PacSurf2022 Session CF-MoE1: Charles S. Fadley Memorial Session I: Advances in X-ray Spectroscopy and Analysis

In 1988-89 the author had the great pleasure of doing postdoctoral work in the Fadley group. Experiencing the impressive work with an electron spectrometer with a multi-channel detector (VG ESCALAB5) [1] initiated decades of thinking about a further parallelization of photoelectron recording techniques. Today we find powerful implementations that extrapolate those early ideas that have resulted in orders of magnitude increased recording efficiency.

Thanks to the excellent time structure of modern photon sources such as lasers, HHG sources, Synchrotrons and FELs, time-of-flight (ToF) detectors can recorde a wide energy band in parallel. Additional full-field imaging of (k_x,k_y) momentum patterns has led to a novel ARPES technique, termed ToF momentum microscopy (ToF-MM). A fast delay-line detector records (x,y) position and ToF for each electron with count rates up to 8x10⁶ cps. ToF-MM enabled new types of experiments, such as fast tomographic-like Fermi surface and -velocity mapping [2].

The developments had a dramatic impact on experiments with very low signal intensities. In particular, spin-resolved experiments suffer from the low figure-of-merit of spin detectors. Significant advances have been made with the advent of multichannel spin recording [3]. Bulk-sensitive photoemission using hard X-ray excitation suffers from low cross sections and enhanced phonon scattering. The ToF-MM approach was extended to energies >7keV and an order of magnitude higher (E_B,k_x,k_y) phase-space acceptance. This led to new types of (H)ARPES [4] and X-ray photoelectron diffraction (XPD) experiments [5]. The large k-field-of-view provides 4D bulk spectral functions ρ(E_B,k) in a single measurement without scanning the photon energy.

In fs time-resolved photoemission (tr-ARPES), weak signals of transient excited states are recorded with high efficiency, both for FELs [6] and HHG sources [7]. In parallel to the tracking of the electronic structure, rapid changes in the geometric structure can be observed using tr-XPD. Resolving the intrinsic time scale of ultrafast processes and further parameters like pump-laser fluence coined the term “multidimensional photoemission” for such experiments. For further reading, see current reviews [4,5,7].

[1] J. Osterwalder, C. S. Fadley et al., J. Electron Spectrosc. Relat. Phenom. 48, 55 (1989); [2] K. Medjanik et al., Nat. Mat.16, 615 (2017) and J. Synchr. Rad. 26, 1996 (2019); [3] M. Kolbe et al., Phys. Rev. Lett. 107, 207601 (2011); [4] G. Schönhense, H.-J. Elmers, J.Vac. Sci. Techn. A 40, 020802 (2022); [5] O. Fedchenko et al., J. Phys. Soc. Jpn. 91, 091006 (2022); [6] F. Pressacco et al.. Nat. Commun. 12, 5088 (2021); [7] O. Karni, I. Esin, K. M. Dani, Advanced Materials (2022), doi:10.1002/adma.202204120in print.

We discuss the understanding of multiplet splitting and shake-up satellite structure for transition metal compounds. Chuck Fadley was an early pioneer for the former (1).

Gupta and Sen (2) expanded atomic multiplet splitting theory to include spin orbit coupling, providing predictions for the 2p XPS splitting features for TM cations. Such calculations do not include ligand-field or bonding effects in solids. Later semi-empirical approaches attempted to include such effects by scaling parameters and by invoking a charge transfer model to describe shake-up satellites. A free program, CTM4XAS (3), incorporates all this to simulate the cation 2p XPS spectrum.

Though CTM4XAS may roughly describe the BE spread in the cation 2p XPS, it completely disagrees with the description obtained from our full ab initio cluster calculations (Fe₂O₃ and NiO) of the electronic characteristics of the major features. The strong experimental satellite intensity for the Fe2p spectrum of Fe₂O₃ and other high spin Fe(III) compounds, at ~8.5 ev, has long been assigned to shake-up; and has been used to verify the presence of Fe(III) (4). Similarly, in the Ni2p spectrum of NiO, the satellite at ~6.5ev is used to verify the presence of Ni(II). Our calculations confirm that these satellites do have a large amount of shake character, and that the leading edge of the “main” peak (lowest BE) is essentially “normal”, ie it has no shake involvement (5, 6). CTM4XAS reverses these assignments, claiming that main peak is a shake state (charge transfer in their parlance) and the intense satellite represents the “normal” transition. Furthermore, our calculations show that once shake excitation is included, the intensity distribution of all structure in the 2p spectrum is highly modified, and there is no region, except the leading edge of the main 2p_3/2 peak, that is 100% "normal" (5). Everywhere else there is significant mixing of “normal” and shake. We expect this mixing to hold for other TM compounds.

References

1. C. S. Fadley, D. A. Shirley, A. J. Freeman, P. S. Bagus, and J. V. Mallow, Phys. Rev. Lett., 1969, 23, 1397-1401.
2. R. P. Gupta and S. K. Sen, Phys. Rev. B, 1974, 10, 71-77; P. Gupta and S. K. Sen, Phys. Rev. B, 1975, 12, 15-19.
3. F. De Groot and A. Kotani, Core level spectroscopy of solids. CRC Press, Boca Raton, 2008.
4. C. R. Brundle, T. J. Chuang, and K. Wandelt, Surface Sci. 48, 29 (1974).
5. P. S. Bagus, C. J. