Atom Probe Tomography

Atom Probe
Tomography
Essential
Knowledge
Briefings First Edition, 2017
2 Atom Probe Tomography
Front cover image: three-dimensional image of a NiFeAlCr-based

superalloy showing three phases in the microstructure: a primary
Fe-rich phase (alpha-1), a secondary B2 NiAl phase (alpha-2), and
a secondary alpha-1 precipitate phase within the alpha-2 phase. The
secondary alpha-1 precipitates do not form within ~15nm of large-
scale primary regions, ie there is a denuded zone or precipitate-free
zone near the primary alpha-1 phase. Detailed knowledge of
how these complex microstructures vary with composition and
processing allows better understanding of precipitation processes
and leads to improvements in the engineering of such materials.
The Local Electrode Atom Probe (LEAP®) data was acquired at 50 K
in voltage pulsed mode. Only a fraction of the ions are displayed
for clarity.
© 2017 John Wiley & Sons Ltd, The Atrium, Southern Gate,
Chichester, West Sussex PO19 8SQ, UK
Microscopy EKB Series Editor: Dr Julian Heath
Spectroscopy and Separations EKB Series Editor: Nick Taylor
Atom Probe Tomography 3
CONTENTS
4 INTRODUCTION
6 HISTORY AND BACKGROUND
12 IN PRACTICE
25 PROBLEMS AND SOLUTIONS
30 WHAT’S NEXT?
About Essential Knowledge Briefings

Essential Knowledge Briefings, published by John Wiley & Sons,
comprise a series of short guides to the latest techniques, applications
and equipment used in analytical science. Revised and updated
annually, EKBs are an essential resource for scientists working in
both academia and industry looking to update their understanding
of key developments within each specialty. Free to download
in a range of electronic formats, the EKB range is available at
www.essentialknowledgebriefings.com
INTRODUCTION
Atom probe tomography (APT) has always offered the
enticing prospect of being able to determine the identity and
position of nearly every atom in a material, providing the ulti-
mate in elemental analysis. Actually realizing that prospect has
taken much time and effort. But with the latest APT systems able
to generate three-dimensional (3D) information at near atomic
resolution and possessing a chemical sensitivity that can reach
parts per million, the technique is now being used to analyze
everything from the oldest minerals known to man to the latest
nanomaterials.
APT involves applying either ultra-fast voltage pulses or laser
pulses to a needle-shaped specimen, stripping away atoms located at
the tip of the needle and converting them into charged ions in a
process known as field evaporation. These ions are then accelerated
by an electric field towards a position-sensitive detector that
registers the time it takes each ion to travel from the sample to the
detection system, as well as its impact position.
Because the energy applied to the atoms at the surface is
known, each ion’s mass-to-charge ratio can be calculated from its
travel time, and this mass-to-charge ratio is usually sufficient to
determine the identity of the original atom. APT is a destructive
technique that removes successive layers of atoms from a sample.
The impact position and order of detection can reveal the original
location of each atom in the sample. Combining these sources of
information produces a 3D image showing the atomic structure
and chemical composition of the sample, making APT particularly
effective at analyzing buried features and interfaces between
different regions.
When APT was originally developed, its applications were

limited, as it was relatively slow and could only analyze good
electrical conductors like metals. Much has changed over the past
10–15 years; APT can now rapidly generate useful information
on many different types of samples, including semiconductors,
ceramics, and even complex structures such as geological and
biomineral materials.
This Essential Knowledge Briefing (EKB) will introduce APT
and highlight its impressive abilities. It begins with a brief history
of the technique, including the most important developments
that transformed the 1D atom probe into a high-resolution and
highly sensitive 3D instrument. This is followed by a detailed
explanation of the steps involved in conducting an analysis
with modern-day atom probes, from sample preparation to 3D
reconstructions.
The EKB also considers practical issues and potential
problems that may occur during an analysis and offers possible
solutions. In addition, it details several examples of how this
technique is currently being used by scientists in their research,
and concludes by looking forward to potential future develop-
ments, from automated sample preparation to improved detector
performance.
HISTORY AND BACKGROUND

