The Issue of Pu 5f Occupation

Transcription

1 4/11/2018 Tobin, U. Wisc. Oshkosh Page 1 of 6 The Issue of Pu 5f Occupation J.G. Tobin University of Wisconsin-Oshkosh, Oshkosh, WI, 54901, tobinj@uwosh.edu A brief review of the status of our understanding of the occupation of the Pu 5f states is provided herein, with an emphasis upon experimental measurements. 1.INTRODUCTION: Initially it was believed that the actinide series was a 6d filling series [1], based upon the atomic volumes [2] and associated measurements. However, it became clear that physical properties such as the atomic volume and bulk modulus could be explained in terms of the 5f behavior. [3] Today, the prevailing view, of the electronic filling in the actinides series through Am, is as shown in Table I. Here, we have included the 7p with the 6d and 7s, to be consistent with the spd trivalent picture, as also utilized in the lanthanides. Using Density Functional calculations with a 5f 5 model, Soderlind and Sadigh achieved the notable success of being able to predict the atomic volumes of all six solid phases of Pu. [6] Unfortunately, their model also predicted large and unphysical magnetic moments for Pu. [2] It should be noted that within essentially all atomic pictures, whether in the Russell Saunders or jj limits, a 5f 5 occupation will give rise to a magnetic moment in Pu. [7,8] In the effort to remove the magnetic moments from the calculated Pu results, there were many approaches that utilized an increased 5f occupation, as high as 5f 6. [9 11] Obviously, within a jj limit picture, a 5f 6 filling would produce complete magnetic cancellation. However, a Pu 5f 6

2 4/11/2018 Tobin, U. Wisc. Oshkosh Page 2 of 6 occupation, and even a near 5f 6 occupation, is inconsistent with the earlier experimental work using Pu 4d and 5d X ray absorption spectroscopy (XAS) and related electron energy loss measurements (EELS). [12 16] The earlier 4d and 5d XAS and EELS experimental results and subsequent measurements will be described next. 2.EXPERIMENTAL SPECTROSCOPIC RESULTS: 2.1 4d/5d XAS and EELS Figure 1.(left) Here are shown representative spectra for the 5d and 4d XAS of Pu and U. [17] All of the Pu samples were from metallic samples. All of the 4d XAS spectra are from metals:alpha-pu and alpha U. The U5d spectrum is from UO 2 but this illustrates a key point: all of the U5d XAS, regardless of form, appears to have the same spectra shape. [18] This was first observed by Kalkowski, Kaindl, Brewer and Krone in the 1980s. [19] The central points of the 4d/5d XAS and EELS results can be summarized as shown in Figure 1 and Reference 17. (Also see [18-19]) Here the Pu 5d spectrum has the absence of a pre peak and the 4d spectrum has the diminishment of the 4d 3/2 peak relative to the 4d 5/2 peak. All of the effects observed in the Pu spectra arise because of the large spinorbit splittings in both the 4d/5d and 5f levels and the strong electric dipole selection rule that forbids transitions from the d 3/2 initial state into the 5f 7/2 final state. In the case of the 5d, the fingerprint of the 5f 5 configuration is the absence of the pre-peak and the presence of both the 5d 5/2 and 5d 3/2 edges in the main part of the spectrum, which is often referred to as the giant resonance, following the terminology used for lanthanides. [14] The prepeak disappears because the 5f 5 configuration has almost filled the 5f 5/2 substructure. [12,14] With the transfer of an electron from the 4d levels, the 5f 5/2 substructure is now filled (5f 6 ) and angular momentum coupling between the 5d core hole and the 5f levels is precluded. This observation rules out the 5f 4 initial state. However, because both the 5d 5/2 and 5d 3/2 features are observed in the 5d giant resonance, it is also clear that the

