Practical Guide For Curve Fitting in XPS
Practical Guide For Curve Fitting in XPS
Practical Guide For Curve Fitting in XPS
photoelectron spectroscopy
Cite as: J. Vac. Sci. Technol. A 38, 061203 (2020); https://doi.org/10.1116/6.0000377
Submitted: 05 June 2020 • Accepted: 18 September 2020 • Published Online: 06 October 2020
COLLECTIONS
Paper published as part of the special topic on Special Topic Collection: Reproducibility Challenges and Solutions
Practical guide for x-ray photoelectron spectroscopy: Applications to the study of catalysts
Journal of Vacuum Science & Technology A 38, 033204 (2020); https://
doi.org/10.1116/1.5140747
© 2020 Author(s).
TUTORIAL avs.scitation.org/journal/jva
George H. Major,1 Neal Fairley,2 Peter M. A. Sherwood,3 Matthew R. Linford,1 Jeff Terry,4 Vincent Fernandez,5
and Kateryna Artyushkova6,a)
AFFILIATIONS
1
Department of Chemistry and Biochemistry, Brigham Young University, C100 BNSN, Provo, Utah 84602
2
Casa Software Ltd., Bay House, Teignmouth, United Kingdom
3
Department of Chemistry, University of Washington, Seattle, Washington 98195
4
Illinois Institute of Technology, 3101 S. Dearborn St., Chicago, Illinois 60616
5
CNRS, Institut des Matériaux Jean Rouxel, IMN, Université de Nantes, F-44000 Nantes, France
6
Physical Electronics, 18725 Lake Drive East, Chanhassen, Minnesota 55317
Note: This paper is part of the Special Topic Collection on Reproducibility Challenges and Solutions.
a)
Electronic mail: [email protected]
ABSTRACT
The use of peak fitting to extract information from x-ray photoelectron spectroscopy (XPS) data is of growing use and importance. Due to
increased instrument accessibility and reliability, the use of XPS instrumentation has significantly increased around the world. However, the
increased use has not been matched by the expertise of the new users, and the erroneous application of curve fitting has contributed to
ambiguity and confusion in parts of the literature. This guide discusses the physics and chemistry involved in generating XPS spectra,
describes good practices for peak fitting, and provides examples of appropriate use along with tools for avoiding mistakes.
I. INTRODUCTION Curve fitting, also known as peak fitting,4 is the process used
Over the past 30 years, x-ray photoelectron spectroscopy (XPS) to extract information from the spectral data for a number of tech-
has become the most widely used surface analysis tool and has been niques. Although the details of curve fitting depend on the techni-
an essential component of many research studies.1 Curve fitting has que in question, the curve fitted spectra generally contain
been widely used for more than 50 years for extracting chemical overlapping peaks. Each of these peaks is represented by a function
information from the overlapping features in high-resolution XPS that reflects the physical process involved in generating the original
spectra.2 Despite computational advances and higher accessibility of signal. XPS data interpretation and representation range from a
software resources, it has been challenging to develop a chemically rudimentary understanding/extraction of the elements present in a
and physically meaningful approach to curve fitting. The absence of material to advanced peak fitting and background analysis that
a distinct theoretical description of XPS fitting has led to the publica- reveal chemical states and sample morphologies.
tion of erroneous conclusions about surface chemistry.2,3 In an In XPS, it is convenient to identify two spectral regions,
ongoing study of XPS data in three high profile journals,3 it was namely, the core region (electrons with binding energies, BEs,
observed that roughly 70% of the papers using XPS analyzed the greater than 30 eV) and the valence band region (BE < 30 eV). In
data using some type of curve fitting. Furthermore, errors, miscon- the core region, the spectral features arise from photoelectrons gen-
ceptions, and bad curve-fitting practices accounted for most of the erated from core energy levels (atomic orbitals), which are charac-
serious problems in both the measured XPS data and the spectral teristic of the individual atoms in the sample.5 In contrast, the
analysis that were identified in more than 30% of the papers ana- features in the valence band region arise from photoelectrons
lyzed. This guide is intended to help address this important problem. generated from energy levels that typically involve the chemical
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-1
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
interactions between the atoms in chemical bonding (the molecular information that are critical for avoiding common mistakes in
orbitals).5 The spectral features of both the core and valence regions curve fitting. Important useful topics addressed include dealing
are sensitive to the chemical environments of the atoms in a sample. with overlapping peaks, fitting transition metal spectra, appropri-
In the core region, chemical shifts in the BEs yield a series of typi- ately constraining spin–orbit split peak ratios and energy separa-
cally overlapping peaks when the sample has atoms in different tion, and others.
chemical environments, e.g., oxidation states.6 The resulting data can
be analyzed by curve fitting. Curve/peak fitting is often the only way
to extract quantitative information from these spectra. II. BASIC APPROACH FOR FITTING XPS SPECTRA
The objective of curve-fitting high-resolution core XPS spectra A peak fitting model is defined in terms of component peaks
with a set of component peaks is to separate the photoemission and a background algorithm. The component peaks are specified
signal originating from distinct elemental or chemical states. using lineshapes, which are mathematical functions, and fitting
Parameters of importance extracted from curve fitting are the photo- parameters that permit a component peak to vary in a variety of
emission peak BE, full width at half maximum (FWHM), lineshapes, ways, including position, FWHM, area, Lorentzian character,
and area. The position of a component peak provides evidence for degree of asymmetry, and Gaussian character (The first four of
assigning it to an elemental or chemical environment. The FWHM these parameters will apply to all peaks, the latter two in some
and peak lineshapes are indicators of the chemical and physical envi- specific cases.) Sets of component peaks are summed (sum of all
ronment of the atom. Photoemission peak lineshapes can vary from component peaks) and then added to a background to form an
simple, narrow, symmetric shapes to complex structures characteris- approximation to the original data envelope.
tic of the different oxidation states of metals.5 Quantitative informa- The procedure of fitting component peaks to reproduce exper-
tion about the concentrations of the chemical states identified is imental photoemission spectra utilizes mathematical algorithms to
inferred by measuring the relative area of each component peak, the minimize a figure-of-merit as a measure of the closeness of the
calculation of which requires the separation of zero energy-loss mathematic model to the experimental data. The fitting algorithm
signals from the inelastically scattered background. The determina- incrementally alters the adjustable parameters, which leads from a
tion of accurate approximations to the background signal is also an current state to a termination state that represents a local minimum
integral part of the XPS peak fitting process. with respect to the figure-of-merit. The ability of the functions to
Curve fitting is sometimes misdescribed as being deconvolu- produce a reasonable fit from an initial guess in the iterative curve-
tion.4 Deconvolution is a different process to curve fitting. fitting process can vary. It is crucial to choose an initial guess that
Deconvolution is a process that attempts to remove the broadening is appropriate based on the physics and chemistry of the instru-
that results from instrumental factors, separating the “true” spec- ment and the sample. A poor set of initial guesses can cause the
trum from the observed spectrum. The process results in a new fitting algorithm to direct the fit into a local minimum that is not
spectrum with the data altered by the deconvolution process. representative of the sample physics and chemistry. That is, the
Deconvolution will not be discussed in this paper. curve-fitting process can give a mathematically good or even excel-
Our goal is to demonstrate useful, practical approaches for lent fit even though the component peaks in the final fit are not
analyzing XPS core regions (high-resolution or narrow scans) and chemically or physically reasonable.