The history of the atom probe technique can be traced back
to the early 1950s when German physicist Erwin Wilhelm Müller
developed a new type of microscope, known as the field ion
microscope (FIM). Utilizing the same basic approach as modern
APT, this instrument applied a positive voltage to a needle-shaped
specimen to evaporate desorbed gas ions from the surface,
producing a field ion image visualized directly on a phosphor
screen. With this microscope, Müller and his PhD student
Kanwar Bahadur became the first scientists to observe individual
atoms directly.
One of the main limitations of FIM, however, was that
although it could produce an image of the individual atoms in
a sample, it couldn’t determine their chemical identity. To
introduce this ability, Müller, Panitz and colleagues added a
time-of-flight (TOF) mass spectrometer to their microscope in
1969. A TOF mass spectrometer can determine the mass-to-
charge ratio of ions with a known energy based on the time it
takes for them to travel the fixed distance between the sample
and the detector (TOF), with heavier ions taking longer to travel
than lighter ions. The identity of the original atom can then
be calculated from the ion’s mass-to-charge-state ratio; for
example, a mass-to-charge ratio of 54 Da (dalton, or equivalently
AMU) corresponds to a specific isotope of iron (54Fe), while 206 Da
corresponds to a lead isotope (206Pb).
For this technique to work, Müller and his colleagues now
needed to apply the voltage as a rapid series of pulses, so that each
voltage pulse could mark the start of a TOF measurement for the
evaporated ions. This reliance on voltage pulses meant that APT
was initially restricted to analyzing good electrical conductors,

particularly metallic alloys: refractory metals, steels, aluminium
alloys, nickel super alloys, metallic glasses, etc.
In 1980, Tien Tsong and Gary Kellogg from Sandia National
Laboratories in the USA found a way to conduct field evaporation
with laser pulses, opening up APT to non-conducting materials
such as semiconductors, oxides, ceramics, polymers, and geological
and biological materials. The first commercial APT systems, which
were voltage-pulsed, started to appear in the mid-1980s.
The next breakthrough came in the late 1980s with the devel-
opment of a position-sensitive detector by Alfred Cerezo, Terence
Godfrey and George Smith at Oxford University in the UK, which
transformed Müller’s 1D atom probe into a rich and detailed 3D
atom probe. The position-sensitive detector was simultaneously
able to measure the TOF of ions and their point of impact, which
was directly related to the original location of the atom in the
sample. Now researchers could analyze not only a sample’s
chemical composition, but also its morphology and the spatial
distribution of its elements.
This novel version of APT was extremely successful, but still
had limitations: it could take days to collect data and had a limited
field of view (around 25 nm). These limitations were overcome in
the early 2000s with the introduction of a local counter electrode.
By reducing the distance between the counter electrode and the
specimen to as low as 50 µm, higher electric fields could be
produced with lower voltages, allowing an increase in the voltage
pulse rate and expanding the field of view. Known as Local
Electrode Atom Probe (LEAP®), this approach is now employed
by nearly all modern commercial systems (Figure 1).
Local
Detector electrode
Load lock Laser objective
Laser source
Horizontal
transfer
Laser conditioning
Buffer/storage
Ion
Specimen Ion pump flight
TSP
Vertical
transfer Cryo
cooling Flight path details
Figure 1. Schematic of a modern atom probe
Such modern systems also benefit from further advances in the

detection instrumentation. Once an ion evaporates from a sample,
it now first hits a transducer amplifier known as a microchannel
plate (MCP), which converts the single ion into a flow of electrons. By
amplifying the signal in this way, the MCP is able to detect single
ions in both time and location with high precision and high speed.
The structure of the microchannel plate is relatively simple,
consisting of a thin metallized glass disc penetrated by micro-
channels with a high secondary electron coefficient, which can
amplify the signal by a factor of about 1000. To further increase
this amplification, some systems possess two of these plates,
producing an assembly that can achieve an amplification of
between 100 000 and 1 000 000 electrons.
The resulting cloud of electrons then travels through a delay
line detector, creating electromagnetic (voltage) pulses that can be
processed by standard electronics. Modern APT detectors contain
Unphysical
region
Figure 2. Analytical resolution versus detection limit for APT and various other analytical
techniques
three serpentine wires arranged in different directions, allowing

them to pinpoint, with great accuracy, the original position of
each ion in the sample. The latest APT systems have a lateral
resolution of 0.3–0.5 nm and a depth resolution of 0.1–0.3 nm; this
compares favourably to other analytical techniques (Figure 2).
Alongside these advances in instrumentation have come
associated improvements in analyzing, interpreting and recon-
structing the resultant data. Generating meaningful information
from the raw APT data offers a number of challenges. For
instance, different ions can have similar or identical mass-to-
charge ratios, making identification difficult, while variations
in the field required to evaporate different atoms can interfere
with accurately determining their relative positions in a sample.
Figure 3. A single windows-based program calibrates, reconstructs and analyses the raw
APT data
Nevertheless, the latest APT software can generally deal with

these and other challenges, and present the resultant analytical
information in a range of different formats (Figure 3). These
include raw mass spectra in which ions with distinct mass-to-
charge ratios appear as individual peaks, 3D images of the sample
showing overall elemental composition, individual slices through
those 3D images, the distribution of specific elements in the
sample, and the chemical composition of distinct volumes.
All of these advances have led to APT becoming a viable

