Determining Composition of Thin Films of Uranium Oxide by
X-Ray Photoelectron Spectroscopy
Kristi Adamson, Shannon Lunt, Richard Sandberg, Elke Jackson,
Dr. David Allred, Dr. R. Steven Turley
Brigham Young University
Abstract
We have studied thin films (100-200Ǻ) of uranium oxide created through DC
magnetron sputtering. The oxidation of the uranium surface has been examined through
x-ray photoelectron spectroscopy (XPS). This work shows that the surface does not
oxidize immediately, but over a period of several weeks. By comparison with the work
of Teterin [1] (“A Study of Synthetic and Natural Uranium Oxides by X-Ray
Photoelectron Spectroscopy.” J. Phys. Chem. Minerals. 1981), our thinner samples are a
mixture of UO2 and γ-UO3, with γ-UO3 becoming more prominent as the sample has
more time to oxidize. The surface of the sample oxidizes more quickly than the rest of
the sample.
Introduction
X-ray photoelectron spectroscopy (XPS) is a surface sensitive method for
determining composition of thin films utilizing the photoelectric effect. Aluminum is
bombarded with electrons to create a stream of x rays. As these low energy x rays are
quickly absorbed by air, this process must be done at a pressure no greater than 10-6 torr.
These x rays then impinge on our films. If the x ray has sufficient energy, an atom is
ionized, and an electron is ejected from the sample. An electron detector determines the
kinetic energy of the electron.
By knowing the original energy of the x ray and the kinetic energy of the emitted
electron, the energy with which the electron was bound to the atom, Eb, can easily be
determined by the equation:
Eb=hv-Ek
where hv is the energy of the incident x ray and Ek is the kinetic energy of the ejected
electron. From this binding energy, information can be obtained about the relative
abundance of different elements in a sample and its approximate chemical composition.
The depth from which the electrons are sampled depends on two factors: the
kinetic energy of the electron and the electron escape angle. Beyond 100 eV, an electron
with greater kinetic energy can travel through more material before undergoing inelastic
scattering and thus losing its energy. Hence, weakly bound electrons come from deeper
in the sample.
As can be seen qualitatively from the figure below, the sampling depth is also a
function of the electron escape angle.
e-
θ=0°
θ=60°
e-
λ
2λ
1
Figure 1. The dependence of sampling depth on escape angle of electrons. [2]
Electrons escaping at an angle near normal can travel further within the sample
before scattering and thus the sampling depth will be greater. [3] Electrons escaping at a
near grazing angle will give more information about the surface.
Background
Several studies have been done of uranium oxides in bulk using XPS in an
attempt to understand oxidation rates and bonding parameters of oxygen and uranium [1],
[4], [5]. An effort of special interest is determining which electrons in uranium are
involved in the bonding process and the effects these bonds have on the energy of various
uranium electron bands. As the uranium becomes more oxidized, the molecular orbitals
shift in binding energy. These shifts are used to qualitatively understand the behavior of
sample oxidation.
Teterin [1] extensively studied various uranium oxides using XPS. The positions
of the uranium 4f5/2 and 4f7/2 energy peaks are most helpful in determining composition
as it changes with the different amounts of oxygen. The graph below from Teterin shows
how the peaks shift to higher binding energies with increased oxidation.
1
http://goliath.inrs-ener.uquebec.ca/surfsci/arxps/sampling.html
Figure 2. Relative intensity versus binding energy for electrons in the U4f5/2 and 4f7/2 orbitals [1]
The oxygen 1s energy peak also shows great variance for different oxidation
states. As can be seen from the graph, it is actually composed of two closely positioned
peaks that vary in intensity as oxidation changes. The higher energy peak becomes more
prominent as composition changes from UO2 to UO3.
Figure 3. Relative intensity versus binding energy for electrons in the O1s orbital. [1]
Data
We produced thin films with thicknesses of approximately 100Å by DC
magnetron sputtering by depositing uranium on a silicon wafer in the presence of oxygen.
We assumed that the uranium reacted with the oxygen before deposition, thus giving a
homogeneous composition throughout the film.