what common mistakes should be avoided when curve fitting. A figure-of-merit, e.g., χ 2 or the standard deviation of the
Although there are several types of common errors in peak fitting, residuals, is a single number that is often used to guide adjustments
effective and useful approaches to peak fitting vary somewhat with to the numerous fitting parameters in the peak fitting process. The
the analysis objective, the type of sample, and the complexity of the fitting process may involve numerous paths though the parameter
spectrum. Fitting may be used to verify the composition of a pure space with the possibility of hitting and getting stuck in a number
sample (Sec. V), to separate overlapping peaks for quantitative analy- of local minima. It is, therefore, possible to obtain a range of out-
sis (Sec. VI B 1), to obtain the relative amounts of each phase from a comes depending on the starting point of the fit. There are a
sample with multiple phases of the same element (Sec. VI B 2), and number of additional checks that can be used to ensure that the
to examine in great detail the chemical spectrum for an element or physicochemical quality of a peak fit is adequate and that the current
phase for comparison with theory. This guide describes multiple fit is not just the best mathematical solution. These include the mon-
complications that can appear with fitting the core region, including itoring of the quality of the data reproduction with respect to the
multiplet splitting, conduction band interactions, and the use of noise; the applicability of the curve-fitting model to data of similar
achromatic X-radiation. origin; the generation of accurate peak areas for spin–orbit compo-
The following sections of the paper highlight important topics nents; and the accuracy of peak ratios for peaks from different ele-
for consideration during peak fitting. Section II describes a funda- ments or oxidation states of an element in a known chemical moiety.
mental approach to peak fitting, while Sec. III describes typical Optimization algorithms mathematically fit peaks to data. They
peak fitting errors observed in the literature and offers recommen- lack significant input from physics and chemistry, so optimization
dations to avoid them. Section IV addresses issues to think about based on a single value figure-of-merit may not yield comprehensive,
before starting to curve fit spectra (instrument setup, adequate data scientifically meaningful results. Physical and chemical information
collection, charge correction, lineshapes, and phenomena impacting can be added to a peak model through the selection of the line-
lineshapes and widths). Section V shows the fitting of poly(ethylene shapes, the number of component peaks within a peak model, and
terephthalate) (PET) as an example of fitting a reference spectrum the parameter constraints, all of which present a means by which
from pure material. Section VI summarizes a variety of tools and known relationships are imposed on otherwise independent
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-2
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-3
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
TABLE I. (Continued.)
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-4
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
TABLE I. (Continued.)
component peaks. That is, the likelihood of achieving a chemically peak fitting that include appropriate considerations of the physical
reasonable fit is increased if the initial guesses are based on solid sci- and chemical nature of the relevant spectra. Some of the errors
entific principles. listed below are superficial and do not impact the conclusions of a
The art of fitting data with peaks is in selecting the appropri- paper. However, some fundamentally alter the interpretation of the
ate number of component peaks, making use of appropriate line- data and the conclusions drawn from it. Table I lists the errors in
shapes, and limiting the set of fitting parameters through XPS peak fitting that we have observed in the scientific literature
parameter relationships to produce peak models capable of measur- and suggests ways to avoid them. A more detailed discussion of the
ing physically meaningful quantities from XPS spectral forms. The errors can be found in a report.7
challenging part arises in selecting the correct physical solution
from the set of potential mathematical solutions. In practice, this is IV. XPS MEASUREMENT
best done by obtaining verification of the fitting model by repeating
these measurements on equivalent samples to confirm that the The chemical information that can be determined from curve-
observed changes are representative of the samples. Data collected fitting XPS spectra can only be physically meaningful if the data
should include measuring data from standard samples to verify the are appropriately collected, and proper consideration is made of all
performance of an instrument and understand the limitations of the physical and chemical contributions to the photoelectron
XPS when analyzing similar materials of known composition. spectra.
III. COMMON MISTAKES IN XPS PEAK FITTING A. Instrument calibration and performance
Following the principles outlined in this guide should help the Energy calibration of the instrument is a prerequisite for accu-
analyst avoid many of the errors listed below. Nevertheless, these rate peak energy identification. Any instability within an instru-
errors are regularly seen in the scientific literature.3 Of course, ment has the potential to alter lineshapes and accurate kinetic
there are many examples in papers of proper XPS analysis and energy registration. Regular instrumental calibration and check-up
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-5
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
are critical to ensure that the instrument is performing within the common error.3 Another possible error is to acquire data that have
acceptable limits of reproducibility and accuracy9 needed for physi- a poor signal-to-noise (S/N) ratio and to smooth it before curve
cally meaningful interpretation of curve fit results of spectra. fitting. (Note that the S/N ratio varies across an XPS survey spec-
Figure 1 shows two Nb 3d photoemission spectra collected trum.3) Smoothing may introduce changes in the spectral shape
from a sample that was mainly metallic with a small amount of and width;14,15 changing the underlying physics, and may lead to
NbO and NbO2 present on the surface. The bottom spectrum dis- erroneous interpretation of curve fitted spectra.
played an incorrect (compressed) spin–orbit splitting for the metal-
lic component and a second component that did not correspond to C. Accurate charge correction
one of the expected oxide components. These problems were iden-
tified by comparison with Nb spectra published in the literature Accurate BE identification relies not only on the instrumental
(previously collected data and resources described in Sec. VI A) energy scale calibration but also on the accuracy of charge correc-
and were due to a malfunctioning pass energy (PE) control board. tion if charge compensation is employed during data acquisition.
The correct spectrum in Fig. 1 (top) has all three expected compo- Charge compensation is often required when measuring XPS
nents (Nb, NbO, and NbO2). spectra from insulating materials. A charge builds up near the
Even though modern XPS instruments are becoming more surface of these materials as the photoelectrons are emitted. This
stable and reliable, it is crucial to perform regular energy scale and usually shifts the photoelectron energy and distorts peak shape. For
linearity calibrations as prescribed by the standards.11,12 Recently, a details on charge compensation, please refer to the XPS Guide:
procedure that allows the performance and calibration of an XPS Charge compensation and BE referencing for insulating samples.16
instrument to be checked rapidly and frequently was published.13
D. Lineshape of an XPS peak
One of the most influential aspects of a successful peak model
B. Adequate data acquisition is the lineshapes or the shapes/functions appropriate for modeling
The quality of information extracted from curve-fitting XPS the photoemission process after measurement by an XPS instru-
spectra can only be as high as the quality of the experimental data. ment. These lineshapes must accurately represent, or at least closely
The spectra acquired are not always of sufficient intensity and reso- approximate, the signal measured by an instrument. Therefore, it is
lution to perform useful peak fitting. Instrumental contributions to essential to understand the role of the instrumentation in modify-
the spectrally resolvable chemistries are discussed below. The ing peaks.
analyst should also consider the signal-to-noise of the spectra pro- The XPS process is relatively simple.17 X-ray light in the
duced. For elements with low sensitivities and/or that are present intrinsic process causes an electron to move from an initial, occu-
in low concentrations, overinterpretation of noise as chemistry is a pied energy level to a final, virtual energy level that has energy
FIG. 1. Nb 3d spectrum collected using a PHI 5600XPS system (Ref. 10). The bottom spectrum was collected with a malfunctioning PE board. The top spectrum was col-
lected after the PE board had been replaced. A Shirley background was used.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-6
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
greater than the Fermi level of the solid under study. The electron as additional intrinsic, processes can occur after the intrinsic XPS
is then lost from the virtual final state energy level and escapes process that can give additional spectral features.
