Background Selection

In general, three types of backgrounds are used: 1) a simple straight line or linear background, 2) the Shirley background in which the background intensity at any given binding energy is proportional to the intensity of the total peak area above the background in the lower binding energy peak range[1] (i.e. the background goes up in proportion to the total number of photoelectrons below its binding energy position) and 3) the Tougaard background (or Tougaard universal cross-section approach) which is a methodology for integrating the intensity of the background at a given binding energy from the spectral intensities to higher kinetic energies[2]. The simple linear background suffers from large peak area changes depending on the position of the chosen end points and is the least accurate. The Tougaard background is the most accurate but suffers from complications in practical use, particularly if there are numerous peak overlaps. The Shirley background falls somewhere in between for accuracy, however its ease of use has resulted in its widespread adoption.

References:
[1] M.P. Seah Quantification of AES and XPS in Practical Surace Analysis by Auger and X-ray Photoelectron Spectroscopy ed. D Briggs & M.P.Seah , John Weley & Sons, Chichester UK, 1983 p. 204.
[2] Neal Fairly, XPS lineshapes and Curve Fitting in Surface Analysis by Auger and X-ray Photoelectron Spectroscopy ed. D Briggs & JT Grant, IM Publications, Chichester UK, 2003, p. 398.

Precision and Accuracy in XPS

Precision in XPS is quite good given sufficient signal/noise in the spectra obtained. A series of analyses of the same area of a sample surface will give essentially the same result. However, one must be aware that in some cases the sample surface itself can change over time due to X-ray damage. Adsorption/desorption of carbonaceous materials and water can also occur.

Accuracy in XPS will depend on a number of things. One of them is the accuracy of the relative sensitivity factors (RSF) used for quantitation. For many of the common elements the RSF values are quite good while some of the rarer elements are less accurate. For a pristine, homogeneous silicon dioxide surface, XPS analysis will give 66.7% O and 33.3% Si. However surfaces are, in general, never as easy as that. Less photoelectrons will escape the surface unaffected by inelastic losses as one goes deeper into the surface. As such, layering of oxides or contaminant films, discreet particles, island structures, and topographic changes all will change the photoelectron yields for the elements present as compared to a homogenous bulk sample.

Both precision and accuracy will suffer as the peak intensity diminishes and/or with decreases in signal to noise. Changes in endpoint selection can cause large variations in the peak area measured for smaller peak sizes. Elements with a low photoelectron cross-section, such as nitrogen and boron, are particularly susceptible to errors of this type. The background type (linear, Shirley and Tougaard) can also affect the peak areas measured, again affecting accuracy. For most elements/core lines with moderate to strong photoelectron cross-sections, precision may vary by only a few tenths of a percent. This is a minor amount if it is a major element present at the surface (10-100 at.% ) but becomes a proportionally larger error as the amount decreases (less than a few percent). For elements/core lines with lower photoelectron cross-sections (N, B) this error can increase to a percent or more.

Information on XPS detection limits can be found here.

Aluminum

Aluminum oxides and hydroxides (Al2O3, Al(OH)3 and AlOOH) are extremely difficult to differentiate by XPS. A compilation of Al 2p3/2 values and modified Auger parameter values (Al 2p - KL23L23(1D)) are presented below. Unfortunately, statistical speaking, the values for the varying oxides and hydroxides overlap each other.
Table 1. Al 2p3/2 binding energy values [1].

Table 2. Al 2p-KL23L23(1D) modified Auger Parameter values [1].

Reference:
[1] C.D. Wagner, A.V. Naumkin, A. Kraut-Vass, J.W. Allison, C.J. Powell, J.R.Jr. Rumble, NIST Standard Reference Database 20, Version 3.4 (web version) (http:/srdata.nist.gov/xps/) 2003.

