembly
of polyelectrolytes. The book provides excellent cov-
erage of both fundamental principles and potential
applications of polyelectrolyte multilayers.
• Rao, C. N. R., Müller, A., and Cheetham, A. K. (Eds.).
2005. The Chemistry of Nanomaterials: Synthesis,
Properties, and Applications (Volumes 1 and 2). John
Wiley & Sons, Weinheim, Germany. This book contains
some of the best reviews on quantum dots, nanotubes,
and nanowires (synthesis and properties). The book
also contains a good chapter on oxide nanoparticles.
• Rao, C. N. R., Müller, A., Cheetham, A. K. (Eds.).
2007. Nanomaterials Chemistry: Recent Developments
and New Directions. John Wiley & Sons, Weinheim,
Germany. This book provides good coverage of mostly
inorganic nanomaterials. The book describes the use of
nanomaterials for some interesting applications such as
supercapacitors, molecular machines, and transistors.
• Gompper, G., and Schick, M. (Eds.). 2006. Soft Matter.
John Wiley & Sons, Weinheim, Germany. The first
volume in this book explores polymer melts and mix-
tures. The second volume focuses on complex colloidal
suspensions. Both volumes are heavy on theoretical
studies and are recommended for graduate students
seriously interested in computational chemistry.

1. (a) Using antibodies A and D, which bind specifically
to an antigen X at different sites, and a UV-vis
spectrophotometer, design a series of experi-
ments to quantify the concentration of X in solu-
tion. Explain all steps/calculations.
(b) What kind of repulsive and attractive interactions
would you expect between two zwitterionic lipid
bilayers? Between a zwitterionic lipid bilayer and
solid support? How are they different?
(c) Consider a situation where quantum dots are
injected into a cell for imaging—will the size of
quantum dots interfere with cell functions or pose
any problems in this experiment?
2. Consider the three molecules shown in Question 11,
Chapter 2. Langmuir films that can be constructed
from these molecules show hysteresis in their Π-A iso-
therm. Hysteresis occurs when the compression iso-
therm differs from the expansion isotherm. Hysteresis
in the isotherm for these molecules is largest for the
one with the strongest end group dipole moment.
Explain this observation.
3. Using the following materials and your knowledge of
self-assembled monolayers, design a way to immobi-
lize an antibody (containing primary amines) on a gold
surface. The surface is functionalized through a pro-
cess of incubation in the given solutions, and QCM is
then used for the immobilization of the antibody itself.
Give details as to why and how each material binds
to the underlying layer. Qualitatively predict what
the QCM mass versus time profiles will look like for
each deposition step. Materials: (a) An aqueous solu-
tion of 1-ethyl-3-[3-dimethylaminopropyl]carbodiim-
ide hydrochloride (or EDC), (b) mercaptoundecanoic
acid in anhydrous ethanol, (c) an aqueous solution
of N-hydroxysuccinimide (NHS), (c) a gold-coated
substrate.
4. Consider the fabrication of a polyelectrolyte multilayer
composed of PEI and PAZO on a glass substrate. After
each deposition step, the mass of the film can be mea-
sured by QCM and thickness by ellipsometry. Predict
how these values change for a 10-bilayer film as the
following parameters are varied: (a) Decreasing the pH
of the polyelectrolyte solution from 7 to 4, (b) increas-

ing the salt concentration from 0 to 100 mM, and (c)
increasing the deposition temperature from 25°C to
30°C. For (a) you will need to look up various pKa val-
ues for groups on each polyelectrolyte.
Appendix
TOTAL INTERNAL REFLECTION (TIR)