from the solid at a specific energy. This energy is subsequently In choosing a Gaussian–Lorentzian mix, it is important to
measured in the spectrometer. The uncertainty principle gives the note the following points:
initial energy level a Lorentzian energy distribution (discussed in
Sec. IV E). The basic shape of the XPS peak recorded in a spec- • Typically, it is found that compounds have a more Gaussian
peak shape than metals, which usually have considerable
trometer is modified by instrumental and other factors (such as
Lorentzian character. One reason for this is the presence of
phonon broadening) to give a Gaussian contribution.18 This leads
to a measured lineshape that is a convolution of a Gaussian and a optical phonons in compounds, whereas metal has principally
Lorentzian function. This function, known as the Voigt function,19 low-energy acoustic phonons leading to an increase in vibrational
most accurately reflects the nature of this intrinsic XPS process. broadening in compounds and thus more Gaussian character.
This topic is discussed in Sec. IV E.
XPS peaks can be expressed by Voigt functions that vary in
• In the curve-fitting process, the analyst starts the process by sug-
their amounts of Gaussian and Lorentzian character. This can be
done by identifying a Gaussian (ghm) and a Lorentzian (lhm) gesting possible values for the various peak parameters such as
FWHM. BE, FWHM, Gaussian/Lorentzian character, and peak heights. It
A parameter called the “Gaussian/Lorentzian mix,” glmix, may is usually necessary to fix some of these parameters in order for
the iterative process to converge. The analyst will determine
be defined as
typical values for Gaussian–Lorentzian mix for their instrument,
but typical values for many instruments are given below:
lhm o Compounds: glmix below 10% for Voigt functions and M = 0.5
glmix ¼ ,
lhm þ ghm for pseudo-Voigt product functions and M = 0.2 for
pseudo-Voigt sum functions.
o Metals and elements: glmix above 60% for Voigt functions and
which takes the value 0.0 for a pure Gaussian and 1.0 for a pure
M above 0.8 for pseudo-Voigt functions and M above 0.6 for
Lorentzian. Thus, the Voigt function varies the amount of
pseudo-Voigt sum functions.
Gaussian and Lorentzian character by using Gaussian and
Lorentzian FWHM values that are usually different. Figure 2 shows These represent good starting guesses for the Gaussian–Lorentzian
a plot of the Voigt function with glmix = 0–0.9999. It should be mix values.
noted that Gaussian–Lorentzian sum and product functions, which
approximate the Voigt function, called pseudo-Voigt, have also E. FWHM of the XPS peak
been widely used in XPS peak fitting.20 In these pseudo-Voigt func-
tions, there is a mixing ratio (M), which controls the amount of The FWHMs of the Voigt function peaks used in core XPS
Gaussian and Lorentzian character, typically M = 1.0 for a pure peak analysis depend on several factors discussed below.
Lorentzian and M = 0.0 for a pure Gaussian, but sometimes these
M values are reversed. As discussed in Sec. IV F, extrinsic, as well 1. Lifetime of the core hole—the natural linewidth
The photoelectron peak is broadened to give a Lorentzian
shape that depends on the lifetime of the core hole resulting from
the photoelectron process. The core hole corresponds to an excited
state, and various processes fill this hole, especially x-ray or Auger
emission. There are several tabulations (Refs. 22 and 23 and refer-
ences therein) of natural linewidths, a number of which are based
upon experimental data from x-ray emission spectroscopy and
Auger spectroscopy and calculations. Krause and Oliver pub-
lished23 a tabulation of core energy levels from atomic number 10
to 110, and many other papers report natural linewidth values,
including those using XPS data.22
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-7
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
core levels can be excited by higher energy x rays compared to soft consistent with a balance between the appropriate signal to noise
Al Kα x rays. For example, for Al, a soft x ray can excite up to the ratio and analysis time available, used to get the best data for curve
2s orbital only, but a hard x ray can excite up to the 1s orbital. fitting.
Figure 3 shows the Al 2p and Al 1s core regions obtained
using a Cr x-ray source, which has an energy of 5414.7 eV. The
kinetic energy of Al 2p photoelectron is ∼5336.8 eV, which is much 4. Resolution of the spectrometer used to measure the
higher than the kinetic energy of Al 1s electron (∼3850.3 eV). The photoelectron energy
spectra in Fig. 3 are overlayed to the same maximum and intensity Most XPS instruments use hemispherical analyzers. Usually,
for comparison. In Fig. 3(a), the Al 2p curve fitted with two asym- the photoelectrons have their energy adjusted in a lens system so
metric spin–orbit components separated by 0.44 eV and at 2:1 that they enter the hemispherical analyzer with particular kinetic
intensity ratio. The causes and treatment of asymmetry are dis- energy called the PE. The analyzer settings, including PE, contrib-
cussed in Sec. VI F. In Fig. 3(b), Al 1s and Al 2p3/2 components of ute to the resolution function of the instrument and hence to the
Al 2p transition are overlayed to the same maximum and intensity energy resolution and photoemission peak shape. Therefore, it is
for comparison. The FWHM of Al 1s spectrum is 1.35 eV while important to determine the energy resolution for the operating
that of Al 2p3/2 component is 1.08 eV, demonstrating the increase mode used to acquire spectra that will be curve fitted.
of lifetime broadening for the deeper core level.
5. Chemical environment of the atom from which the
3. Effect of the spectrometer used to measure the pho- core photoelectron was ejected
toelectron energy on the linewidth The chemical environment of the atom from which the core
The spectrometer that measures the kinetic energy of the pho- photoelectron was ejected has a substantial effect on the FWHM.4
toelectrons adds additional width to the natural linewidth, as well In particular, the intrinsic photoelectron process couples with
as giving Gaussian character to the peak lineshape. This means that phonons in the solid, resulting in broadening of the core peak and
the observed linewidth is generally substantially different from the giving the peak Gaussian character. Metal phonons are often domi-
natural linewidth necessitating the use of the Voigt function nated by acoustic phonons of low energy, whereas metal com-
described above. Tables of natural linewidth (e.g., Ref. 22) can be pounds such as oxides and carbonates have extensive optical
helpful in setting the Lorentzian component of the Voigt function phonons of much higher energy resulting from multiple solid-state
in a curve fit of a spectrum but rarely describe the measured peaks vibrational modes.25 Thus, metals normally have substantially nar-
themselves. Typically, the analyst should obtain the spectrum with rower FWHM values than metal compounds and more Lorentzian
the best resolution available from the specific spectrometer, character. Figure 17 in Sec. VI F for the Al 2p region illustrates this
FIG. 3. Core XPS spectra of aluminum metal obtained using Cr Kα photon energy of 5414.7 eV. Spectra obtained using PHI Quantes instrument. (a) Al 1s and Al 2p
spectra overlayed to the same maximum and intensity for comparison. The Al 2p curve fitted with two spin–orbit components separated by 0.44 eV and at 2:1 ratio; (b) Al
1s and Al 2p3/2 component of Al 2p transitions overlayed to the same maximum and intensity for comparison. A Shirley background was used.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-8
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
—it can be seen that the oxide peak has a substantially greater approaches for background inclusion, and a forthcoming paper8 is
FWHM than the metal peak when measured at high resolution. devoted to this topic. There are three major approaches:
• The removal of the background to yield a spectrum with an
F. Existence of extrinsic processes as well as additional approximately horizontal background. This approach alters the
intrinsic processes experimental data with the subsequent curve fitting being made
The photoelectron process can be simplified into a three-step on the altered data. This can lead to errors depending upon the
model:26 chosen background model. It also deprives the viewer of the
ability to compare the original experimental data with the curve
• Step 1: The excitation of the electron from the initial energy level fit. This approach is not recommended.