analytical technique in many different fields, including nano-
sciences and microelectronics, and being applied to the study of
samples that would once have been impossible. For example, in
geology APT is now being used to study rocks and fossils. In
medicine and dentistry, it is being used to analyze the junction
between prosthetic and biological tissue, to determine how
kidney stones are formed, and to assess the development of
cavities in teeth.
IN PRACTICE
Before the advent of modern commercial instruments, data
collection was the main bottleneck in atom probe analyses, taking
days to obtain sufficient data. However, with all the improvements
to APT introduced over the past few years, this step can now take
less than an hour. One part of the process that remains less automated,
and therefore needs more operator input, is sample preparation,
which can be a delicate and time-consuming operation.
To ensure effective field evaporation, the targeted region of
interest (usually around 100 x 100 x 100 nm3 in volume) needs to
be needle shaped. One commonly used method for producing this
shape is known as electropolishing, which can be applied to
metals and metal alloys. An example electropolishing method
uses a sample with a thin, wire-like geometry that is dipped into a
thin layer of electrolyte floating on top of a dense inert liquid,
such that only the mid-section of the sample is immersed in the
electrolyte. Applying an electrical current then causes material
to be removed, or polished, from this immersed mid-section. A
number of other geometries and chemistries can also be used for
electropolishing depending upon the material being analyzed.
As more and more material is removed from the mid-section
(necking), the sample eventually breaks to leave one end with the
required needle shape and sharpness (Figure 4f). By using gradually
less concentrated solutions of the electrolyte to ensure a con-
trolled break, this method can also produce a needle shape at both
ends. Although the electropolishing process is often monitored
using an optical microscope, the sample can also be checked with
a transmission electron microscope (TEM) or a scanning electron
microscope (SEM).
Figure 4. Specimen preparation: (a–c) standard FIB lift-out and mounting of a specimen;
(d–e) sharpening the sample with the region of interest left at the very apex; (f) electropolishing
of a wire geometry sample
Electropolishing may be relatively easy and cheap to per-

form, but it is not suitable for every material, particularly
non-conducting materials or where the region of interest is very
site specific. This is why a second method known as focused ion
beam (FIB) milling, which involves essentially carving samples
into the desired shape using a beam of ions, is often used. Almost
any material can be shaped this way, including semiconductor
devices, carbon filaments, ceramics, insulators, and even biologi-
cal and geological samples. FIB milling can also produce the
needle shape in specific areas of a sample, allowing APT to target
grain boundaries, interfaces or even single transistors. FIB is
often combined with SEM in commercial systems, allowing the
milling process to be monitored in real time. Because such
FIB-SEM systems can also perform other types of analysis, such
as energy dispersive X-ray spectroscopy (EDX), a whole range of
information can be gathered during sample preparation.
With FIB milling, APT samples are typically prepared using

the lift-out method. This involves using an ion beam to carve out a
small wedge of material, which is extracted, mounted on the end of
a sample carrier and then milled into the required needle shape.
The ion beam typically employs currents of between 50 pA and
10 nA; higher currents produce faster milling, while lower currents
achieve a higher resolution.
In a process that is very similar to the FIB preparation of
samples for TEM analysis, the region of interest is first covered
with a protective layer to avoid accidental damage. The ion beam
is then used to cut angled trenches on either side of this wedge-
shaped region of interest, followed by a cut at the base of the wall
formed by the initial trenches. This cantilevered section is then
removed with a micromanipulator and attached to a carrier tip
with FIB-deposited material (eg tungsten, platinum). The same
steps can be repeated to produce multiple samples from the same
wedge of material.
At this stage, the sample still needs to be milled into a needle
shape. This is typically done using a multi-step process: first, an
annular milling pattern produces a cylinder around 1–2 µm in
diameter; next, patterns with sequentially lower diameters and
lower beam energies are applied to obtain a conical shape with a
radius of curvature of less than 100 nm at the apex (see Figure 4).
Once the sample is ready for analysis, it is placed in the
specimen mount, which is then inserted into the ultra-high
vacuum (UHV) system and cooled to cryogenic temperatures of
20–100 K. The analysis starts when a standing positive voltage is
applied to the sample, together with either nanosecond or shorter
laser pulses or nanosecond voltage pulses. Older instruments were
only able to collect thousands to tens-of-thousands of ions for