When light is shone upon the boundary between two medi-
ums or substances of different refractive indexes (n1 ≠ n2), the
light passes through the boundary into the second medium
and is refracted toward the normal if n2 > n1 and away from the
normal if n1 > n2. This refraction obeys Snell’s law:
n1 sin θ1 = n2 sin θ2
(A.1)
where n1 and n2 are the refractive indexes of the two sub-

stances, θ1 is the angle of incidence, and θ2 is the angle of
refraction. In cases where the refractive index of the second
medium is less than that of the first (i.e., n1 > n2), there exists
an angle, called the critical angle θcritical, where the angle of
refraction is 90°, or where the light in the second substance
is refracted along the interface between the two substances.
The critical angle θcritical between any two substances can be
calculated by applying Snell’s law and performing some basic
algebra to get
⎛n ⎞
θcritical = arcsin ⎜ 2 ⎟ (A.2)
⎝ n1 ⎠
When incident light strikes the interface between the sub-

stances at an angle that is greater than the critical angle, the
light is reflected from the surface. This phenomenon is called
total internal reflection.
EVANESCENT WAVES
A description of total internal reflection using classical phys-
ics says that the incident light is completely reflected from the
surface between the two substances. In reality, however, some
of the energy from the incident light actually penetrates the
second medium to a small extent. This “portion” of the light
that enters the other medium is called an evanescent wave. The
evanescent wave (or evanescent field) decays rapidly, moving
281
282 Appendix
away from the interface between the two substances according

to the following equation:
− x /d p
E x = E 0e (A.3)
where Ex is the electric field amplitude of the evanescent wave

at a distance x from the interface, E0 is the electric field at
the interface, and dp is the penetration depth, defined as the
distance at which Eo is reduced to 1/e of its original value. If
the conditions of TIR that generate an evanescent wave are
known, then dp can be calculated as
λ
dp =
2
2 ⎛ n2 ⎞ (A.4)
2πn1 sin θincidence − ⎜ ⎟
⎝ n1 ⎠
where θincidence is the angle of incidence, and n1 and n2 are the

refractive indexes of the two mediums. We see that the pen-
etration depth dp of the evanescent wave can be calculated for
a given wavelength of light. In the visible region, the values
of dp typically range from 50 to 100 nm. Thus, the evanescent
wave can be an excellent probe of the area near the boundary
between the two substances (i.e., it is an excellent probe of
surface modifications), and as such serves as the basis for a
variety of nanomaterial characterization techniques such as
SPR, ATR-FTIR, and DPI.
Glossary
Adsorbate: A term used to describe an atom, ion, or molecule

that is either deposited onto a surface or is the species
to be adsorbed.
Adsorbent: Surface upon which an adsorbate is deposited.
Adsorption isotherm: A plot of how the amount of adsorbate
on a given adsorbent changes with pressure (for gas-
phase deposition) or concentration (for liquid-phase
deposition).
Aggregation: Buildup of material at an interface. The term also
describes the formation of clusters or an increased
local concentration of molecules (e.g., micelle) in bulk
phases.
Aggregation number: Number of molecules that comprise a
single micelle.
Amphiphile: Molecule (typically a surfactant) containing both
hydrophilic and hydrophobic domains.
Anisotropic: Something in which the refractive index depends
on the orientation of the propagating wave. The term
also refers to an assembly in which all of the molecu-
lar components point in the same direction.
Anti-Stokes shift: (a) Like the Stokes shift, this is the differ-
ence between absorbed and emitted photons. However,
in this case the emitted photon must exceed the energy
of the exciting photon. This extra energy stems from
the dissipation of thermal energy within crystalline
structures.
Anti-Stokes shift: (b) The frequency shift of Raman scattered
light to a higher energy.
Atomic force microscopy (AFM): A sensitive imaging tech-
nique, which exploits a small piezoelectric tip moni-
tored by a laser to detect minute changes in surface
topology.
Attenuation: The gradual loss in intensity due to flux through
a medium.
Ballistic transport: Refers to a nanostructure whose mean-free
path is less than its width. Ballistic transporters (typi-
cally metal nanowires) have very high conductance
values compared to those whose wire widths fail to
exceed the mean-free path.
283
284 Glossary
BET isotherm: An adsorption isotherm that describes multi-