We have examined the x-ray photoelectron spectra of thin films of uranium oxide
to determine their composition and how it changes as the films are exposed to
atmosphere. Figure 4 shows the oxygen 1s peak of the sample two weeks after they were
produced.
9K
8.5K
8K
7.5K
7K
6.5K
6K
5.5K
5K
4.5K
540.9
538.9
536.9
534.9
532.9
530.9
528.9
526.9
524.9
522.9
Figure 4. O1s peak of uranium oxide sample 2 weeks after deposition
532.6 eV
In comparing these graphs with those found by Teterin [1] it can be seen that there are no
exact matches. Rather, it appears that our samples are some mixture of oxidation states,
most likely a mixture of UO2 and UO3. Two weeks after creation, the thin films showed
a mixture of approximately half UO2 and half UO3. Figures 5-7 below show the oxygen
1s peak as the uranium oxide is continually left exposed to the atmosphere.
22K
21.5K
21K
24K
23K
20.5K
20K
19.5K
19K
18.5K
18K
17.5K
17K
16.5K
16K
15.5K
15K
14.5K
14K
13.5K
13K
540.9
22K
21K
20K
19K
18K
17K
16K
15K
14K
13K
538.9
536.9
534.9
532.9
530.9
528.9
526.9
524.9
522.9
Figure 5. 13 weeks after deposition
540.9
538.9
536.9
534.9
532.9
530.9
528.9
526.9
524.9
522.9
Figure 6. 18 weeks after deposition
18K
17K
16K
15K
14K
13K
12K
11K
10K
9K
540.9
538.9
536.9
534.9
532.9
530.9
528.9
526.9
524.9
522.9
Figure 7. 25 weeks after deposition
These graphs show that as the films have more time to oxidize, the higher energy peak
becomes more intense while the lower energy peak diminishes, indicating a higher
oxidation state.
The uranium 4f5/2 and 4f7/2 peaks were also studied to verify oxidation states.
Figure 8 shows the spectrum two weeks after deposition.
382.42 eV
392.97 eV
12K
11K
10K
9K
8K
7K
6K
5K
4K
3K
2K
1K
400.9 398.9 396.9 394.9 392.9 390.9 388.9 386.9 384.9 382.9 380.9 378.9 376.9 374.9 372.9
Figure 8. U4f peaks of uranium oxide 2 weeks after deposition
392.52 eV
380.92 eV
It can be observed from Figures 9--11, that as the films were exposed to
atmosphere, the uranium peaks shifted to lower energies for at least the first 18 weeks,
but then the trend began to shift to higher energies.
40K
38K
45K
36K
34K
40K
32K
30K
35K
28K
26K
30K
24K
22K
25K
20K
18K
20K
16K
14K
15K
12K
10K
10K
8K
6K
411.9
406.9
401.9
396.9
391.9
386.9
381.9
376.9
Figure 9. 13 weeks after deposition
411.9
406.9
401.9
396.9
391.9
386.9
381.9
376.9
Figure 10. 18 weeks after deposition
30K
28K
26K
24K
22K
20K
18K
16K
14K
12K
10K
8K
6K
411.9
406.9
401.9
396.9
391.9
386.9
381.9
376.9
Figure 11. 25 weeks after deposition
All of these figures are from XPS scans with an incidence angle of 55º to the
surface normal. To observe whether a different spectrum was produced with different
incidence angles, the uranium 4f5/2 and 4f7/2 peaks were also examined with an angle of
75º from normal.
30K
30K
28K
28K
26K
26K
24K
24K
22K
22K
20K
20K
18K
18K
16K
16K
14K
14K
12K
12K
10K
10K
8K
8K
6K
6K
411.9
406.9
401.9
396.9
391.9
386.9
381.9
376.9
411.9
406.9
Figure 12. 25 weeks after deposition
Incidence Angle=55ْ
401.9
396.9
391.9
386.9
381.9
376.9
Figure 13. 25 weeks after deposition
Incidence Angle =75ْ
We see from Figures 12 and 13 that with a greater incidence angle, and an
effectively more shallow sampling depth, the uranium peaks shift to slightly higher
energies.