each analysis, but today’s systems can achieve a significantly
higher rate of detection, collecting hundreds of millions of ions.
As APT is a destructive technique, the sample will inevitably
change shape during analysis. In fact, as soon as the first few layers
are field-evaporated, the radius of the tip tends to increase and
the sample becomes blunter. For this reason, the standing voltage
needs to be increased gradually to maintain a constant electric field
and to avoid fluctuations in the evaporation rate. In modern APT
systems, this gradual increase in voltage is controlled automatically.
The raw data from an analysis includes: an x,y value for the
position of each impact; the voltage applied at the moment of
evaporation; and the TOF. This raw data is typically analyzed
with the commercial Integrated Visualization and Analysis
Software (IVASTM) from CAMECA®, but other software tools, such
as GPM3D-soft created by researchers at the University of Rouen
in France, are being developed as well. Figure 5 shows examples of
the type of data that can be produced by an APT analysis.
Most of the time, the process of determining chemical
composition from the spectral data is relatively straightforward.
Exceptions occur when two or more ions impact the detector in
close proximity in space and time, or when the spectral peaks
of two ions directly overlap. Situations where this may occur
and possible ways to solve it are discussed in the Problems and
Solutions chapter.
Optimizing an APT analysis for a specific material can be
time consuming, as it involves optimizing a whole range of sample
preparation and analysis conditions, including specimen size and
shape, acquisition pulse rate, base temperature and detection rate.
60
50
Concentration (%)
40
30
20
10
0
0 5 10 15 20 25
Distance (nm)
Figure 5. Example APT data showing the ability to observe microscale and nanoscale
features with the same resolution and compositional accuracy
Optimum conditions tend to differ for different materials, and

determining them is mainly a matter of trial and error (see Table 1
in Problems and Solutions for a list of analysis conditions). Once
these conditions have been defined for a material, however, the
same methodology can be adopted for every analysis. One reason
for this simplicity is that, generally speaking, APT does not
require any chemical standards, as the delay line detector detects
all ions with equal probability. This means there is no need to
include sensitivity factors in the analysis.
Reconstructing data for a 3D analysis can be more complex.
There are several methods for doing this, but the most commonly
used method assumes the needle-shaped sample has a hemispherical
end, on which the position of the detected ions is back-projected.

This assumption is not consistent with complex samples con-
taining different layers or specific features with wildly variable
evaporation efficiencies; however, these oversimplified assump-
tions often still yield reasonable reconstructions in most cases.
There are occasions where the variety of features in the sample
is too extreme to get minimally accurate reconstructions with
the standard method, so more complex mathematical models
are required.
CASE STUDY 1. Combining APT and EDX

Researchers are becoming increasingly aware that the key
to improving efficiency in electronic devices is to be found at the
nanoscale. However, current methods for analyzing materials
at the nanoscale, such as X-ray diffraction and TEM, cannot
guarantee accurate readings in terms of composition and structure.
So, when Bastien Bonef and his team at the Université Grenoble
Alpes in France wanted to analyze nitride-based alloys at this
scale, they decided to adopt a new approach, combining APT
with EDX.1
In theory, APT seems to be the ideal technique for determin-
ing chemical composition and 3D structure at the nanoscale, but
current limitations in data reconstruction for layered samples (see
Problems and Solutions) meant that it might not prove that
reliable for nitride-based alloys such as AlGaN.
‘In the case of APT on nitrides, the problem is can we trust
the quantification?’ explains Bonef. ‘In most cases, APT will give
you the correct quantification, but it’s not always the case.’ This is
why EDX is seen as a welcome addition to the system. ‘Combining
EDX with APT just makes the analysis much more reliable and
using EDX on the structure gives us a reference to perform the 3D
reconstruction of the APT volume.’
Preliminary results seem promising. While APT could clearly
identify two separate layers in the alloys, the results improved when
the analysis conditions for APT were fine-tuned using information
provided by EDX. This included defining the most efficient laser
energy to use during data collection, as well as a way to calibrate
the subsequent 3D reconstruction.
‘People were originally skeptical about APT because of the time

required to prepare the sample, perform the analysis and the data
treatment,’ says Bonef. ‘In the case of nitrides, the time required to
get to the final data is not that long and indeed is much faster than
getting this information from TEM.’
1. Bonef B, Lopez-Haro M, Amichi L, et al. Composition analysis of

III-nitrides at the nanometer scale: comparison of energy dispersive
x-ray spectroscopy and atom probe tomography. Nanoscale Res Lett
2016;11:461. (http://dx.doi.org/10.1186/s11671-016-1668-2)
a b c
Capping layer
AlGaN 7%
6 nm
AlGaN 5%
12 nm
X 20
GaN
3 nm
AlGaN 12%
3 nm
Substrate
5 nm 20 nm
(a) Schematic of the AlGaN superlattice showing its chemical composition. (b) Scanning
transmission electron microscopy image of one period of the superlattice. (c) APT 3D
reconstruction of four periods of the superlattice
CASE STUDY 2. Applying APT to geological