Chromium

Fitting parameters for chromium 2p3/2 spectra are based on spectra taken from a series of well-characterized standard compounds [1] (updated and expanded in reference [2], charge-referenced to C 1s at 285.0 eV in [1] and 284.8 eV in [2]). Fitting parameters that can determine Cr(0), Cr(III) oxide, Cr(III) hydroxide and Cr(VI) components were determined and take into account asymmetry in the metal peak, a broader envelope of peaks attributed to multiplet splitting of the Cr(III) compounds and a single peak (no unpaired electrons) for Cr(VI) compounds. Cr(III) species can be further divided into the oxide, which will show discrete multiplet structure, and the hydroxide which shows only a broad peak-shape. The asymmetry in the metal peak is defined here by a Gaussian/Lorentzian product formula modified by an asymmetric form (supplied by CasaXPS software[3]) and is based on spectra from an argon ion sputter cleaned pure metal surface. The FWHM for the metal will depend on the instrument used and should be measured for a particular instrument type. Analysis of the metal peak will also give a good estimate of multiplet splitting peak-widths as the FWHM of the metal peak generally matches that of the individual multiplet peaks under similar spectrometer conditions. Peak-widths for the Kratos Axis Ultra set at a pass energy of 20 eV are around 0.88 eV for the metal and five individual Cr(III) oxide multiplet peaks, while the hydroxide peak is around 2.6 eV. Quantification of Cr(VI) species (single peak at 579.5 eV from average of literature data, FWHM of 1.3 to 1.5 eV to incorporate a variety of Cr(VI) species) is limited by the overlap with the multiplet splitting of the Cr(III) species. This likely raises the detection limits for Cr(VI) in a mostly Cr(III) matrix to around 10% of total chromium. Any contribution attributed to Cr(VI) below that should be treated as “not detected”. An example of this fitting is presented in Figure 1. Fitting parameters are presented in Table 1. A CasaXPS ready example can be found here.  Fitting parameters for a number of other species (e.g. CrCl3, Cr2S3) can also be found in [1]. Figure 1. Fitting of a Cr 2p3/2 peak for a decorative chrome plated part [1].

Table 1. Curve-fitting parameters for Cr 2p3/2 spectra [2].

References:

[1] M.C. Biesinger, C. Brown, J.R. Mycroft, R.D. Davidson, N.S. McIntyre, Surf. Interface Anal. 36 (2004) 1550.
[2] M.C. Biesinger, B.P. Payne, A.P. Grosvenor, L.W.M. Lau, A.R. Gerson, R.St.C. Smart, Appl. Surf. Sci. 257 (2011) 2717.

[3] N. Fairley, 1999-2007 http://www.casaxps.com

C 1s - Carbonates

The C 1s binding energy average for carbonate ((CO₃)^2-) is 289.3 +/- 0.6 eV (average of 18 carbonates, + 4.5 eV above adventitious carbon at 284.8 eV). Calcium carbonate is at 289.5 eV +/-0.1 eV or + 4.7 eV above adventitious carbon). Note that silver carbonate has a much lower binding energy at ~288.5 eV.

Chromium Nitride (CrN)

The Cr 2p3/2 spectrum of chromium nitride (CrN) can be fit using the following peak shape parameters (for a 20 eV pass energy spectrum). 

Peak shape: LA(3.4,25,17)
Peak width: 2.10 eV
Position: 574.6-574.8 eV

Note that other reports suggest a higher binding energy position of up to 576 eV [1,2].

The peak shape is asymmetric and much broader then the metallic peak. There is a slight hump near the top of the peak on the lower binding energy side of the peak.  Assessment of the amount of nitride present in the Cr 2p spectrum should be taken in light of the amount of nitride seen in the N 1s spectrum and the total amount of N and Cr in the survey spectrum. 

The spectrum presented below is consistent with that reported on in Reference [3] taken using a lower resolution instrument.  The spectrum from [3] also shows an asymmetric peak shape and has a FWHM of 2.9 eV. However, the lower binding energy hump is less discernible.

Cr 2p spectrum of chromium nitride (CrN).

References:
[1] O. Nishimura, K. Yabe. J. Electron Spectrosc. Relat. Phenom. 49 (1989) 335.
[2] A. Lippitz, Th. Hubert, Surf. Coat. Technol. 200 (2005) 250.
[3] I. Milosev, H.H. Strehblow, B. Navinsek, P. Panjan, Surface Science Spectra, 5 (1998) 138.

Aluminum Oxide Thickness Measurement

If the metal:oxide ratio can be determined for a thin film oxide sample (~0-9 nm) and if the inelastic mean free path (IMFP, λ) of the metal (λm) and oxide (λox) is known (or can be calculated), the oxide film thickness can be calculated using the calculations of the type developed by Strohmeier [1] and Carlson [2] defined as follows:

d=λoxsinθ ln(((NmλmIox)/(NoxλoxIm))+1) (Eq. 1)

where θ is the photoelectron take-off angle, Iox and Im are the area percentages of the oxide and metal peaks from the high-resolution spectrum, and Nm and Nox are the volume densities of the metal atoms in the metal and oxide, respectively.

For an aluminum oxide film on an aluminum substrate, the Al 2p spectrum has well separated oxide and metal peaks (example shown in Figure 1) and Io and Im values can be readily ascertained. Using the λ and N values proposed by Strohmeier [1], an Excel based aluminum oxide thickness calculator is presented. Just input the percentage of metal from the high-resolution spectrum to get a film thickness in Angstroms. (Note: you must download the file to Excel to use it - it is locked in Google Docs).

Figure 1. Al 2p XPS spectrum of a thin film Al oxide on Al metal with a calculated oxide thickness of 3.7 nm.