layer film formation. This isotherm assumes that phys-
isorption on the substrate is infinite, that no interlayer
interactions exist, and that each layer can be described
by the Langmuir model.
Black lipid membrane: Lipid membrane assembled on an
aperture in a hydrophobic surface. The membrane is
referred to as “black” because it is dark when exposed
to light as photons reflected off the back half of the
bilayer destructively interfere with those bouncing off
the front half.
Bottom-up synthesis: In this approach, nanostructures are
built by continuously extending a thread of bound
molecules. Subsequent layers are built atop previous
ones exploiting intermolecular forces.
Bulk phase: Solvent region above or below an interface. The
term is used in context to adsorption. For example,
molecules move from the bulk phase and aggregate on
the surface.
Cassie-Baxter wetting: A wetting state in which water rests
upon nano- or micro-sized “pins” with air spaces in
between.
Charge overcompensation: Refers to the tendency of a charged
layer to attract excess counter ions during self-assem-
bly. In this way, layers can be subsequently built atop
one another with the overcompensated charge acting
as a medium for continued growth of molecules of
alternating charge.
Charge reversal: Integral to the iterative process of electro-
static self-assembly. A charged molecule can aggregate
atop a preexisting surface of opposite charge. The
resulting surface is the opposite charge of the original
layer, an instance of charge reversal. See also charge
overcompensation.
Chemisorption: Process of using chemical bonds to tether
molecules to a surface.
Chromophore: The region of a molecule that absorbs light and
undergoes an electronic transition.
Column chromatography: Method of purifying and separating
one chemical compound from another.
Conjugation: Alternating single and double bonds in linear
carbon chains. All carbon atoms are sp2 hybridized and
the electrons in the π-bonds are actually delocalized
along the entire carbon chain. See also pi-bonding.
Glossary 285
Constructive interference: A phenomenon in which light waves

combine to produce a more intense wave, or brighter
light area, than the individual waves. This type of
interference occurs when the waves are in-phase.
Contact angle: Angle formed between a drop of liquid and a
solid interface.
Contact angle hysteresis: The difference between a droplet’s
receding and advancing contact angle.
Converse piezoelectric effect: The ability of a crystal to oscil-
late due to applied alternating current.
Cooperative adsorption: An adsorption processes whereby the
presence of one adsorbate enhances the adsorption of
another. Also refers to the property set of positively and
negatively charged molecules, which do not deposit on
a surface individually, to stabilize electrostatic inter-
actions and allow for the deposition of both molecules
on a surface simultaneously.
Critical micelle concentration (CMC): Concentration of sur-
factant molecules in solution above, in which micelles
begin to spontaneously form.
Debye length: The characteristic decay length of the electro-
static potential for the Debye-Hückel model. Also used
as a rough approximation for the size of the electric
double layer.
Desorption: Process of a molecule bound to a surface becom-
ing detached from that surface and returning to the
bulk phase.
Destructive interference: A phenomenon in which light waves
combine to produce a less intense wave, or dimmer
light area, than the individual waves. This type of
interference occurs when the waves are completely
out-of-phase.
Direct piezoelectric effect: The ability of a crystal to produce
an electric field in response to mechanical stress.
Dispersion forces: Force between two instantaneously induced
dipoles.
Dissipation: Value for relating the energy lost to the energy
stored over the course of an event.
DNA microarray: DNA microarrays are surfaces that have
been coated with specific oligonucleotide sequences
and have been particularly useful in gene expression
studies. DNA microarrays vary in the number of oli-
gonucleotides as is determined by the nature of the
study; diagnostic DNA microarrays generally use tens
286 Glossary
of oligonucleotides, whereas those for research and