samples
When he first encountered a LEAP atom probe almost 10 years
ago, geologist Alberto Pérez-Huerta at the University of Alabama
in the USA was instantly fascinated by this technology. Up to
that point, the university’s atom probe had solely been used to
study metals and alloys. Very quickly, however, Pérez-Huerta
and his colleagues saw the potential for applying APT to
geological samples in order to probe the chemistry of minerals at
the nanoscale.
‘This instrument was designed for metals, but it has improved
so much that you can do non-conductive materials, such as minerals
and ceramics and insulator materials,’ says Pérez-Huerta. ‘We
believe it gives us the capacity to look at the atomic scale composi-
tion of any mineral out there.’
The reason APT is so valuable to geologists is that as well as
revealing a mineral’s chemical composition, it can also help deter-
mine how the mineral formed. For example, explains Pérez-Huerta,
‘we can look at volcanic rock and understand things about the
eruption, or we can look at the chemistry of coral and understand
about climate changes in the past’.
After a few tests to ensure APT could reproduce the composi-
tion of standard minerals, like carbonite,1 Pérez-Huerta and his
colleagues are now set on expanding the type of geological samples
analyzed with APT. This list already includes minerals like pyrite,
magnetite, quartz and calcite, as well as biominerals like biocalcite
and bioaragonite from seashells and fossils like egg shells. They are
now even contemplating using APT to analyze bone tissue.
One practical limitation lies in sample preparation. ‘It’s very

difficult to prepare samples, it takes a long time and we have to be
very careful because many of these minerals have cleavage planes
and are brittle,’ says Pérez-Huerta. ‘Minerals start forming these
nanocracks and when you put them in the LEAP and start ionizing
them, these nanocracks are going to be areas of failure.’
Despite these problems, Pérez-Huerta is determined to get the
rest of the geological community to realize the impressive potential
of APT. ‘It still blows my mind we can see atoms in three dimen-
sions,’ he admits.
1. Pérez-Huerta A, Laiginhas F, Reinhard DA, et al. Atom probe tomography

(APT) of carbonate minerals. Micron 2016;80:83–9. (http://dx.doi.org/
10.1016/j.micron.2015.10.001)
Example of the FIB lift-out of a single geological crystal for analysis by APT
CASE STUDY 3. Admiring diamonds with APT

Dieter Isheim at Northwestern University in Evanston, USA,
first became curious about diamonds a few years ago. Debat-
ing whether the nanoscale diamonds found in some meteorites
predated the formation of the sun, he and his team decided to
measure the ratios of two different isotopes of carbon (carbon-12
and carbon-13). Presolar material is known to be characterized by
isotopic abundances markedly different to those found in substances
formed after the sun formed. As one of the only analytical tech-
niques capable of distinguishing isotopes in individual nanoscale
objects, APT was an obvious choice for this work.
During their analysis, however, the researchers noticed a
unique behavior when the nanodiamonds were placed under a high
electric field. ‘With metals, [APT] is easy because metals have such
a high concentration of free charge carriers that the electric field
is always concentrated at the surface’, says Isheim. ‘The problem
with diamonds is that they’re dielectric. This means the electric
field penetrates into it and it’s not concentrated at the surface, and
the residual field at the surface is not strong enough to cause field
evaporation.’
The solution that he and his team came up with involved
applying a pulse energy significantly higher than normal, in order
to promote impact ionization. This way, explains Isheim, ‘you
can create charge carriers which can push the electric field to the
surface, so that field evaporation can happen’.
Using this novel APT method, Isheim and his team discovered
that the vast majority of these meteoritic nanodiamonds didn’t form
before the sun after all,1 and have since gone on to use the same
a Carbon atoms
7.2 nm
Platinum atoms
160 nm
Carbon-13 Carbon-12
b
75 nm
(a) Individual nanometer-sized meteoritic nanodiamonds (carbon atoms: black

dots), atomically resolved in an APT reconstruction.1 For preparation purposes, the
nanodiamonds were dispersed in a platinum matrix (orange dots). (b) Atomically and
isotopically resolved APT reconstruction of a nano-layered diamond material with
alternating layers of carbon-13 (blue) and carbon-12 (red) atoms2
method to test diamonds in other applications. For example,

specially designed diamonds with alternating layers of carbon-12
and carbon-13 are being investigated as novel materials for micro-
electronics. While diamond may not have been a popular material
in this field so far, its thermal, mechanical and electrical proper-
ties make it an ideal choice for increasing computer reliability and
allowing the use of high-power systems.
‘It turns out the layers are not perfectly isotopically pure and
with APT we can measure not only the layer composition but also
the transition at the interface,’ says Isheim. In this case, APT can
accurately determine the concentration and distribution of each

isotope, which is essential for understanding how these specially
designed diamonds perform under different conditions.2
1. Heck PR, Stadermann FJ, Isheim D, et al. Atom-probe analyses of

nanodiamonds from Allende. Meteorit Planet Sci 2014;49:453–67.
(http://dx.doi.org/10.1111/maps.12265)
2. Mukherjee S, Watanabe H, Isheim D, et al. Laser-assisted field evapora-
tion and three-dimensional atom-by-atom mapping of diamond isotopic
homojunctions. Nano Lett 2016;16:1335–44. (http://dx.doi.org/10.1021/
acs.nanolett.5b04728)
PROBLEMS AND SOLUTIONS

APT may be a particularly sensitive analytical technique at
the nanoscale level, but it still has some important limitations.
For a start, researchers still don’t completely understand the
process of field evaporation and how it works with different
elements. This can have a major impact, especially for samples
made up of different layers or containing a particular feature with
a different chemical composition, as the data analysis can become
complicated at the interface between these different regions.
A poor understanding of factors affecting field evaporation
also means that detection efficiency may be lower than expected.
To detect an atom that has been removed from the surface of the
needle tip, field evaporation needs to be synchronized with a laser
pulse and the atom needs to be evaporated as a charged ion. This is
not always the case. Researchers suspect some atoms evaporate as
neutral molecules, which can’t be detected, or evaporate between
distinct voltage or laser pulses, preventing an accurate measure-
ment of TOF. For example, nitrogen (N) can evaporate as N2;
while gallium can evaporate at energies below those provided by a
voltage or laser pulse.
In addition, unlike with other analytical techniques, it can
be difficult to know accurately the magnification for each sample.
Although APT can determine chemical composition and 3D
structure very effectively, it can struggle to determine the size of
a specific feature accurately or precisely. One possible solution is
to synergistically combine APT with other analytical techniques,
such as TEM. In other cases, it may be possible to calculate the size
of a particular feature, such as a precipitate or a cluster, by knowing
the number of atoms collected and the shape of the feature.
In practical terms, one of the most common problems with

APT analyses is sample fracture. This can occur because the sample
is under high mechanical stress during analysis, and any weak
point, such as a crack or a surface notch, may induce fracture
before the analysis is completed. Typically, metals and alloys are
strong enough to withstand voltage pulsing, but laser pulsing,
which field-evaporates ions at lower stress levels, is often used
with weaker materials.
In an ideal world, all APT experiments would extract data
with maximum quality and minimum analysis time. However,
each sample is unique and so operators need to define the optimum
conditions for each analysis (Table 1). This is particularly impor-
tant when planning experiments with types of samples not
previously analyzed.
For certain samples, analysis speed has to be sacrificed. Slower
pulse rates may be required for samples containing high-mass
elements with slow TOFs or elements with low thermal diffusivity,
as a fast pulse rate can produce spectral peaks with long tails.
Similarly, high detection rates maximize data quality for most
samples, but not when the analyzed material is prone to field-
evaporating multiple ions at the same time, or if the sample is
weak and prone to early failure.
In most cases, lower temperatures provide better data quality,
by reducing the impact of residual gas present in the analysis
chamber and harmonizing the fields required to evaporate
different elements, which in turn reduces spectral background.
Some samples, however, provide better yields at higher temperatures;
for example, aluminium alloys perform better at temperatures of
around 20–40 K, whereas steels are typically analyzed at 50 K.
Acquisition Variable General Metrics Comments Primary

parameter trend trade-off
Pulse rate f (Hz) ++ Background Faster is better Almost none

– m/n range Exceptions: except for
– Wrap around high mass ions practical
– Base temperature with long TOF considerations
offset effects (laser) or extremely for TOF of all
poor thermal potential ion
diffusivity species
materials in
laser mode
Base T (K) ++ Analysis yield Lower leads to Data quality
temperature –– MRP better spatial versus
–– Background resolution analysis yield
– Heat flow (laser) Higher leads to
– Surface diffusion better yield
Detection DR ++ Background Faster is better Data quality
rate (ions/ – Analysis yield Exception: versus
pulse) – MRP poor yielding analysis yield
– Multi-hit materials and
performance materials prone
to multi-hit
behavior
Pulse PF (%) ++ Compositional Higher is better Data quality
fraction accuracy Exception: versus
(voltage + Background extreme PF may analysis yield
mode) – Potential for PF limit MRP and
decay pulse amplitude
– High cycle fatigue is hardware
limited
Laser pulse LPE +++ Analysis yield Higher leads to Data quality
energy (nJ or pJ) ++ Background better yield versus
(laser mode) + MRP Lower leads to analysis yield
+ Effective pulse better spatial
fraction resolution
– Complex ion
generation
–– Surface migration
Table 1. General experimental parameter considerations, data quality trends and trade-offs
(parameter increases , trends improves +, or worsens – the metrics)
Peak overlaps between elements at different charge states