References:
1. B.R. Strohmeier, Surf. Interface Anal. 15 (1990) 51.
2. T.A. Carlson, G.E. McGuire, J. Electron Spectrosc. Relat. Phenom, 1 (1972/73) 161.

Cu(0):Cu(II) or Cu(I):Cu(II) Calculations

The presence of the well known shake-up satellite found in Cu 2p spectra as an indication of the presence of Cu(II) species is well known. Recently, a study of the surface chemistry of the flotation separation of chalcocite (Cu₂S) from heazelwoodite (Ni₃S₂) employed a fitting procedure and calculation that quantifies the amount of Cu(II) species present on the surface of Cu(I) sulfide[1] as first developed by Jasieniak and Gerson[2] and now described in reference [3] and [4]. The calculation takes into account the photoelectron yield from both the main 2p_3/2 peak and the shake-up peak and is based on main peak/shake-up peak ratios derived from Cu(OH)₂ standard spectra.

Quantification of the amount of Cu(II) species on a Cu(0) or Cu(I) containing surface does appear to be possible. If, for example, a Cu metal surface is oxidized to Cu(II), the shake-up structure associated with the Cu(II) species can be used for a Cu(0):Cu(II) quantification. Alternatively if Cu(II) species and Cu(I) species are present, the Cu(I):Cu(II) ratio can be determined. This method of Cu(0):Cu(II) (or Cu(I):Cu(II)) determination depends on shake-up peaks that are present in the spectra of d⁹ Cu(II) containing samples but are absent in d¹⁰ Cu(0) (or Cu(I)) spectra. Shake-up peaks may occur when the outgoing photoelectron simultaneously interacts with a valence electron and excites it to a higher-energy level. The kinetic energy of the shaken-up core electron is then slightly reduced giving a satellite structure a few eV below (higher on the calculated BE scale) the core level position[5]. Hence, these electrons are part of the total Cu 2p emission and should be included in both total Cu and relative chemical state speciation. For example, the main emission line (A) in Figure 1 contains both Cu(II) (A1) and Cu(0) (A2) contributions but the satellite intensity (B) is entirely from Cu(II). The total intensity from Cu(II) species is represented in the combination of the signals from the direct photoemission (A1) and the shaken-up photoemission (B).

Accurate Cu(0):Cu(II) ratios for samples containing a mixture of Cu(0) and Cu(II) rely on determining an accurate ratio of the main peak /shake-up peak areas (A1_s/B_s) for a 100% pure Cu(II) sample. With a reliable value of A1_s/B_s obtained for Cu(OH)₂ or CuO (where all copper present is in the Cu(II) state), the relative concentrations of Cu(0) and Cu(II) species present on a surface that contains both species can be obtained by the following simple equations[3,4]:

% Cu(0) = A2/(A+B)*100 = (A-A1)/(A+B)*100 = (A-(A1_s/B_s)B)/(A+B)*100

% Cu(II) = (B+A1)/(A+B)*100 = B(1+(A1_s/B_s))/(A+B)*100

where B is the area of the shake-up peak and A is the total area of the main peak.

In order to determine accurate values of A1_s/B_s, seven Cu 2p_3/2 analyses of pure Cu(OH)₂ were obtained. Analyses were carried out on the various Cu(OH)₂ samples at acquisition times of generally less than a few minutes as it has been shown that reduction of Cu(OH)₂ can occur after extended X-ray exposure[6]. Our studies suggest that after X-ray exposures of 3 h, up to 10% of Cu(OH)₂ has been reduced to Cu(I). At pass energies of 20 eV and 40 eV, A1_s/B_s values of 1.57±0.1 and 1.59±0.1 were found, respectively. A similar analysis of a pure CuO sample was also carried out and gave a A1_s/B_s value of 1.89±0.08 (20 eV pass energy). Figure 1 shows spectra for a sputter cleaned metal surface, CuO and Cu(OH)₂ standards used for A1_s/B_s determination and a spectrum of the native oxide on a pure metal surface with the amount of oxidation of the surface calculated[3,4].

It should be noted that the peak-shape and main peak to shake-up peak separation is quite different for Cu(OH)₂ and CuO (Figure 1). This is useful (along with the O 1s signal if only Cu species are present) in determining which A1_s/B_s value to use for Cu(0):Cu(II) (or Cu(I):Cu(II)) calculations. If the Cu(0) or Cu(I) signal is relatively strong, (and the sample is conducting) some assessment of which is present in the sample may be made based on the BE of the 2p_3/2 peak[3,4].

An excel spreadsheet calculator that uses the equations above for Cu(II) determinations can be found here. (Note: you must download the file to Excel to use it - it is locked in Google Docs).