3 4/11/2018 Tobin, U. Wisc. Oshkosh Page 3 of 6 configuration cannot be 5f 6. For 5f 6, there is only one feature in the giant resonance, associated with the 5d 5/2 initial state. This is because the 5d 3/2 to 5f 7/2 transition is forbidden and the 5f 5/2 subshell is now full. The 5d XAS measurements upon Am confirm this analysis. [14, 20] Similarly, the diminishment of the 4d 3/2 peak in the Pu spectrum is driven by the selection rules that forbid the 4d 3/2 to 5f 7/2 transitions. [14, 5] Furthermore, for the Am 4d 3/2 peak, the peak is almost gone, lost in the noise. [14, 16] Using the atomic models developed by G. van der Laan, based upon Cowan s codes [21], it is possible to get quantitative agreement between experiment and theory for Pu and Am. [15], by merely postulating that Pu is 5f 5 and Am is 5f 6. [15, 16] Essentially all of the 4d and 5d XAS and high energy 4d and 5d EELS measurements in the actinides are in complete agreement, [12 16] with only one notable exception. [18] Because of the large lifetime broadening in the actinides, there is a relative paucity of fine structure with which to test this further, but such fine structure does exist in the corresponding transitions in the lanthanides. An example of the complete agreement between the XAS, EELS and atomic theory is shown in Reference 22, using Ce as the sample. It should be noted that the high energy of the excitation beam (300,000 ev) is a key component to the convergence of the EELS behavior to the XAS limit. The strength of the agreement between 4d/5d XAS experiment and theory puts limitations upon how far the Pu 5f occupation can diverge from 5. [14] For example, a 5f 5.5 configuration, assuming two components of 5f 5 and 5f 6 natures and that they would contribute equally to the spectra, would imply that that there would be significant changes in the spectra. In all likelihood, even a Δn = 0.3 deviation, n(5f) = (5.0 +/ 0.3) would be observable. This will be used later in the final analysis. Before going on to other techniques, it is important to note explicitly that for the 4d/5d XAS and EELs, the results for alpha Pu and delta Pu are almost identical, as can be seen in Reference 14. These results indicate strongly that the n(5f) values for alpha Pu and delta Pu are very nearly the same, if not completely identical. 2.2 Supporting Photoelectron Spectroscopy Results The 4d/5d XAS and EELs results are supported by other measurements, including Photoelectron Spectroscopy (PES). Both high-resolution PES [23], performed in house at LANL, and Resonant PES [24] support the 5f 5 picture. 2.3 RXES Results Perhaps the most powerful additional measurement is another variant of XAS, the L 3 (2p) Resonant X ray Emission Spectroscopy, (RXES), performed by Booth and co workers. [25, 26] The RXES experiment is a two photon experiment with both the incoming and emitted photons in the hard X ray regime. Above the core level threshold, the RXES becomes a type of higher resolution XAS, where the improved resolution comes from the detection of an emitted photon generated in a core to core decay. Below threshold, the experiment resembles a type of Raman spectroscopy, going through virtual states. Near threshold, the states accessed include those above the Fermi energy but below threshold. RXES is a two-step process involving an excitation by an incident photon out of the initial, or ground, state. For example, a 2p electron to a state above the Fermi level, such as a 6d state. The second step involves the decay of a 3d electron that fills the 2p hole and emits a photon with different energy. [26.] The sensitivity to n(5f) comes from the long observed effect that L 3 XANES edges in actinides shift with electronic configuration. [27,

4 4/11/2018 Tobin, U. Wisc. Oshkosh Page 4 of 6 28] Using this approach, Booth et al. arrived at the n(5f) determinations for Pu shown in Table II. Table II α Pu δ Pu(Ga) n(5f) f4:f5:f :0.46: :0.38:0.45 n(5f) / / 0.15 f4:f5:f :0.56: :0.46:0.46 n(5f)(this work) Here, f4, f5, and f6 are the fractional occupations of the 5f 4, 5f 5 and 5f 6 states, respectively. While the mixed configuration approach agrees with the analysis propagated by Shim, Haule and Kotliar, [29] using a variant of DMFT, it is problematic when compared to the 4d/5d XAS and EELS analysis. (It should also be noted that SHK also predicted a greater dominance of 5f 5 component. Moreover, their value for n(5f) for delta Pu is 5.2.) It is unclear why effects associated with the minority configurations do not appear in the 4d/4d XAS and EELS spectra. Nevertheless, the estimates of n(5f) of 5.2 for alpha Pu and 5.3 for delta Pu agree well with the earlier estimates for the 4d/5d XAS and EELS measurements. 2.4 Cluster Results Finally, it is useful to consider the possibility of linking the atomic and bulk configurations for Pu. One way to do this is through the utilization of cluster calculations. Pu cluster calculations have been carried out by Ryzhkov et al. [30,31] While the properties of the center-most Pu should reflect bulk behavior, averaging over the entire cluster should give a measure of the size dependent effects. Shown in Reference 30 is a comparison of PES results from Pu [24] with predictions for the central part of a cluster. Good agreement is obtained, confirming the validity of the approach. A summary of the size dependences and the previously discussed results is shown in Figure 2. From the 4d/5d XAS and EELS, the best estimate of the n(5f) for Pu is 4.7 n(5f) 5.3. From the L 3 (2p) RXES, one obtains the estimate that n(5f) is 5.2 to 5.3, with 5.2 for alpha Pu and 5.3 for delta Pu.. Interestingly, the cluster calculations agree with completely, with an asymptotic approach of n(5f) toward 5.2 to 5.3 at larger sizes. Furthermore, one also observes that the correct behavior is produced for the 6d, 7s and 7p occupations, moving from a 7s 2 for the atom to a strong 6d occupation in the bulk, consistent with the formation of 6d related bonding. [32] The 7p and 7s populations are consistent with elements of delocalization. 3. FINAL CONCLUSIONS Thus, it has been successfully shown how the Pu5f population goes from 6 in the atom to a value near 5 in the bulk, with the best estimate of the bulk n(5f) being 5.2 to 5.3, with 5.2 for alpha Pu and 5.3 for delta Pu. It should also be noted that the one unfortunate exception [16, 18] has been overturned by a more recent result. [33] Acknowledgment: Based upon LLNL-BOOK , by JG Tobin, Nov 2014.