screening can have hundreds of thousands of oligo-
nucleotides on a single microarray.
Electrical double layer: The diffuse layer of counter ions
in a solution, which are associated with a charged
surface.
Evanescent field: Standing waves formed at the boundary
between two media with different wave motion prop-
erties. They decay exponentially with distance from
the boundary in which the waves are formed.
Exciton: A quasiparticle created by irradiation of photoac-
tive materials and consisting of an unpaired electron
and an imaginary “electron hole.” As incident light
is exposed to a dye interface, an in-phase oscillation
among all the layer’s oscillators occurs and an excited
domain (the exciton) forms, having absorbed the
energy from the incident photons.
Extinction coefficient: Also called molar absorptivity, it is a
parameter that helps define how strongly a substance
absorbs light. It is often given at a given wavelength
per mass unit or per molar concentration.
Fluorescence: A relaxation process in which the molecule
relaxes by reemitting light, generally of lower energy
than the light it absorbed. More specifically, it refers to
the light emitted when an electronic transition occurs
between electronic states of the same spin multiplicity
(e.g., singlet-singlet transitions).
Fluorescence interference contrast microscopy (FLIC): Monitors
the calculable modulation of fluorescence intensity due
to the interaction between a reflecting surface and fluo-
rescent objects in order to attain nanometer-accurate
height measurements. In this method, a film on a reflec-
tive surface is capped with a fluorescently tagged entity
and the assembly’s size is quantified based on the spe-
cific interaction between the surface and fluorophore.
Fluorescence resonance energy transfer (FRET): Mechanism
of energy transfer between chromophores whereby an
excited donor transfers an electron over a short (<10nm)
distance to an acceptor molecule through dipole−dipole
coupling.
Fluorophore: Fluorescent unit of a molecule, which absorbs
and emits energy at a specific wavelength.
Fractal: Geometric shape that can be split into sections that
are at least partially identical to a smaller replica of
the entire structure.
Glossary 287
Free electron model: In this model, the electron does not exist
as a discrete particle moving along the line. Rather, it
resembles a standing wave whose exact form depends
on the value of n.
Gravimetry: The measure of the strength of gravitational
fields.
H-type aggregate: One-dimensional molecular assembly in
which the dipole moments are aligned parallel to each
other but perpendicular to the line joining their cen-
ters. This is sometimes referred to as the “face-to-face
arrangement.”
Hard sphere model: A way to determine atomic radius by
assuming atoms in a solid are hard spheres and pack
closely together.
HOMO: Highest occupied molecular orbital.
Hydrodynamic radius: The radius of a sphere that diffuses at
the same rate as the molecule. Because most molecules
are not spherical, this radius is often smaller than the
effective rotational radius.
Hydrogen bonding: Attractive interaction between a hydrogen
atom and another electronegative atom.
Hydrophobic effect: Tendency of non-polar molecules to aggre-
gate in polar solvents to reach a thermodynamically
favored energy state.
Induced dipole interactions: Force existing between a perma-
nent dipole and a neighboring induced dipole.
Interface: The two-dimensional region of space at which two
different phases contact each other.
Interferometry: The study of the ways in which light waves
interact or interfere with each other.
Intermolecular force (F): Forces acting between sets of mol-
ecules, such as hydrogen bonding or dipole−dipole
interactions.
Interpenetration: In polyelectrolyte multilayer films, this
is the tendency of polycation and polyanion to com-
mingle to form highly homogeneous assemblies rather
than distinct, stratified layers. This process shields
the excess charge within the distinct layers, allowing
for tighter packing.
Ion-ion forces: Attractive or repulsive interaction between
ionic species.
J-type aggregate: One-dimensional molecular assembly in which
the dipole moments of the individual monomers are
aligned parallel to the line joining their centers. This is
sometimes referred to as the “end-to-end arrangement.”
288 Glossary
Laminar flow: The flow of fluid in parallel layers.

