Schilke Lab

 X-Ray Photoelectron Spectroscopy (XPS)

Schematic diagram of an XPS system.
(See training documents for details)

XPS peak shift database (NIST) / Binding energy tableΔ

X-ray photoelectron spectroscopy (XPS) is a surface-analysis technique that is extensively used to characterize the atomic composition of a vacuum-compatible sample. It is sensitive to all elements except H and He, provides quantitative and qualitative compositional data, and examines the top 1-10nm of the sample. Chemical state information, including ionization state and chemical bonding, can be determined by observing shifts in the location of peaks. Shown at right is a schematic diagram of the XPS instrument. See below for an explanation of the theory and operation.

An XPS “survey spectrum” identifies all atoms at the surface, and can typically be obtained in just a few minutes; high-resolution scans for quantification or identification of trace elements can take up to several hours. Representative images of these results are shown below. A variety of post-processing can be performed on this data, including calculation of compositions, layer thickness, and in some cases, even an estimate of the surface coverage and morphology.

Composition depth profiles can be created by changing the “takeoff angle” (the incident angle of the X-ray source), or by performing multiple scans, destructively sputtering the surface to reveal lower layers between each scan. More complicated experiments are also possible.

The nearest XPS instrument is housed at the University of Oregon in the CAMCOR XPS Facility. The institutional fee is approximately $40 per beam-hour, so batching of samples is prudent.

For scheduling and training, contact

// <![CDATA[
Nix.decode(“”);
// ]]>Dr. Stephen Golledge// <![CDATA[
Nix.decode(“”);
// ]]>

. We also have a set of XPS training documents generously provided by CAMCOR.

Theory of Operation

Very briefly, the XPS instrument projects a narrow X-ray beam of known energy onto the surface of a sample in an ultra-high vacuum. The photoelectric effect causes electrons to be ejected from the surface. These electrons have a characteristic kinetic energy that is dependent upon the atom’s quantum structure. The difference between the input X-ray energy and the released electron is called the Binding Energy (BE)Δ, measured in electron volts. Changes in the electronic structure caused by chemical bonding are manifested as shifts in the BE of the elements, and can be used to identify species on the surface.

The electrons emitted from the surface are then collected and focused by a magnetic lens, and passed into a hemispherical analyzer. Only electrons with the correct kinetic energy can reach the detectors, where they are counted and plotted as a function of binding energy (i.e. the opposite of their kinetic energy).

There are many good references about XPS. A short list is given here:

• Wikipedia article on XPS
• Presentation on ESCA/XPSΔ from the 2009 NESAC/Bio workshop (content restricted to McGuire/Kelly lab members)

Example Spectra and Interpretation

An example of an XPS survey scan of a sample of polyurethane is shown at right. The labeled peaks are characteristic of the atoms carbon, nitrogen and oxygen. The Auger peak (at the far left) is from electron-electron interactions within an atom. Interpretation of Auger peaks is somewhat complicated, so we will ignore it here.

The sloping background is characteristic of XPS spectra, and is due to the occasional ejection of electrons from the surface that have lower kinetic energy because they have interacted with atoms on the surface. The large “dip” and shape of the background to the left of each element can be used to model the surface distribution and layer thickness. Other features are due to resonances and other effects.

The exact position and shapes of the peaks (all from 1s-orbital electrons) provides information about the chemical state of the atoms in the sample, but the survey does not have enough resolution to be useful. For this, we turn to high-resolution scans of the peaks corresponding to individual atoms. These high-resolution scans for C_1s, N_1s, and ‘O_1s are shown below:

In these scans, the different chemical states of the constituent atoms can be clearly seen. At left, three peaks have been fitted to the data, corresponding to the three different types of carbon bonds in the urethane bonds, highlighted at right. The “shake-up” peak is due to intra-atom energy transfer due to π-π^* resonances. In the nitrogen spectrum, we see only the urethane bond, while the oxygen peak shows two species corresponding to the two oxygen types.

Quantitative Analysis

Since XPS is a surface-sensitive technique, we can make use of that fact to quantify the effective thickness of an adsorbed layer, just as in ellipsometry. In this case, however, we can make use of the attenuation of electrons from the underlying surface by an overlayer. Just as with a translucent paint, the thicker the overlayer is, the less of the surface signal will be seen.

This method has been used by many authors, including Schlapak, et al. (Langmuir 2006, 22, 277-285), who determined the thickness of immobilized PEO layers on silicon. We can apply the Beer-Lambert Law to derive the following equation:

where λ is the inelastic mean free path (IMFP) of an electron within the overlayer structure. This term is similar to an extinction coefficient in normal photometry; it is a measure of how quickly the signal drops off as the distance increases. The angle θ is the take-off angle (usually 0° unless you set it otherwise). The values of I_film and I_substrate are the background-subtracted peak intensities of the film-carrying and bare substrate, respectively. The inelastic mean free path of an electron in PEO was given as 3.4nm by Schlapak, et al., but the QUASES modeling software (Sven Tougaard) gives ~28Å (2.8nm).

For instance, consider an immobilized PEO polymer film on SiO₂. The PEO-free surface will have a strong Si_2p peak from the surface. When the PEO is immobilized, electrons released by the silicon atoms will be attenuated while passing through the PEO layer, and the Si_2p peak intensity will decrease. By inserting the baseline-adjusted values into the equation above, the average effective film thickness can be computed. From this value and a bulk density for PEO, the mass/cm² of the film can be calculated, and from this, the areal density and other goodies.