Discussion
From the two observed peaks, we see two different trends. The oxygen 1s peak
indicates that the sample is moving to a higher oxidation state. Figure 14 shows how the
composition changes with time. It appears as though the sample will completely oxidize
to UO3 with time.
Percent Composition
Composition
100
80
60
40
20
0
0
10
20
30
Time (weeks)
Figure 14. Oxidation states of uranium oxide over time. The top line represents the percentage of UO3 while the bottom represents
that of UO2.
The uranium 4f peaks, however, show a different trend. If the sample is shifting
to a higher oxidation state, the uranium peaks should move to a higher energy. Yet, we
can see from the charts that the peaks moved to a lower energy, indicating that the sample
is moving to a lower oxidation state. This trend occurred for the first 18 weeks, after
which the peaks began to shift to higher energies. This is due to the fact that, as noted
previously, the sampling depth of XPS is a function of the electron’s kinetic energy. The
kinetic energy for the oxygen 1s electrons is approximately 950 eV while that of the
uranium 4f electrons is 1110 eV. From the graph below [6], the difference in sampling
depth can be determined. (The y-axis is measured in nanometers, while the x-axis is in
electron volts.)
The sampling depth of an electron as a function of energy [6]
This difference in kinetic energy equates to a difference of about 10Ǻ in sampling
depth. Hence, the oxygen 1s peaks are showing us the oxidation of the surface, while the
uranium 4f peaks are showing the oxidation behavior of the layers just below the surface,
and it is obvious that the two trends are not similar.
The scans run at various angles are in agreement with what was determined by the
uranium 4f and oxygen 1s peaks. The scan done at 55º (which samples a greater depth)
shows a slightly lower oxidation state than the scan done at 75º.
Conclusions
At this point, it is clear that the uranium oxide samples continue to oxidize as they
are exposed to atmosphere. More time is required to determine if the uranium oxide
approaches a stable oxidation state. It appears that the samples will completely oxidize to
γ-UO3, rather than U3O8 or U3O7, which has been suggested by previous literature [7].
This is most likely due to the variances in behavior between bulk samples and thin films.
It is also determined that the oxidation state is not constant throughout the sample
but changes with depth. While the surface shows increased oxidation, the layers below
show, at first, a decrease in the amount of oxygen. This behavior was unexpected and is
not yet completely understood. More investigation is required in this area to determine
the behavior of the uranium oxides below the surface.
Bibliography
[1] Teterin, Yu. A., V. M. Kulakov, A. S. Baev, N.B. Nevzorov, I.V. Melnikov, V. A.
Streltsov, L. G. Mashirov, D. N. Suglabov, A. G. Zelenkov. “A Study of Synthetic and
Natural Uranium Oxides by X-Ray Photoelectron Spectroscopy.” J. Phys. Chem.
Minerals. Vol. 7, 151-158 (1981).
[2]
Painter, Royston, http://goliath.inrs-ener.uquebec.ca/surfsci/arxps/sampling.html,
March 16, 2002.
[3] Baird, J. R. and C. S. Fadley, “X-Ray Photoelectron Angular Distributions with
Dispersion-Compensating X-Ray and Electron Optics.” J. Elec. Spec. Relat. Phenom.
Vol. 11, 39-65 (1977).
[4] Allen, Geoffrey C. and Nigel R. Holmes. “Characterization of Binary Uranium
Oxides by Infrared Spectroscopy.” Applied Spectroscopy. Vol. 48 (4), 525-530. (1994).
[5] Teterin, Yu. A. “Structure of X-Ray Photoelectron, Emission and Conversion
Spectra and Molecular Orbitals of Uranium Compounds.” J. Structural Chem. Vol. 39
(6), 850-857 (1998).
[6]Seah, M.P., W. A. Dench. “Compilation of Experimental Data Determined with
Various Electron Energies for a Large Variety of Materials.” Surf. Interface. Anal. I
(1979).
[7] Cordfunke, E. H. P. The Chemistry of Uranium. Elsevier Publishing Company,
Amsterdam. 1969: 64-97.