Langmuir adsorption isotherm: Equation relating the con-
centration or pressure of an adsorbate to the degree of
deposition on an adsorbent.
Langmuir-Blodgett deposition: The transfer of a monolayer to a
substrate using a trough, which compresses molecules
(often surfactants) on the surface of a liquid buffer.
The hydrophobic tails of the molecules will assemble
on the surface of the substrate as it is passed through
the compressed layer of surfactant. Multilayer films
can be created by switching direction of the substrate
entering the compressed layer.
Layer-by-layer deposition: Process of building nanoscale
assemblies by exploiting intermolecular or chemical
attraction forces. A common type of layer-by-layer
assembly (LbL) is electrostatic self-assembly, where,
for example, a positively charged polyelectrolyte
assembles spontaneously on an oxidized silica or
silanol surface.
Lennard-Jones potential energy curve: Mathematical approxi-
mation describing the interaction between two neutral
molecules. The curve describes both attractive and
repulsive interactions and the distance at which these
interactions cancel each other out.
Lipopolymer: Consists of a soft hydrophilic polymer layer with
lipidlike molecules at their surface. These structures
can insert into a phospholipid membrane and tether to
the polymer spacing.
Lithography: See nanolithography.
LUMO: Lowest unoccupied molecular orbital.
Mean-free path: Relates to the distance an electron travels
between subsequent collisions with other moving
particles.
Micelle: Spherical structures formed from the aggregation of
surfactant molecules above the critical micelle con-
centration (CMC).
Microemulsion: A clear solution of liquid water, oil, and sur-
factant without phase separation of hydrophobic and
hydrophilic entities.
Molecular Orbital (MO): Mathematical function that describes
the wavelike behavior of electrons in a molecule.
Molecular polarizability: The degree of electron density dis-
tortion due to an electric field.
Nanowire: A nanostructure that has a diameter on the scale of
a nanometer (10−9 m) and an unrestricted length.
Glossary 289
Nanolithography: Refers to a wide variety of nanoscale surface

manufacturing techniques for creating patterned sur-
faces, such as those found in semiconductor circuits.
Nonlinear optical: A term that indicates the dielectric polar-
ization of the media responds in a nonlinear manner
to the electric field of light.
Oligonucleotide: A short nucleic acid polymer comprised of
fewer than 20 bases.
Overlap repulsion: The interaction that results from the fact
that two atoms or molecules cannot occupy the same
point in space.
P-polarized light: Electric field vector polarized along the
plane perpendicular to the plane of incidence.
Path length: The distance in which light moves through a
material.
Pauli Exclusion Principle: States that no two electrons can
have the same four quantum numbers.
Phonon: A quantum mechanical description of a special type
of vibration, known as the normal modes in classical
physics. Phonons are the discrete amounts of energy
crystalline structures that these modes of vibration
can absorb. In this aspect of quantized energy, the
phonon is analogous to the photon.
Phosphorescence: A relaxation process when an excited elec-
tron undergoes nonradiative intersystem-crossing to a
slightly more stable excited state, called a triplet state,
before relaxing and releasing a photon of light.
Photobleaching (FRAP): Photobleaching is the process of
destroying a fluorophore with intense light. FRAP, or
Fluorescence Recovery After Photobleaching, is the
method quantifying lateral diffusion rates by monitor-
ing fluorescence under a microscope and measuring
the time it takes for a photobleached segment of a fluo-
rophore film to reorganize. Fluorophores reorganize,
intact ones replacing the inactivated ones. The hole in
fluorescence created by photobleaching disappears at
a rate determined by the lateral diffusion.
Photorelaxation: Process of an excited state chromophore dis-
sipating energy through release of a photon.
Physisorption: Process of exploiting weaker molecular forces
to deposit molecules on the surface. Generally weaker
and less specific than chemisorption.
Pi bonding (or π-bonding): Bond resulting from the overlap of
atomic orbitals that are in contact through two areas
of overlap above and below the internuclear axis. For
290 Glossary
example, in sp2 hybridized carbon atoms, the unhy-