Figure 1. Cu 2p spectra for a sputter cleaned Cu metal surface (bottom), Cu₂O standard (2^nd from bottom, a small amount of Cu(II) was found in this sample), CuO standard (3^rd from bottom), Cu(OH)₂ standard (4^th from bottom) used for A1_s/B_s determination and a spectrum of a native oxide on a metal surface (top) with the proportion of Cu(0) and Cu(II) calculated [3,4].

References:

[1] M.C. Biesinger, B.R. Hart, R. Polack, B.A. Kobe, R.St.C. Smart, Miner. Eng. 20 (2007) 152.

[2] A.R. Gerson, M. Jasieniak, The Effect of Surface Oxidation on the Cu Activation of Pentlandite and Pyrrhotite, in: W.D. Duo, S.C. Yao, W.F. Liang, Z.L. Cheng; Long H. (Eds.), Proceedings of the XXIV International Minerals Processing Congress, Science Press Beijing, Beijing, China, 2008, pp. 1054-1063.

[3] M.C. Biesinger, L.W.M. Lau, A.R. Gerson, R.St.C. Smart, Appl. Surf. Sci. 257 (2010) 887-898.

[4] M.C. Biesinger, Surf. Interface Anal. 49 (2017) 1325-1334.

[5] J.F. Watts, J. Wolstenholme, An Introduction to Surface Analysis by XPS and AES, Wiley, Rexdale (2003) 71.

[6] W.M. Skinner, C.A. Prestidge, R.St.C. Smart, Surf. Interface Anal. 24 (1996) 620.

Titanium

Initial fitting parameters for the titanium 2p peak were developed using averaged binding energy (BE) data and 2p1/2 – 2p3/2 splitting data from the NIST XPS Database[1] (Table 1). As well, data from readily available standard samples (metal, TiO2) [2,3] were used to clarify the peak-widths, splitting (Δ=6.05 eV for Ti(0), Δ=5.72 eV for Ti(IV)) and shapes (asymmetric for the metallic component) (Table 2) [3]. An example of the use of these parameters is presented for a mixed oxidation state titanium-containing sample in Figure 1. Although C 1s set to 284.8 eV can be used as an internal charge correction it is also possible in this case to use the Ti 2p3/2 metal peak set at 453.7 eV or the clearly defined Ti(IV) (TiO2) 2p3/2 peak set at 458.6 eV. This removes the uncertainty associated with charge correcting to adventitious C especially in situations where the adventitious overlayer is not in good electrical contact with the titanium containing species underneath. The Ti 2p1/2 peak for each species is constrained to be at a fixed energy above the Ti 2p3/2 peak. The intensity ratio of the Ti 2p3/2 and Ti 2p1/2 peaks are also constrained to 2:1. The FWHM’s for the metal and Ti(IV) peaks are derived from the standard sample analyses. The FWHM’s for Ti(II) (at a BE of 455.4 eV) and Ti(III) (at a BE of 457.2 eV), which are likely structurally loosely ordered, are constrained to have equal width to each other and are generally slightly broader than the well ordered Ti(IV) oxide peaks [2,3]. A CasaXPS ready (.vms file) of mixed titanium species can be downloaded here. Further information and expanded discussion can be found in reference [3].

Figure 1. Ti 2p spectrum of a heat-treated Ti-apatite composite using peak fittings derived from Tables 1 and 2.Table 1. Literature values (from [1]) for Ti 2p3/2 spectra.Table 2. Spectral fitting parameters for Ti 2p: binding energy (eV), percentage of total area, FWHM value (eV) for each pass energy, and spectral component separation (eV).


References:
[1] C.D. Wagner, A.V. Naumkin, A. Kraut-Vass, J.W. Allison, C.J. Powell, J.R.Jr. Rumble, NIST Standard Reference Database 20, Version 3.4 (web version) (http:/srdata.nist.gov/xps/) 2003.
[2] M.C. Biesinger, B.P. Payne, B.R. Hart, A.P. Grosvenor, N.S. McIntye, L.W.M. Lau, R.St.C. Smart, Quantitative Chemical State XPS Analysis of First Row Transition Metals, Oxides and Hydroxides, IVC-17/ICSS-13 and ICN+T2007, Stockholm July 2-6, 2007, Journal of Physics: Conference Series 100, 012025 (2008).
[3] M.C. Biesinger, L.W.M. Lau, A. Gerson, R.St.C. Smart, Resolving Surface Chemical States in XPS Analysis of First Row Transition Metals, Oxides and Hydroxides: Sc, Ti, V, Cu and Zn, Applied Surface Science, 257 (2010) 887-898.

Converting From Atomic Percent to Weight Percent and Vice Versa

Sometimes it is necessary to convert from atomic percent to weight percent (or vice versa) to be able to compare XPS data to data from other techniques such as EDX.  This converter (Excel based spreadsheet) allows you to easily convert between the two values. (Note: you must download the file to Excel to use it - it is locked in Google Docs).