to the final virtual orbital state. Processes associated with this • The addition of a fixed background to background subtracted spec-
step are called intrinsic processes. trum before peak fitting, where this background is not subtracted
• Step 2: The transfer of the photoelectron to the surface. Processes from the spectrum, and it is shown in the final fit. This approach is
that occur after the first step are called extrinsic processes. mathematically identical to the previous one, but it allows the
• Step 3: The escape of the photoelectron from the surface into the viewer to see the original data, the background, and the fit together.
vacuum. • In the active approach, the background contains parameters that
The first two steps are sometimes called the two-step model. are included as potentially variable factors in the fit. This method
In addition to the key intrinsic process, there are intrinsic pro- also preserves the data in its original form but allows for different
cesses that involve the ejection of a photoelectron, as follows. backgrounds in different regions of the spectrum. In some insula-
tors and gases,30 e.g., in near ambient pressure XPS,31–34 the XPS
• Plasmons25 can give rise to additional peaks in a spectrum. Both background can be linear, but most XPS backgrounds in metals
bulk and surface plasmons are possible. For example, aluminum and semiconductors are nonlinear. The active approach allows for
metal has a rich plasmon structure.27 Plasmons can also be addressing these differences in complex materials.
extrinsic.
• Shake-up5 processes in which a valence electron and a core elec- We do caution the reader that a simple linear background between
tron are excited. If the valence electron is promoted to another the starting and ending points in a spectrum is incorrect and will
valence energy level in the sample, this is called a shake-up cause substantial errors in the peak fitting of spectra where the
process. If the valence electron is ejected as an additional photo- measured background is not flat. As discussed in the Practical
electron, the process is called a shake-off process. Guides for X-ray Photoelectron Spectroscopy: Quantitative XPS,35
• Multiplet splitting5 arises from samples that have unpaired elec- the line describing the background should, ideally, go through the
trons because there is more than one final state depending upon middle of the noise in the data before and after the peak.
the number of spin up and spin down electrons in these final
states. Transition metal compounds are often paramagnetic and, V. EXAMPLE CURVE FIT OF REFERENCE SPECTRA
in this case, display multiplet splitting in their core XPS data FROM PURE MATERIALS—PET
(discussed in Sec. VI C).
We fit here a reference spectrum taken at high resolution from
• High Binding Energy Assymetry28 Conduction Band interactions
an XPS database of organic polymers—PET.36 PET is an important
give a tail of approximately exponential shape to the core XPS of
reference polymer that is used for verifying instrument performance,
conducting materials (Sec. VI F).
including charge neutralization validation.4,13,37,38 Monochromated
Extrinsic processes include Al Kα x rays (1486.3 eV) were used to collect these data on a Scienta
ESCA300 spectrometer. The energy spread of the monochromator is
• Plasmons, which can also be intrinsic. predicted to be 0.26 eV. A PE of 150 eV was used, and high energy
• Inelastic processes often occur between ejected photoelectrons resolution was achieved by using an appropriate analyzer entrance
and other atoms and molecules. These processes cause loss of aperture width. The FWHM of the C 1s envelope was determined to
energy of the photoelectron and represent the principal contribu- be 0.67 eV. Low-energy electrons were used for charge compensation.
tion to the background in XPS.29 There are effective software The PET structure is shown in Fig. 4. Based on its chemical
models for the study of this loss structure, and these studies can formula (C10H8O4)n, there is 71 at. % of C and 29 at. % of O in
provide valuable information. These processes can have a large PET. There are 10 carbons in the repeating unit of the polymer, 4
effect on peak ratios as a function of emission angle. of which are present in the aromatic ring (labeled 1) with a BE of
284.7 eV. Two other carbons (labeled 2) in the aromatic ring are
secondary shifted carbons such as C*—C(vO)O, which will con-
G. Background subtraction tribute to the BE of ∼285.3 eV. Two carbons that are singly bonded
Most XPS spectra have a significant background, and account- to oxygen (labeled 3) have BE of 286.2 eV. The other two carbons
ing for the background contribution is an important aspect of the are in carboxyl (ester) groups (labeled 4), and their BE is 288.7 eV.
curve-fitting process. The choice of the type of background and Out of the four total oxygens, half are present in the OvC part of
selection of end points has the most significant impact on the area the carboxyl/ester group (labeled 1) at 531.6 eV, and the other half
under the curve, and hence the ratios of the individual peaks used to are in the O—C part of the carboxyl/ester group (labeled 2) at
curve fit the spectrum. There have been numerous studies devoted to 533.2 eV.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-9
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
FIG. 4. High-resolution C 1s and O 1s spectra of PET fitted using symmetrical GLS (fixed M = 0.1) peaks. (a) and (b) Initial set of peaks added to initiate the curve fit. (c)
and (d) Final curve fitted spectra. Residuals are displayed below the spectra. A Shirley background was used.
The C 1s and O 1s spectra are accompanied by a survey scan Once the necessary component peaks have been added into the
confirming that no other elements are present at the PET surface. envelope, constraints on the adjustable parameters need to be estab-
When a Shirley background is used for quantification of the high- lished. If the goal is to understand the chemical state information in
resolution C 1s and O 1s spectra, 69 at. % C and 31 at. % O are the spectra, adding parameter constraints will improve the stability
detected. Slightly different concentrations due to the possible pres- of the peak model with respect to the noise. However, with each new
ence of adventitious carbon38 are observed. constraint, the output from the optimization converges ever more
Figure 4 shows C 1s and O 1s high-resolution spectra (narrow closely to the solution defined by these constraints until the useful-
scans) with component peaks added to initiate a curve fit. The orig- ness of fitting peaks to the data is lost. If the goal is to measure the
inal data from the “High Resolution XPS of Organic Polymers: The intensity changes for a series of samples from the same study and a
Scienta ESCA300 Database”36 have been fit using the commercial peak model is well formed, then a rigid peak model that is heavily
software MultiPak 9.9.39 Four component peaks are added to the constrained will tend to return more consistent intensities for data
carbon spectrum, reflecting the number of different chemical envi- with similar chemistry and experimental conditions. Somewhere
ronments that contribute to the resolvable BE shifts discussed near the middle of these two extremes is the domain for most photo-
above. Due to the aromatic structure of the benzene ring, a emission peak fitting.
shake-up signal is observed above 290 eV, which should represent FWHM constraints and the indirect constraints imposed by
ca. 7% of the total carbon signal based on the reference data.36 This lineshape selection are the two main parameters used in peak
shake-up contribution is fit with one or more peaks. fitting. Other types of constraints, such as spin–orbit intensity ratio
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-10
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
for spin–orbit splitting, are discussed in Sec. VI C. For spectra with and 58% from O2 (533.2 eV) (with shake-up peak at 538.5 eV and
well-defined peak resolution, e.g., the C 1s narrow scan of PET, the contamination peak at 535.0 eV excluded from the total area).
