Banner for the Material Measurement Laboratory at NIST

The Surface and Microanalysis Science Division


Interactive Standard Test Data


Objective

In keeping with NIST's tradition of providing high-quality reference materials and data, we are providing well-characterized spectral data for assessing quality in computer-based data analysis procedures. Known as Standard Test Data, the data presented here are simulations of X-ray photoelectron spectra (XPS-STD).

Like NIST's Standard Reference Materials (SRMs), Standard Test Data are used to assure the quality of the chemical measurement process. While SRMs are used to assess the complete chemical measurement process from sample preparation/measurement through data evaluation, Standard Test Data are used solely to assess the quality of the data evaluation domain; in particular, computer-based data-analysis procedures that convert instrument responses to relevant chemical information.

From this site, you will be able to download the XPS-STD. Then, using your own data analysis software, you determine the number of peaks in the spectrum, positions of the peaks, and their intensities (areas). After completing the data analysis, you can then upload the peak parameter estimates for an on-line determination of the error in your results in terms of bias and random error. By minimizing bias and random error, customers may determine which analysis procedures produce high accuracy and precision.


Approach

Since the XPS-STD are simulations, peak parameters such as peak location and peak intensity are known from the models. They are not estimates. The test data are designed to help the analyst assess data analysis procedures from the errors in the analyst's peak parameter determinations. Performance is measured by the deviations of the analyst's parameter estimates from the true values for these parameters. In addition to assessing data-analysis procedures, the test data are designed to help the analyst decide which algorithms to use under specific circumstances.

There are 220 spectra in the XPS-STD set. The simulations are spline polynomial models of measured polymer carbon 1s spectra, which were provided by Dr. D. Briggs, formerly of ICI Wilton Research Center, U.K. Figure 1 shows the measured spectra of the three polymers with a single main C 1s peak used in this study.

Figure 1 Figure 1. Measured C 1s XPS spectra of poly(dimethylsiloxane) (PDMS), poly(ethylene sulfide) (PETHS) and p-quaternary phenyl[poly(phenylene)oligamer] (QUAT). Click on the figure for a larger view.

Among the 220 XPS-STD spectra, 180 are doublets and 40 are singlets. The doublet spectra were constructed by adding two modeled C 1s spectra to simulate a spectrum of overlapping peaks that might be measured from a specimen consisting of two polymers. Following a factorial design, the doublet spectra present 1) varying degrees of overlap between peaks, 2) varying levels of relative intensity between peaks, and 3) varying levels of Poisson noise. Figure 2 shows the factorial design with three levels for the separation between peaks, three levels for the relative intensities of the peaks, and two levels for the Poisson noise.

Figure 2 Figure 2. Factorial design of doublet spectra in the XPS-STD set. Click on the figure for a larger view and additional details of the design.

Among the doublet spectra, there are 18 spectral conditions, i.e., 18 factor combinations. In addition to the doublets, two sets of singlets (PDMS and PETHS) were simulate as individual polymer C 1s peaks at the two Poisson noise levels for a total of 4 singlet spectra. The 22 doublet and singlet spectra were replicated 10 times, each with the same amount of fractional Poisson noise for a total of 220 spectra.


Results of the XPS-STD Pilot Study

Curvefits of the XPS-STD by 20 analysts using a variety of software procedures, revealed peak binding energies that deviated substantially from the true binding energies. These deviations from true values, calculated as biases and random errors, are the evaluated results of the Study Group. Customers using this web site can compare the bias and random error in their data analysis with the statistics for the Study Group.

For additional information see the Pilot Study.


Procedure for Using the XPS-STD

Here is an overview of the steps involved in using the XPS-STD:

  1. Download the XPS-STD consisting of 10 sets of 22 spectra for a total of 220 test spectra. In addition to the test spectra, two reference spectra are included. One reference spectrum contains the measured values for one of the polymers, so this spectrum exhibits a single main C 1s peak. The other reference spectrum, which also exhibits a single C 1s peak, contains the spline-modeled values for the same polymer. These reference spectra may be used to determine initial curve-fitting parameters.
  2. The order of the 220 test spectra is randomized. Analyze all 220 spectra or any multiple of 22 (i.e., complete sets) to determine the number of peaks in the spectra, binding energies of peaks (i.e., location) and intensities of peaks. You must, however, analyze the sets in order. In other words, for example, analyze sets 1,2,3,4,5 or 2,3,4,5,6 but not 1,3,4,5,6. Use whatever data analysis software you prefer. Since XPS line shapes may be approximately Gaussian-Lorentzian, most analysts have used algorithms that fit non-linear Gaussian and/or Lorentzian functions with a least-squares regression procedure.
  3. Upload your data analysis results to determine the amount of bias and random error.
  4. Indicate your curve-fitting approach.
  5. Compare your errors to the statistics of the Study Group that used the same curve-fitting approach.
  6. To minimize data analysis errors, modify your data analysis procedures, reanalyze the XPS-STD, and resubmit results for evaluation.

Download from here a zipped set of the XPS-STD Data Set. Within this zip file, there are 10 groups of 22 data files, thus a total of 220 files. Also included are the two reference files, "Raw.ref" and "Model.ref" which are used to determine initial parameters for analysis. After downloading this file, you must unzip it with an unzipping utility, such as Winzip.

Submit your data analysis results and get the On-Line Evaluation of an XPS-STD Analysis.


References

J.M. Conny, C.J. Powell, and L.A. Currie, "Standard Test Data for Estimating Peak-Parameter Errors in X-Ray Photoelectron Spectroscopy. I. Peak Binding Energies," Surf. Interface Anal., 26, 939-956 (1998).

J.M. Conny and C.J. Powell, "Standard Test Data for Estimating Peak-Parameter Errors in X-Ray Photoelectron Spectroscopy: II. Peak Intensities," Surf. Interface Anal., 29, 444-459 (2000).

J.M. Conny and C.J. Powell, "Standard Test Data for Estimating Peak-Parameter Errors in X-Ray Photoelectron Spectroscopy: III. Errors With Different Curve-Fitting Approaches," Surf. Interface Anal., 29, 856-872 (2000).


Who We Are

The XPS-STD project was inspired by Lloyd Currie who pioneered the concept of Standard Test Data (L.A. Currie, J. Res. Natl. Bur. Stand., 90, 409, 1985; L.A. Currie, J. Res. Natl. Bur. Stand., 93, 193, 1988). The XPS-STD were designed by Joseph Conny and Cedric Powell. This website was designed by Joseph Conny. Programs (Java) for data retrival and data evaluation were written and implemented by Jeffrey Lee.


Please send questions or comments about XPS-STD or this web site to Joseph M. Conny

The National Institute of Standards and Technology (NIST) is an agency of the U.S. Department of Commerce.

Privacy policy / Security policy / Accessibility / Disclaimer


Last Updated: March 2, 2012