bridized 2p orbitals on neighboring carbons overlap to
form the π-bond.
Pi−pi stacking: Packing of large, often aromatic molecules
whereby favorable overlap across pi systems of differ-
ent molecules occurs.
Plasma oscillations: Rapid oscillations of the electron density
in a conducting medium.
Plasmon: The smallest unit of plasma oscillations; used for
quantization.
Poisson-Boltzmann equation: Calculates the actual distribu-
tion of the counter ions at equilibrium.
Polarizability: Defined as the susceptibility of an atom or mol-
ecule’s electron cloud to distortion by an external elec-
tric field.
Polyelectrolyte: Polyelectrolytes are polymers, or chains of
molecules, which contain free ions that make them
electrically conductive. Soluble in water, polyelectro-
lytes become charged when in solution, and are often
countered by a salt ion of opposite charge. The amount
of charge on a polyelectrolyte determines whether it
is classified as strong or weak. Strong polyelectrolytes
are fully soluble, whereas weak polyelectrolytes, with
fractional charge, are only partially soluble.
Polymer cushioned phospholipid bilayers: Allow bilayers to
effectively be decoupled from the surface by minimiz-
ing interactions with the underlying substrate.
Potential energy (V): Energy stored within a system due to its
position in a force field.
Quantum dots: Quantum dots are inorganic semiconductor
nanoparticles, typically 2−10 nm in size. The excitons
(electron-hole pairs) of quantum dots are confined to
three dimensions. Quantum dots typically consist of a
core, shell, and final coat, as shown in Figure 5.10. They
are characterized by the nature of each of these layers,
their size and aspect ratio, their quantum efficiency in
optical materials, and their coercivity in magnetic mate-
rials. Properties of an ensemble of quantum dots are
additionally determined by particle size distribution
and differences in morphology within the ensemble.
Raman scattering: The scattering of light in which the light
increases or decreases in energy, often due to a change
in vibrational energy of the scattering molecule.
Rayleigh scattering: The scattering of light without a change
in its energy. The degree of scattering is a relationship
Glossary 291
between the wavelength of light being scattered and

the electric polarizability of the particles doing the
scattering.
Refractive index: Ratio of the velocity of light in a vacuum to
the velocity of light in a medium.
Resonant frequency: The frequency at which the local ampli-
tude of oscillation is a maximum.
S-polarized light: Electric field vector polarized along the
plane of incidence.
Scanning tunneling microscopy (STM): A sensitive imag-
ing technique that exploits a small tunneling current
between a surface and a tip to detect minute changes
in surface topology.
Second-harmonic generation: A nonlinear optical effect in
which the frequency of light is doubled when passed
through a material.
Self-assembled monolayer (SAM): A surface coverage that forms
when molecules are spontaneously attracted to a func-
tional surface, often by electrostatic or van der Waals
forces. SAMs can also result through chemisorption.
Self-assembly: The process whereby molecules in the bulk
phase spontaneously organize either in the bulk phase
or on a surface.
Sigma bonding: A strong covalent bond characterized by sym-
metry with respect to rotation about the bond axis.
Small unilamellar vesicles (SUVs): Tiny (50−200 nm), single-
walled vesicles of uniform diameter.
Soft sphere model: More realistic way of modeling an atom,
this model assumes atoms are compressible to some
degree and do not have completely rigid boundaries.
Spectroscopy: The study of the absorption of transmission of
light through a medium as a function of either wave-
length or frequency.
Spot density: Number of spots that can be placed in a given
area of a DNA microarray, varies greatly among in situ,
non-contact, and contact printing methods.
Steric hindrance: Stress in a molecular structure or assem-
bly that arises when oversized functional groups are
forced into too small a space. Such stress can result in
torsional strain.
Stokes shift: (a) The difference between the excitation pho-
ton wavelength necessary to induce an electronically
excited state and the wavelength of the subsequently
emitted photon. (b) The frequency shift of Raman scat-
tered light to a lower energy.
292 Glossary
Sum-frequency generation: A nonlinear optical effect when two