FWHM can be chosen based on the widths of the individual peaks Based on the area under each component peak, the quantifica-
to reproduce the experimental curve as best as possible. Some var- tion of PET can be derived, as shown in Table II. This direct, quan-
iation (ca. 10% of the FWHM or ±0.1 eV) might be justified among tifiable comparison of the carbon and oxygen in the polymer can
different peaks due to small secondary effects. As described above, be used to confirm both the structure of the polymer and that the
a function having both Gaussian and Lorentzian character often curve fit adequately reflects the chemical structure.
best describes the experimentally derived spectra. The
pseudo-Voigt function that is used to approximate the Voigt func-
tion is typically either a product function or a sum function. Both
functions involve the mixing of equal width Gaussian and VI. AVOIDING MISTAKES DURING PEAK FITTING
Lorentzian functions with a mixing ratio (M) defined in the analyt-
ical function. The mixing ratio, M, takes the value 0.0 for a pure A. General information and databases of reference
Gaussian and 1.0 for a pure Lorentzian, though some authors have spectra with information on peak widths, separations,
the reverse definition. During the iterative fit, a fixed GL mixing and satellite positions
ratio can be chosen, or a range of GL mixing ratios can be used There are several resources available to assist in curve fitting,
during optimization to find the best fit. In this example, a sum including complex spectra that contain several chemical compo-
function of Gaussian and Lorentzian (GLS) with M = 0.1 (where nents. The NIST XPS database40 and XPS Spectra (Chemical Shift/
M = 0.0 is for 100% Gaussian) was used. Binding Energy)41 are searchable databases containing line posi-
Similarly, based on the sample chemistry, two peaks to be fit tions, chemical shifts, values for doublet splitting, and energy sepa-
are added to the O 1s spectrum to best approximate the number rations of photoelectron and Auger-electron lines. X-ray
and type of chemical moieties present. That is, as predicted by the Photoelectron Spectroscopy (XPS) Reference Pages42 and the XPS
chemistry of PET, the O 1s spectrum is fit with two distinct peaks Reference Table of Elements43 contain curve-fitting details, BEs,
of approximately equal height and similar FWHM. Due to the aro- FWHMs, spin–orbit splitting values, references, and other practical
matic structure of the benzene ring, some shake-up signature is notes for most of the elements in the periodic table. The XPS sim-
observed above 539 eV. One more peak is added at ca. 535 eV to plified website also contains useful information for narrow scan
account for surface contamination. analysis.44
During an iterative fit, constraints of FWHM and lineshapes There are various valuable databases that contain high-quality
are used to minimize the residuals—the difference between experi- spectra of known materials that can be downloaded. These include:
mental and fitted curves. It is good practice to show the residuals
of a fit. Figure 4 and Table II show the final parameters used to fit • Representative spectra for each element in a book published by
the spectra. The percent area obtained is slightly different from the Physical Electronics and used by analysts for many years.10
data published in the Handbook due to the different lineshapes • The journal Surface Science Spectra (SSS). This journal has an
used in the original Handbook.36 The resulting peak model consists invaluable collection of spectral data with full experimental
of the peak position and percent area that each component peak details. From volume 15 onwards, data can be downloaded as a
occupies out of the total carbon signal. The component peak posi- text file. SSS contains many examples of peak fitting.
tion information is then translated into the chemical state of the • eSpectra®45 is a database and comparative tool that contains the
sample. Curve fitting of the C 1s spectra results in 40.3% of the complete collection of the spectra published in Surface Science
area coming from C1 (284.8 eV), 20.2% from C2 (285.3 eV), 20.2% Spectra as well as from other sources.
from C3 (286.3 eV), and 19.3% from C4 (288.7 eV) (with shake-up • The XPSSurfA database hosted by La Trobe University in
peak at 291.3 eV excluded from the total area). Curve fitting of the Australia contains spectra from high-quality samples that were
O 1s spectra results in 42% of the area coming from O1 (531.7 eV) collected under precise conditions.46
TABLE II. Parameters used to curve fit C 1s and O 1s from PET in Fig. 4.
C 1s
BE, eV 284.8 285.3 286.3 288.7 291.3
FWHM, eV 0.9 1.0 1.2 0.9 2.5
% Area 39.4 19.7 19.7 18.9 2.3
% Area without a shake-up 40.3 20.2 20.2 19.3 0.0
O 1s
BE, eV 531.7 533.2 535.0 538.5
FWHM, eV 1.1 1.4 1.4 2.0
% Area 39.6 55.9 2.0 2.8
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-11
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
B. Use of reference spectra as to why peak assignments are difficult with transition metals. The
first is the combinations of intrinsic and extrinsic processes, such as
1. Resolving overlapping peaks
multiplet, shake-up, and plasmon loss structures, contributing to the
In this example, the N 1s spectrum of N-doped HfO2 overlaps overall photoelectron shape of the pure metals and their different
with the Hf 4p3/2 region. The Hf 4p3/2 peaks appear between 375 coordination states. The second is the overlap of the various chemi-
and 390 eV, while some higher BE components contribute to the cal states usually present in transition metal-based systems. The third
energy range where a signal from N 1s is expected. To solve this is that exposing transient metal compounds to air always results in
problem of overlapping spectra, it is important to acquire the oxidation and formation of both stochiometric and substochiometric
spectra from the reference material, a thick Hf oxide film without oxidation states. Another complication is that these oxides can
nitrogen doping in this case. Figure 5 shows a high-resolution Hf have varying degrees of conductivity, which may affect the position
4p3/2 region in which several peaks are included to fit the experi- and shape of photoelectron peaks due to charging. Early and
mental data. even current databases, such as NIST and Handbook of X-ray
The Hf 4p3/2 region [Fig. 5(a)] has features that occur in the Photoelectron Spectroscopy,10 usually indicate that there are multiple
same region where N 1s signal is present. In this example, the chemical states that can correspond to the same BE and, at the same
purpose of curve-fitting Hf 4p3/2 spectrum is not to extract the time, they provide single BE for a unique chemical environment of
chemistry of Hf but rather to create a mathematically rigid model the transition metal. However, as multiple reports of reference com-
in which peaks overlapping with the N 1s region are constrained in pounds have shown,47–49 complex spectral structure, which is the
position, width, and area to the higher intensity peak in the lower result of an envelope of multiplet peaks and/or other intrinsic and
BE region which is free of overlap with the N 1s region. Two peaks extrinsic effects, is often present with the different chemical states of
due to Hf 4p3/2 in the region where N 1s is located, i.e., between the element. Simply put, the multiple chemical states of a single tran-
390 and 405 eV, are constrained in intensity, position, and width sition element often give radically different spectra.