are possible. In most cases, overlaps can be distinguished based on
the other isotopes present. A classic example is steel. If a particular
steel alloy contains chromium (Cr), this element shares a common

isotope with iron (Fe): 54Cr and 54Fe. As both Cr and Fe also have
three other isotopes, it’s possible to analyze the ratio of each
isotope in the steel alloy and compare them to their natural
abundances, using this isotopic information to identify the
fraction of the peak that corresponds to 54Cr and to 54Fe.
With APT now being used in fields beyond metallurgical
analysis, data interpretation is becoming increasingly complex –
particularly in samples containing complex organic molecules,
where elements may field evaporate in conjunction with hydro-
gen, carbon or oxygen. This can produce a complex mass spectrum
that is difficult to interpret. Nevertheless, even though it may not
be possible to identify single atoms, high-quality 3D reconstruc-
tion can still produce valuable information at the molecular level.
One of the most intriguing aspects of APT is its ability to
create a 3D map of a sample at nanoscale resolution. The calcula-
tions used to produce these maps are based on a number of
assumptions, including the hemispherical surface of the tip. In
homogeneous samples, this is a good assumption, as it’s natural
for the sample to become approximately hemispherical to mini-
mize its surface area as atoms are evaporated.
Problems may arise in complex heterogeneous samples,
such as those containing interfaces and precipitates that result in
heterogeneous evaporation. When the sample contains materials
with different compositions, they can require different strength
fields to evaporate, which means the assumption of a hemispherical
sample is no longer valid.
This is the case for precipitates with a higher melting
temperature than the surrounding matrix. When they reach the
surface of a sample, the tightly bonded ions in these precipitates

will be the last to evaporate, creating a bump on the surface with a
higher magnification than other regions. In this case, assuming
a hemispherical surface will result in scaling errors during the
reconstruction process, although an accurate reading of the
magnification can often be obtained by complementing APT
with a correlative analysis like TEM.
WHAT’S NEXT?
The company CAMECA, headquartered in France with
facilities in the UK and USA, and a business unit of the global
instrument company AMETEK, is now the sole producer of
commercial APT systems, and is still actively working to improve
the technique by increasing throughput, efficiency, acquisition
times and resolution.
One of the last areas to resist automation is sample prepara-
tion. This is because, with APT being applied to an ever greater
variety of materials, more demands are being placed on the
methods for producing a needle shape. At the moment, all these
methods require multiple instruments to achieve this shape, with
heavy involvement from the operator. The future for sample
preparation includes developing methods that can produce such
microtips rapidly, accurately and automatically. This would allow
for the preparation of identical and uniform samples in large
volumes, generating comparable results for each analysis.
Another area where researchers would like to see improve-
ments is in the efficiency of ion detection. At the moment, ion
detection efficiencies of 50–80% are the best that can be expected
with modern instruments. The problem starts at the microchannel
plate before amplification. As this plate is composed of thousands
of channels arranged close together, there is always a chance that an
ion will hit the space in between two channels and get lost. New
designs for microchannel plates are envisioned, using thinner and
stronger materials covered by a secondary electron-emitting layer
made from a metal oxide such as aluminium(III) oxide, silicon
dioxide, zirconium dioxide or magnesium oxide. Researchers are
also testing new superconducting materials for this application.
Further improvements to the 3D reconstruction algorithms

would also be desirable. As described earlier, the most common
reconstruction method assumes a hemispherical end, but this
cannot be applied to all samples. One way to overcome distortions
would be to determine the exact trajectory for each ion, but there
is a large computational cost associated with such analyses at the
moment. Other options include developing new algorithms that
can more accurately describe a particular sample, and combining
APT with TEM to obtain information on the dimensions and
angles of the sample, as well as information about any interfaces
and precipitates that are present.
At the moment, combining APT and TEM on the same
sample volume can be time consuming and complicated, requiring
samples to be transferred from one instrument to the other under
UHV and/or cryogenic conditions. As there are many advantages
to this combination, including more accurate magnification and
reconstruction, researchers are considering ways to physically com-
bine both instruments or to more easily integrate the information
provided by each of them. One idea is to place a TEM on the side
of the APT to obtain information in an easy and convenient way,
as well as to allow the development of a more automated system.
After its expansion into non-conducting and geological
materials, some scientists are now trying to apply APT to biologi-
cal samples, but that raises the problem of how to produce a
needle shape in soft biological material. One solution is to freeze
the sample and prepare a needle tip with FIB, but this process
still has many difficulties. Efforts are underway to make it
possible to transport frozen samples between instruments under
UHV conditions.
Finally, APT also has the potential to become a metrology