5 4/11/2018 Tobin, U. Wisc. Oshkosh Page 5 of 6 Figure 2 Here is shown a summary of the results. See text for details.

6 4/11/2018 Tobin, U. Wisc. Oshkosh Page 6 of 6 References 1. W.H. Zachariasen, J. Inorg. Nucl. Chem. 35, 3487 (1973). 2. J.C. Lashley, A. Lawson, R.J. McQueeney and G.H. Lander, Phys. Rev. B 72, (2005). 3. H.L. Skriver, O.K. Andersen and B. Johansson, Phys. Rev. Lett. 41, 42 (1978). 4. N.W. Ashcroft and N.D. Mermin, Solid State Physics, New York, J.R. Naegele, Actinides and Some of their Alloys and Compounds, Electronic Structure of Solids: Photoemission Spectra and Related Data, Landolt Bornstein Numerical Data and Functional Rel. in Sci. and Tech., ed. A Goldmann, Group III, Volume 23b, Pages (1994); and ref. therein. 6. P. Soderlind and B. Sadigh, Phys. Rev. Lett. 92, (2004). 7. S.Y. Savrasov and G. Kotliar, Phys. Rev. Lett. 84, 3670 (2000). 8. Michael Brooks, multiple communications. 9. A.O. Shorikov, A.V. Lukoyanov, M.A. Korotin and V.I. Anisimov, Phys. Rev. B 72, (2005). 10. A. Shick, et al., Phys. Rev. B 73, (2006) and references therein. 11. A. Svane, L. Petit, Z. Szotek, and W.M. Temmerman, Phys. Rev. B 76, (2007). 12. K.T. Moore, M.A. Wall, A.J. Schwartz, B.W. Chung, D.K. Shuh, R.K. Schulze, and J.G. Tobin, Phys. Rev. Lett. 90, (2003). 13. J.G. Tobin, et al., Phys. Rev. B 72, (2005). 14. J.G. Tobin, et al., J. Phys. Cond. Matter 20, (2008). 15. G. van der Laan, K.T. Moore, J.G. Tobin, B.W. Chung, M.A. Wall, and A.J. Schwartz, Phys. Rev. Lett. 93, (2004). 16. K. T. Moore, et al., Phys. Rev. Lett 98, (2007). 17. J.G. Tobin, S. W. Yu, and B.W. Chung, Top. Catal. 56, 1104 (2013). 18. J.G. Tobin, J. Electron Spectroscopy and Rel. Phen., 194, 14 (2014). 19. G. Kalkowski, G. K. Kaindl, W. D. Brewer, and W. Krone, Phys. Rev. B 35, 2667 (1987). 20. M. T. Butterfield, et al., Phys. Rev B 77, (2008). 21. R. D. Cowan, The Theory of Atomic Structure and Spectra (University of California Press, Berkeley, K.T. Moore, B.W. Chung, S.A. Morton, S. Lazar, F.D. Tichelaar, H.W. Zandbergen, P. Soderlind, G. van der Laan, A.J. Schwartz, and J.G. Tobin, Phys. Rev. B 69, (2004). 23. J. X. Zhu, A. K. McMahan, M. D. Jones, T. Durakiewicz, J. J. Joyce, J. M. Wills, and R. C. Albers, Phys. Rev. B 76, (2007). 24. J.G. Tobin, et al., Phys. Rev. B 68, (2003). 25. C.H. Booth, et al., Proc. Natl. Acad. Sci. 109, (2012). 26. C.H. Booth, et al., Journal of Electron Spectroscopy and Related Phenomena 194, 57 (2014). 27. A.L. Ankudinov, R. Ravel, J.J. Rehr, and S.D. Conradson, Phys. Rev. B 58, 7565 (2998). 28. S. D. Conradson, et al., Inorg. Chem. 42, (2003). 29. J. H. Shim, K. Haule, and G. Kotliar, Nature 513 (2007) 30. M. V. Ryzhkov, A. Mirmelstein, S. W. Yu,B. W. Chung,[c] and J. G. Tobin, International Journal of Quantum Chemistry 113, 1957 (2013). 31. M.V. Ryzhkov, A. Mirmelstein, B. Delley, S. W. Yu, B.W. Chung, J.G. Tobin, Journal of Electron Spectroscopy and Related Phenomena 194, 45 (2014). 32. B. R. Cooper et al., in Handbook of the Physics and Chemistry of the Actinides, edited by A. J. Freeman and G. H. Lander (Elsevier Science, Amsterdam, 1985), V ol J. G. Tobin, et al., Phys. Rev. B 92, (2015).