with respect to the most intense peak at 383 eV. For an N-doped Figures 6(a)–6(c) show spectral signatures for pure metallic
material with two different loadings, the rigid set of peaks is intro- Ni, NiO, and Ni(OH)2, as reported by Biesinger et al.47 The com-
duced into the high-resolution N 1s/Hf 4p3/2 region [Figs. 5(b) plexity of the spectral signatures due to contributions of the multi-
and 5(c)]. The intensity of peaks overlapping with N is controlled ple processes that can occur when the photoelectron escapes the 2p
by the height of the most intense peak, and this leads to a poor fit. orbital is obvious. Therefore, the most accurate approach to solve
This fit, though, is consistent with the Hf, so additional peaks must this problem is to use a curve-fitting parameter either developed by
be added to complete a curve fit. These peaks are the contribution the analysis from pure reference spectra, if they are available, or, to
from N 1s photoelectrons. This approach is successful even in sit- use curve-fitting parameters from reference compounds published
uations when N loading into the sample is very low, assuming that in the literature. The first approach is preferred as it ensures that
the data have been acquired with an appropriate signal to noise. the same spectral and analyzer conditions are used to collect the
data from the samples and reference materials. We emphasize that
the series of peaks used to fit the narrow scans in Figs. 6(a)–6(c)
2. Fitting transition metal spectra based on reference are simply employed to fit the peak envelope to create an overall
spectral envelopes shape that will be used in subsequent peak fitting, more than they
High-resolution spectra of transition metals are among the are intended to have specific chemical meaning.
most challenging spectra to peak fit.47–49 Indeed, it is common to Figure 6(d) overlays the photopeak shapes (or envelopes) that
see peak fitting of transition metals that oversimplify assignments were created by Biesinger et al. based on fitting the narrow scans in
of peaks to individual chemistries and subsequently have erroneous Figs. 6(a)–6(c) for Ni, NiO, and Ni(OH)2.47 The ambiguity of
interpretations of the results. There are several contributing factors finding a unique spectral signature for pure reference material is
FIG. 5. High-resolution region of Hf 4p3/2-N 1s electrons for (a) pure HfO2, (b) highly doped N-HfO2, and (c) low-doped N-HfO2. The peak model in (a) is a mathematically
rigid constrained model used in separating contribution from the Hf 4p3/2 region and N 1s electrons in (b) and (c). A Shirley background was used.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-12
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
obvious. However, through a linear combination of the photoelec- input from other spectral features exists. In this case, care should
tron signatures representing the pure chemical states, curve fitting be taken in deciding physical constraints to be used during the
of the mixed chemical states of the transition metals is possible. optimization of curve fit parameters.
This is essentially a classical least squares (matrix algebra) problem For most elements, participation in chemical bonding with
in which one has the spectrum of the mixture of compounds as other elements causes an increase in BE, especially when the ele-
well as those of the pure components.50 Figure 6(e) shows a curve ments to which they are bonded are more electronegative. That is,
fit of the spectrum from a mixed Ni-Mo oxide, which has had the for elements having multiple possible oxidation states, more oxi-
background removed, using this type of complex photopeak model. dized chemistries will usually show higher BEs. This pattern is well
The relative ratio of the various chemical states of Ni can be established for carbon, where an increased number of oxygen
extracted from this information, assuming that one knows informa- atoms bonded to a carbon atom leads to a steady increase in its
tion on the physical form of the material (powder, spheres, layered BE.6 Here, the presence of more electronegative oxygen atoms
surface structure, etc.) supporting the chemical composition deshields the carbon atom, effectively increasing the nuclear charge
derived by XPS fitting. felt by its electrons, which raises their BEs (and lowers their kinetic
energies in XPS). In many cases, the least oxidized (lowest BE)
form of an element will produce the “cleanest” peaks in an XPS
3. Finding appropriate FWHM values for unknown spectrum that can then be used to select the best FWHM value for
samples when no references are available the fit. However, this will not always be so—it will depend on the
In certain situations, an analyst is confronted with the need to sample chemistry.
curve fit spectra from complex multicomponent materials where Figure 7 shows a high-resolution C 1s spectrum in which the
finding appropriate references may not be possible and limited highest intensity component results from the contribution of
FIG. 6. (a)–(c) High-resolution Ni 2p region for pure reference compounds curve fitted based on Biesinger et al. (Ref. 45) peak models; (d) pure reference spectra over-
layed; (e) Ni 2p for Ni-Mo oxides fitted using developed peak models. A Shirley background was used.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-13
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
FIG. 7. High-resolution C 1s spectrum of PET fitted using peaks with three different widths to match the right-most slope of the spectrum. Grey—experimental line, green
(dark gray)—peak. The upper insets show zoomed-in areas between 282 and 284.5 eV. The linear background was used.
carbon in the lowest oxidation state in the material, which is also The separation between the 3/2 and 1/2 components of the S 2p
the most abundant. In Fig. 7(a), the FWHM of the component orbitals is fixed to 1.16 eV, and so is the area ratio to 1:2.
peak is too narrow, which results in part of the spectral area being It is worth noting that in the case of p, d, and f levels, spin–
under fitted. In Fig. 7(c), the FWHM value is too broad, which orbit splitting occurs, and in a number of cases, the FWHM value
results in overfitting the right shoulder of the spectrum. The best of the higher BE component is greater than the lower BE compo-
match between the component peak and the experimental data is nent, an effect which has been described as a Super Coster–Kronig
shown in Fig. 7(b). The FWHM value obtained from this match is effect.52 When the FWHM of spin–orbit split components differs,
probably an adequate representation of the physical, chemical, and it is essential that the curve-fitting program has the capability of
instrumental contributions to the data. To a good first approxima-
tion, this FWHM could be used for other component peaks that
might be considered in the fit. Nevertheless, the lone peak at ca.
289 eV in this spectrum is another reasonable choice for determin-
ing the FWHM values of the fit components—it seems to have far
less overlap with other signals than the larger peak around
284.5 eV. It would be advisable to consider both FWHM values in
this curve fitting (they should be quite similar). Further analysis
would be needed to determine which would be most appropriate,
or their average, range of values near the average, etc.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-14
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
Metal Oxide
Ti 2p 6.1 5.7
Si 2p 0.63 N/A
Fe 2p 13.1
In 3d 7.6
Mo 3d 3.15
Ag 3d 6.0
Au 4f 3.7
Hf 4f 1.68–1.71
Pt 4f 3.35
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-15
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-16
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
G. Uniqueness plot
33
XPS of Polymers Database. Figure 15 is an example of a peak In this section, the impact of FWHM on a peak model is con-
model consisting of a single component making use of an exponen- sidered. The data analysis in Fig. 18, performed using a so-called
tial tail lineshape to fit an O 1s peak measured from Nylon 6 pre- uniqueness plot,61 shows that a range of plausible figure-of-merit
pared by cleaning with an argon cluster source ion gun. values is obtained through optimization involving specific FWHM
There are two aspects to fitting data using asymmetric line- values for the Ga 3s component. The uniqueness plot demonstrates
shapes, namely, data reproduction and quantification via a compar- how one can select an informed, fixed FWHM parameter for Ga 3s
ison of component area. The example shown in Fig. 16 illustrates for use in fitting curves to an individual spectrum as the only
one of these aspects. means of establishing a physically meaningful peak model.