tool. To do so, analyses would need to be completed within a certain
time, with the results used as feedback during further processing.
This approach would require a considerable amount of effort, but
no more than other metrology tools already available.
Conclusion
Since the turn of the century, the user base of APT has
expanded from just a handful of researchers to thousands who
now use APT on a regular basis, and is predicted to grow rapidly
in the coming years. The rapid increase in the number of publica-
tions featuring atom probe work is a testament to its increasing
popularity.
This is not surprising since, despite still struggling with
certain limitations, the ability of APT to determine the elemental
composition of samples in three dimensions at the nanoscale level
is too appealing to ignore. And this appeal should only grow, as
APT systems become more automated, and their resolution and
sensitivity gradually increase, until scientists really can determine
the identity and position of every atom in any material they desire.
FURTHER READING
Atom Probe Tomography user’s website, run by CAMECA.
(http://www.atomprobe.com)
Gault B, Moody MP, Cairney JM, Ringer SP. Atom probe
microscopy. New York: Springer, 2012. (http://dx.doi.org/10.1007/
978-1-4614-3436-8)
Hono K. Nanoscale microstructural analysis of metallic
materials by atom probe field ion microscopy. Prog Mater Sci 2002;
47:621–729. (http://dx.doi.org/10.1016/S0079-6425(01)00007-X)
Kelly TF, Larson DJ. Atom probe tomography 2012. Ann
Rev Mater Res 2012;42:1–31. (http://dx.doi.org/10.1146/annurev-
matsci-070511-155007)
Kelly TF, Miller MK. Atom probe tomography. Rev Sci
Instrum 2007;78:031101. (http://dx.doi.org/10.1063/1.2709758)
Larson DJ, Prosa TJ, Ulfig RM, et al. Local electrode atom probe
tomography: a user’s guide. New York: Springer, 2013. (http://dx.
doi.org/10.1007/978-1-4614-8721-0)
Lefebvre-Ulrikson W, Vurpillot F, Sauvage X (eds). Atom probe
tomography: put theory into practice. London: Academic Press, 2016.
(http://dx.doi.org/10.1016/B978-0-12-804647-0.01001-9)
Miller MK, Forbes RG. Atom-probe tomography: the local
electrode atom probe. New York: Springer, 2014. (http://dx.doi.
org/10.1007/978-1-4899-7430-3)
Miller MK, Kelly TF, Rajan K, et al. The future of atom probe
tomography. Materials Today 2012;15:158–65. (http://dx.doi.org/
10.1016/S1369-7021(12)70069-X)
Essential
Knowledge
Briefings

Atom Probe Tomography

Uploaded by

Copyright:

Available Formats

Atom Probe Tomography

Uploaded by

Document Information

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

Atom Probe Tomography

Uploaded by

Copyright:

Available Formats

Atom Probe

Front cover image: three-dimensional image of a NiFeAlCr-based

About Essential Knowledge Briefings

When APT was originally developed, its applications were

HISTORY AND BACKGROUND

was initially restricted to analyzing good electrical conductors,

Figure 1. Schematic of a modern atom probe

Such modern systems also benefit from further advances in the

three serpentine wires arranged in different directions, allowing

Nevertheless, the latest APT software can generally deal with

All of these advances have led to APT becoming a viable

Electropolishing may be relatively easy and cheap to per-

With FIB milling, APT samples are typically prepared using

only able to collect thousands to tens-of-thousands of ions for

Optimum conditions tend to differ for different materials, and

end, on which the position of the detected ions is back-projected.

CASE STUDY 1. Combining APT and EDX

‘People were originally skeptical about APT because of the time

1. Bonef B, Lopez-Haro M, Amichi L, et al. Composition analysis of

CASE STUDY 2. Applying APT to geological

One practical limitation lies in sample preparation. ‘It’s very

1. Pérez-Huerta A, Laiginhas F, Reinhard DA, et al. Atom probe tomography

CASE STUDY 3. Admiring diamonds with APT

(a) Individual nanometer-sized meteoritic nanodiamonds (carbon atoms: black

method to test diamonds in other applications. For example,

accurately determine the concentration and distribution of each

1. Heck PR, Stadermann FJ, Isheim D, et al. Atom-probe analyses of

PROBLEMS AND SOLUTIONS

In practical terms, one of the most common problems with

Acquisition Variable General Metrics Comments Primary

Pulse rate f (Hz) ++ Background Faster is better Almost none

Peak overlaps between elements at different charge states

steel alloy contains chromium (Cr), this element shares a common

surface of a sample, the tightly bonded ions in these precipitates

Further improvements to the 3D reconstruction algorithms

Finally, APT also has the potential to become a metrology

You might also like