Introduction

Firstly, thank you for using the EPSRC service here at Cardiff, and we hope that your visit was fruitful and informative as it was for us. However, XPS analysis, although relatively straight forward, should be treated with care, especially if you are unaware of the features present in the spectra, together with transmission function correction and which sensitivity factors etc to use.

Here we will give a (brief!) guide to help you in the right direction on the analysis of your data, using CasaXPS software the demo version of which is available from the CasaXPS website, and will allow you to read and manipulate/analyse your saved data however some limitations (e.g. saving) are present in the demo.

Should you wish to know more about XPS (and surface analysis in general), CasaXPS or quantification issues etc, we recommend the following texts/databases/screen cams:

Screen Cams (Copyright Neal Fairley)

Other Analysis Points

Analysing Data

Important things to remember for the analysis of XPS spectra are the corrections for the relative sensitivity factors of the elements/orbitals present in the sample and also the transmission function of the spectrometers analyser. Proper appreciation of these, will yield the true surface composition (within experimental certainty).

(1) Transmission Correction

The VAMAS format files supplied to you have the Transmission function of the Kratos Axis Ultra-DLD instrument (measured in all spectroscopic modes on a clean gold foil) stored within it, so no further work is required for this step. To ensure that at transmission function is present, call up the Region Properties dialogue (F5) and if the Automatic tick box is ticked, then there is a transmission function present, and this is being used. It is recommended to use this transmission function if using the Kratos modified Wagner sensitivity factors (see below).

Should you wish to use theoretically calculated sensitivity factors, or to use a Tougaard background subtraction, the spectrometer intensity response function must be determined absolutely and you should therefore use the transmission function as that developed by the NPL, and as seen in the paper by Walton and Fairley, then you may use our NPL derived transmission function. It is recommended to use this transmission function if you wish to use Scofield's relative sensitivity factors (see below).

(2) Sensitivity Factors (see screen cams also)

The sensitivity factors supplied with CasaXPS are those tabulated by Scofield (Journal of Electron Spectroscopy and Related Phenomena, Volume 8, Issue 2, 1976, Pages 129-137) , although widely used they require further correction, typically for angular and dependency and also the mean free path (MFP) of the photoelectrons, which is not always overly evident, especially if the geometry of the system is unknown. To overcome this the author of CasaXPS supplies a ready-made database of Kratos sensitivity factors, which are based on those by Wagner (Sensitivity factors for XPS analysis of surface atoms (Journal of Electron Spectroscopy and Related Phenomena, Volume 32, Issue 2, 1983, Pages 99-102) but modified to account for the geometry of the system (60 degrees). Since Wagner (and therefore modified Wagner) factors are based on measurements of many compounds, then the MFP dependence is already accounted for. To load the modified factors, you must do the following:

Should you wish to use the NPL transmission function from section (1) and therefore use the Scofield parameters, then together with the MFP dependency discussed above, you must account for the angular dependence of 60 degrees. This is done as follows:

Other Analysis Methods

(1) Overlayer Thickness

The 'Thickogram', developed by Cumpson (Surf. Interface Anal. 29, 403-406 (2000)) is a useful graphical method for measuring overlayer thicknesses in samples where the overlayer has a different elemental chemistry than the substrate (e.g.: Yttria on top of a silicon wafer). The method is useful as it has the following advantages:

1) Uniform surface contamination is unimportant (e.g. adventitious carbon layer)
2) Instrumental factors common to oxide and substrate cancel
3) Simplistic equation
4) Works over a wide range of film thicknesses

However, with all methods of this type, there are a few guidelines and things to note and adhere to:

1) The emission angle must be between 0 and 60 degrees (i.e. Take-off angle of 90 to 30 degrees); Emission angles ca. 45 degrees are the most accurate
2) Method is applicable to a wide range of Kinetic Energies above ~500 eV
3) Results have an associated error of +/-10% (based on accuracy of attenuation lengths obtained by calculations)

 

So, now we know the rules and is applicability, how do we use it?

Firstly we need the following values (where: o = Overlayer, s= Substrate):

I_o = Intensity of Overlayer Peak (or Peak Area)
I_s = Intensity of Substrate Peak (or Peak Area)
S_o = RSF of Overlayer Peak
S_s = RSF of Substrate
E_o = KE of Overlayer Peak
E_s = KE of Substrate Peak
θ = emission angle, (0 for 90 degree takeoff COS(θ) (=1 for 90 degree takeoff)
λ_o = attenuation length of photoelectrons (from the overlayer) in the overlayer. This can be calculated, for example, from the NIST effective attenuation length database.

 

 

Then:

1) Calculate A = (I_o/S_o) /( I_s/S_s) - add this point to the Thickogram
2) Caculate B = (E_o/E_s) - add this point to the Thickogram
3) Draw a straight line between A & B. Point C is then found on the curve.
4) Thickness (T) is then given by T = C(λ_o).COS(θ)

 

(2) Oxide Overlayer Thickness (Especially Si and Al)

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 λ_m) and oxide (λ_ox) is known (or calculated), the oxide film thickness (Amstrongs) can be calculated using the calculations of the type developed by Carlson (J. Elec. Spect. Relat. Phenom, 1972/73; 1, 161) and Strohmeier (Surf. Interface Anal. 1990, 15, 51) which are defined as follows:

Where θ is the photoelectron take-off angle, I_ox and I_m are the percentage areas of the oxide and metal peaks fitted from the high-resolution spectrum, and N_m and N_ox are the volume densities of the metal atoms in the metal and oxide. Note that this equations assumes I_ox and I_m originate from similar photoelectron energies.

 

 

 

(3) Overlayer Thickness on Particles and Fibres ('XPS Topofactors')

Shard et al. (Surf. Interface Anal. 2009, 41 (7) 541-548) introduced the concept of XPS ‘Topofactors’, which can be used in conjunction with the ‘Thickogram’, to provide overlayer thicknesses on topographic samples of known geometry. Their concept is simple; analysis is performed with the sample normal directed towards the XPS analyser, the equivalent planar thickness is calculated from the Thickogram and the Topofactor applied to the result to provide the actual thickness.