photons of different frequencies combine to produce a
photon of the sum frequency when passed through a
material.
Superhydrophobic surface: Highly hydrophobic surface with
water contact angles exceeding 150°.
Supramolecular chemistry: The study of chemical compounds
comprised of a distinct number of cooperating mol-
ecules. Supramolecular machines are comprised of a
“cooperating set” of molecules and exploit intra- and
intermolecular forces to complete a specialized task.
Surface functionalization: Deliberate attachment of specific
molecules to a surface in order to allow for the specific
binding of a subsequent adsorbate.
Surface tension: The work done in increasing the area of a sur-
face by transporting a molecule from the bulk phase
to that surface. A measurement of the cohesion of like
molecules at an interface. The tendency of a surface to
contract due to energetic factors.
Surfactant: Compounds that lower the surface tension of a
liquid and form micelles above the critical micelle
concentration (CMC).
Thio-: A prefix that indicates an oxygen atom in the common
compound is instead sulfur. For example, ether has a
general structure ROR’ and a thioether would have a
general structure RSR’.
Thiol: Thiols are compounds having the structure RSH (where
R ≠ H).
Top-down synthesis: In this approach, structures are synthe-
sized starting from a bulk substrate and chiseled down
until only a field of desired assemblies remains.
Total intermolecular potential energy: The sum of all the dif-
ferent interactions between two molecules.
Transducer: A device that converts one type of energy to
another.
Transmittance: The ratio of the intensity of a light beam after
passing through a sample to the original intensity of
the light beam.
Van der Waals interaction: Attraction or repulsion between
molecules due to interaction with covalent bonds or
electrostatic forces.
Viscoelasticity: The property of materials to both resist flow
when a stress is applied and rapidly return to their
original state once the stress is removed.
Glossary 293
Waveguide: An object that confines and directs waves by

means of total reflection of the wave within itself.
Wenzel wetting: A wetting state in which water rests upon
a surface whose morphology has been altered so that
in a given area, water is in contact with more surface
than if the surface were completely flat.
Physical sciences / nanotechnology
“The writing ... is very fluid. The problems and figures are good. Overall, I
learned a great deal….”
—Professor Lisa Klein, Rutgers University
“I believe the textbook will serve students well in their goal to gain a greater
understanding of why nanoscaled systems are of great interest, how they are
fabricated, and how they are characterized using a wide variety of analytical
instrumentation very commonly found in university and industrial settings.”
—Professor Marcus D. Lay, University of Georgia
With a selective presentation of topics that makes it accessible for students

who have taken introductory university science courses, Understanding Nano-
materials is a training tool for the future workforce in nanotech development.
This introductory textbook offers insights into the fundamental principles that
govern the fabrication, characterization, and application of nanomaterials.
Provides the Background for Fundamental Understanding
Assuming only a basic level of competency in physics, chemistry, and biology,

the author focuses on the needs of the undergraduate curriculum, discussing
important processes such as self-assembly, patterning, and nanolithography.
His approach limits mathematical rigor in the presentation of key results and
proofs, leaving it to the instructor’s discretion to add more advanced details or
emphasize particular areas of interest.
With its combination of discussion-based instruction and explanation of

problem-solving skills, this textbook highlights interdisciplinary theory and
enabling tools derived from chemistry, biology, physics, medicine, and engineering.
It also includes real-world examples related to energy, the environment, and
medicine.
Author Malkiat S. Johal earned his Ph.D. from the University of Cambridge in England. He later served
as a post-doctoral research associate at Los Alamos National Laboratory, New Mexico, where he
worked on the nonlinear optical properties of nanoassemblies. Dr. Johal is currently a professor and
researcher at Pomona College in Claremont, California. His work focuses on the use of self-assembly
and ionic adsorption processes to fabricate nanomaterials for optical and biochemical applications.
73109
6000 Broken Sound Parkway, NW

Suite 300, Boca Raton, FL 33487
711 Third Avenue
an informa business New York, NY 10017
w w w. c r c p r e s s . c o m 2 Park Square, Milton Park
Abingdon, Oxon OX14 4RN, UK
w w w. c r c p r e s s . c o m

Vdoc - Pub Understanding Nanomaterials

Uploaded by

Copyright:

Available Formats

Vdoc - Pub Understanding Nanomaterials

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Vdoc - Pub Understanding Nanomaterials

Uploaded by

Copyright:

Available Formats

understanding

Boca Raton London New York

CRC Press is an imprint of the

No claim to original U.S. Government works

International Standard Book Number-13: 978-1-4398-9112-4 (eBook - PDF)

Chapter 1 A Brief Introduction to Nanoscience . . . . . . . . . . . 1

Chapter 2 Intermolecular Interactions and

2.4.1 The Particle in a Box Model . . . . . . 37

Chapter 3 Rudiments of Surface Nanoscience . . . . . . . . . . . 55

Chapter 4 Characterization at the Nanoscale . . . . . . . . . . . 97

4.3 Ellipsometry . . . . . . . . . . . . . . . . . . . . . . . 108

4.6.5.2 Attenuated Total

4.9.4.1 History and Principles

Chapter 5 Types and Uses of Some Nanomaterials . . . . . . 207

5.4.2.2 Reactive Methods in

5.10.5 Arrays of Supported Bilayers

advanced-level classes, although I strongly suggest training

Malkiat S. Johal is Asso­

1.1 THE NEED FOR NANOSCIENCE EDUCATION

much of the material emphasizes liquids and surface science).

1.2 THE NANOSCALE DIMENSION AND

CH3(CH2)11 group OH group

Figure 1.1 An assembly of 48 dodecanol molecules forming a hypothetical aggregate

such as polymers and proteins have sizes approaching 10 nm;

his legendary speech, “There’s Plenty of Room at the Bottom,”

usually unknown, is another example of an extended entity.

1.4 SUPRAMOLECULAR SCIENCE

micellar nano-reactor synthesis), and physics (e.g., organic

1.5 SOURCES OF INFORMATION ON NANOSCIENCE

• Deffeyes, K. S. and Deffeyes, S. E. Nanoscale: Visualizing

book focuses on the technical and business aspects of

building blocks is driven by a combination of thermodynamic

where r is the distance between the two molecules and may

govern self-assembly processes, protein-protein interactions,

2.1.1 Ion-Ion Interactions

where ε0 is the permittivity of free space (8.854 × 10 –12 m–3 kg–1

Equation 2.3 is obtained by differentiating Equation 2.2, and

Example 2.1 The Coulombic Energy Between Ions

The ionic radius of Na+ is determined to be 95 pm and the ionic

rNa–Cl = 95 pm + 181 pm = 2.76 × 10 –10 m

Each ion has a formal charge of +1 or –1, so

qNa+ = ze = (+1)(1.60217 × 10 –19 C) = 1.60217 × 10 –19 C

qCl– = ze = (–1)(1.60217 × 10 –19 C) = –1.60217 × 10 –19 C

Then, using Equation 1.2,

One example of ion-ion interactions can be found in the for-

in thickness on its surface. In a network of surface SiO2 groups,

2.1.2 Ion-Dipole Interactions

Polar molecules such as H2O are often characterized by

where L is the distance separating the partial positive and par-

Table 2.1 Dipole Moments of Some Common Molecules

Figure 2.4 A schematic depiction of the variables involved in an ion-dipole interaction.

where µ is the dipole moment, qion­ is the charge on the ion, r 12

the interaction between molecules with electrical multipoles

2.1.3 Dipole-Dipole Interactions

where θ1, θ2, and ϕ are defined in Figure 2.5.

Figure 2.5 A schematic depiction of the variables involved in a dipole-dipole interac-

region of the other. This anti-parallel orientation allows the

2.1.4 Interactions Involving Induced Dipoles

Ion Induced Dipole

Malkiat S. Johal is Asso

1.2 THE NANOSCALE DIMENSION AND

where µ is the dipole moment, qion is the charge on the ion, r 12

2.2 ELECTROSTATIC FORCES BETWEEN SURFACES:

where Ψ0 is the potential at the charged surface.

2.2.3 Interactions Between Charged

2.4 SIMPLE MODELS DESCRIBING