The problematic lineshape used in Fig. 16 is a Doniach–Sunjic Each peak model includes parameters for peak position,
profile convoluted with a Gaussian, which certainly allows good data FWHM, and area. FWHM differs from position and area in the
reproduction, but fails to permit accurate quantification as a conse- sense that FWHM depends upon instrumental energy resolution,
quence of an extended asymmetric tail. Care should be taken when data acquisition conditions, as well as on the chemical state of the
using the Doniach–Sunjic lineshape, as component area changes as a sample. Ideally, the BE and relative area for the component peaks
function of the asymmetry parameter, and contributions from will be independent of the instrumental energy resolution.
higher BE components may overlap with an extended tail. Furthermore, there is correlated information between the lineshape
An example of where both data reproduction and quantifica- and FWHM in the sense that lineshape parameters chosen also
tion are important is measuring intensities for an oxide film over alter FWHW for the curve being fitted using the FWHM fitting
an elemental substrate, such as aluminum oxide on aluminum or parameter. By making a uniqueness plot, one can assess how a par-
silicon oxide on silicon. Separation of oxide signal from an elemen- ticular fitting parameter changes the outcome of curve fitting. This
tal signal is best performed using lineshapes for which adjustments technique is applied to the FWHM for the data and curves shown
to component parameters do not have significant consequences for in Fig. 18.
the component area, which is not the case for Doniach–Sunjic Optimization of the peak model shown in Fig. 18 was per-
profiles. For this reason, asymmetry in photoemission peaks is formed using optimization parameters that were free to adjust
typically modeled using lineshapes other than the Doniach–Sunjic within physically meaningful intervals. The standard constraint
profile.59 (discussed in Fig. 8) on the component areas of the S 2p doublet
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-17
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
FIG. 13. High-resolution C 1s and O 1s spectra of PEEK. The colors of the peaks correspond to the colors of the carbon and oxygen atom types in the repeating PEEK
polymer unit above. A background modeled the same was as in Figs. 11 and 12.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-18
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-19
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
FIG. 17. Al 2p spectrum measured by John Walton (University of Manchester) FIG. 18. A material including sulfur and gallium results in spectra with overlap-
from an aluminum foil. An asymmetric lineshape is applied to the narrow alumi- ping component peaks for S 2p and Ga 3s. These component peaks are exam-
num metal 2p doublet components while the oxide components are generalized ples of photoemission with significantly different lineshapes and significantly
Voigt functions. (Ref. 54) The definition of lineshapes is described in Casa different FWHM fitting parameter values. The definition of lineshapes is
Software manual (Ref. 56). A Shirley background is computed using the metal described in the Casa Software manual (Ref. 56). Four-parameter universal
components, which only allows a flat background beneath the aluminum oxide cross section Tougaard background is used.
components. Note how both the aluminum oxide and aluminum metal signals
interact with each other via Lorentzianlike tails to these lineshapes.
o Carbonates and bicarbonates can be distinguished.69
o Different types of phosphates show no difference in the P 2p
region can be especially effective in the identification of different core region, but substantial differences in the valence band
chemical species, where this is particularly true when monochro- region (Fig. 20).70,71
matic X-radiation is used: • Polymers typically give a valence band that is rich in features.
Thus, organic polymers with little or no difference in the C 1s
region often show a valence band with very characteristic
• Metals often have very distinct features in the valence band
features.36
region, for example:
o Molybdenum metal has a distinct four feature pattern at BEs
below 5 eV, which can be explained by the calculated
spectrum.21,66
• Silver metal has a distinct four feature pattern at BEs below 8 eV, J. Reporting of fitting process and parameters
which can be explained by the calculated spectrum.21 To enable others to understand and assess the results obtained
• Oxides often have distinct valence band features, including a from fitting of spectra, it is necessary that both the results and
multiple peak pattern in cases where the core region exhibits one information about the fitting approach, constraints, and assump-
peak with little or no BE shift between different oxidation states, tions and final fitting parameters be included in journal articles or
for example: reports. ISO Standard 19830 describes the minimum reporting
o Molybdenum oxides have distinctly different valence band requirements for peak fitting in XPS.72 Important parameters
spectra.21,67 include spectrometer and x-ray characteristics, the use of charge
o Aluminum oxides and oxyhydroxides cannot reliably be distin- compensation and charge correction, the type of spectral
guished in the Al 2p core region but show substantial differ- pre-preprocessing such as satellite subtraction, smoothing, and
ences in the valence band region.64,68 deconvolution, the type of background used, fitting parameters
• Molecular ions such as sulfates, carbonates, and phosphates have such as lineshapes, FWHM, and χ2, as well as the software used for
distinguishable valence band features, for example: curve fitting.
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-20
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
FIG. 20. Valence Band spectra of phosphorus-oxygen compounds. The experi- REFERENCES
mental spectra are compared with spectra calculated from band structure calcu- 1
lations. A simple phosphate, AlPO4, a ring phosphate Al4(P4O12)3, and a C. J. Powell, Microsc. Today 24, 16 (2016).
2
three-dimensional P/O cluster (P2O5) are shown. P. M. A. Sherwood, Surf. Interface Anal. 51, 589 (2019).
3
M. R. Linford et al., Microsc. Microanal. 26, 1 (2020).
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-21
Published under license by AVS.
TUTORIAL avs.scitation.org/journal/jva
4 40
D. R. Baer and A. G. Shard, J. Vac. Sci. Technol. A 38, 031203 (2020). NIST X-ray Photoelectron Spectroscopy Database, NIST Standard Reference
5
D. Briggs and J. T. Grant, Surface Analysis by Auger and X-Ray Photoelectron Database Number 20 (National Institute of Standards and Technology,
Spectroscopy (IM Publications, Chichester, UK, 2003). Gaithersburg MD, 2000), https://srdata.nist.gov/xps/.
6 41
V. Gupta, H. Ganegoda, M. H. Engelhard, J. Terry, and M. R. Linford, J. Chem. See http://techdb.podzone.net/eindex.html for XPS (X-ray Photoelectron
Educ. 91, 232 (2014). Spectroscopy) Database.
7 42
G. H. Major et al., “Perspective: An assessment of the frequency and nature of M. C. Biesinger, see http://www.xpsfitting.com/ for X-ray Photoelectron
erroneous XPS analyses in the literature,” Appl. Surf. Sci. (submitted). Spectroscopy (XPS) Reference Page (2009–2018).
8 43
M. H. Engelhard, A. H. Gomez, P. M. A. Sherwood, and D. R. Baer, J. Vac. Sci. XPS Reference Table of Elements, Thermo Fisher Scientific (2013–2020).
44
Technol. A 38, 063203 (2020). XPS Reference Table of Elements, Scientific™.
9 45
T. Deegan, “X-ray photoelectron spectrometer calibration and thin film investi- eSpectra visualization too, AIP Publishing LLC.
46
gations on germanium oxides,” Master of science thesis (Dublin City University, A. J. Barlow, R. T. Jones, A. J. McDonald, and P. J. Pigram, Surf. Interface
1998). Anal. 50, 527 (2018).
10 47
J. F. Moulder, Handbook of X-ray Photoelectron Spectroscopy: A Reference M. C. Biesinger, B. P. Payne, L. W. M. Lau, A. Gerson, and R. S. C. Smart,
Book of Standard Spectra for Identification and Interpretation of XPS Data Surf. Interface Anal. 41, 324 (2009).
48
(Physical Electronics, Eden Prairie, MN, 1992). M. C. Biesinger, B. P. Payne, A. P. Grosvenor, L. W. M. Lau, A. R. Gerson, and
11
ISO 16129:2018, X-ray Photoelectron Spectroscopy—Procedures for Assessing R. S. C. Smart, Appl. Surf. Sci. 257, 2717 (2011).
49
the Day-to-Day Performance of an X-ray Photoelectron Spectrometer M. C. Biesinger, L. W. M. Lau, A. R. Gerson, and R. S. C. Smart, Appl. Surf. Sci.
(International Organization for Standardization, Geneva, 2018). 257, 887 (2010).
12 50
ASTM E2735-14, ASTM International, West Conshohocken, PA (2014). R. Kramer, Chemometric Techniques for Quantitative Analysis (CRC Press,
13
J. Wolstenholme, J. Vac. Sci. Technol. A 38, 043206 (2020). Boca Raton, FL, 1998).
14 51
R. Bracewell, The Fourier Transform and Its Applications, 3rd ed. J. H. Scofield, J. Electron. Spectrosc. 8, 129 (1976).
52
(McGraw-Hill Science/Engineering/Math, New York, 1999). N. Mårtensson and R. Nyholm, Phys. Rev. B 24, 7121 (1981).
15 53
G. Major, T. Avval, N. Fairley, and M. Linford, Vac. Technol. Coat. 21, 37 D. O. Scanlon, G. W. Watson, D. J. Payne, G. R. Atkinson, R. G. Egdell, and
(2020), available at https://digital.vtcmag.com/12727/26337/index.html. D. S. L. Law, J. Phys. Chem. C 114, 4636 (2010).
16 54
D. R. Baer et al., J. Vac. Sci. Technol. A 38, 031204 (2020). G. Major, D. Shah, V. Fernandez, N. Fairley, and M. Linford, Vac. Technol.
17
A. Einstein, Ann. Phys. 322, 132 (1905). Coat. 21, 43 (2020), available at https://digital.vtcmag.com/12727/29865/index.
18
P. H. Citrin and D. R. Hamann, Phys. Rev. B 15, 2923 (1977). html.
19 55
W. Voigt, Akad. Wiss. München, Math. Phys. 216, 603 (1912). G. Major, D. Shah, T. Avval, V. Fernandez, N. Fairley, and M. Linford, Vac.
20
V. Jain, M. C. Biesinger, and M. R. Linford, Appl. Surf. Sci. 447, 548 (2018). Technol. Coat. 21, 35 (2020), available at https://digital.vtcmag.com/12727/
21
P. M. A. Sherwood, Surf. Interface Anal. 51, 254 (2019). 26337/index.html.
22 56
J. C. Fuggle and S. F. Alvarado, Phys. Rev. A 22, 1615 (1980). CasaXPS, Casa Software Ltd (2020).
23 57
M. O. Krause and J. H. Oliver, J. Phys. Chem. Ref. Data 8, 329 (1979). E. Desimoni, G. I. Casella, T. R. I. Cataldi, and C. Malitesta, J. Electron.
24
P. L. Lee and S. I. Salem, Phys. Rev. A 10, 2027 (1974); R. C. Rivière, Practical Spectrosc. 49, 247 (1989).
58
Surface Analysis, edited by D. Briggs and M. P. Seah (Wiley, Chichester, UK, M. Schmid, H.-P. Steinrück, and J. M. Gottfried, Surf. Interface Anal. 46, 505
1990), pp. 75–83. (2014).
25 59
D. Roy and D. Tremblay, Rep. Prog. Phys. 53, 1621 (1990); M. Bravo Sanchez, J. A. Huerta-Ruelas, D. Cabrera-German, and
P. M. A. Sherwood, Vibrational Spectroscopy of Solids (Cambridge University, A. Herrera-Gomez, Surf. Interface Anal. 49, 253 (2017).
60
Cambridge, 1972). J. M. Hill, D. G. Royce, C. S. Fadley, L. F. Wagner, and F. J. Grunthaner,
26
C. N. Berglund and W. E. Spicer, Phys. Rev. A 136, 1030 (1964). Chem. Phys. Lett. 44, 225 (1976).
27 61
A. Barrie, Chem. Phys. Lett. 19, 109 (1973). B. Singh, A. Diwan, V. Jain, A. Herrera-Gomez, J. Terry, and M. R. Linford,
28
S. Doniach and M. Sunjic, J. Phys. C Solid State 3, 285 (1970). Appl. Surf. Sci. 387, 155 (2016).
29 62
T. A. Carson, Photoelectron and Auger Spectroscopy, Modern Analytical P. M. A. Sherwood, J. Vac. Sci. Technol. A 14, 1424 (1996).
63
Chemistry (Springer, Boston, MA, 1975). M. O. Krause and J. G. Ferreira, J. Phys. B At. Mol. Opt. 8, 2007 (1975).
30 64
A. Herrera-Gomez, M. Bravo-Sanchez, O. Ceballos-Sanchez, and J. A. Rotole and P. M. A. Sherwood, J. Vac. Sci. Technol. A 17, 1091 (1999).
65
M. O. Vazquez-Lepe, Surf. Interface Anal. 46, 897 (2014). D. S. Jensen et al., Surf. Interface Anal. 45, 1273 (2013).
31 66
D. I. Patel, S. Bahr, P. Dietrich, M. Meyer, A. Thißen, and M. R. Linford, P. M. A. Sherwood, Practical Surface Analysis by Auger and X-ray
Surf. Sci. Spectra 26, 014024 (2019). Photoelectron Spectroscopy, edited by D. Briggs and M. P. Seah (Wiley,
32
D. Shah, S. Bahr, P. Dietrich, M. Meyer, A. Thißen, and M. R. Linford, New York, 1983).
67
Surf. Sci. Spectra 26, 014023 (2019). H. Hixson and P. M. A. Sherwood, Chem. Mater. 8, 2643 (1996).
33 68
M. Salmeron and R. Schlögl, Surf. Sci. Rep. 63, 169 (2008). S. Thomas and P. M. A. Sherwood, Anal. Chem. 64, 2488 (1992).
34 69
K. Roy, L. Artiglia, and J. A. van Bokhoven, Chem. Cat. Chem. 10, 666 (2018). S. Thomas and P. M. A. Sherwood, Surf. Interface Anal. 20, 595
35
A. Shard, J. Vac. Sci. Technol. A 38, 041201 (2020). (1993).
36 70
G. Beamson and D. Briggs, High Resolution XPS of Organic Polymers: The P. M. A. Sherwood, Surf. Sci. Spectra 9, 62 (2002).
71
Scienta ESCA300 Database (Wiley, Chichester, 1992). K. J. Gaskell, M. M. Smith, and P. M. A. Sherwood, J. Vac. Sci. Technol. A 22,
37
C. D. Easton, C. Kinnear, S. L. McArthur, and T. R. Gengenbach, J. Vac. Sci. 1331 (2004).
72
Technol. A 38, 023207 (2020). ISO 19830:2015, Electron Spectroscopies—Minimum Reporting Requirements
38
G. Greczynski and L. Hultman, Prog. Mater. Sci. 107, 100591 (2020). for Peak Fitting in X-ray Photoelectron Spectroscopy (International Organization
39
MultPak 9.9, Ulvac-phi (1994–2016). for Standardization, Geneva, 2015).
J. Vac. Sci. Technol. A 38(6) Nov/Dec 2020; doi: 10.1116/6.0000377 38, 061203-22
Published under license by AVS.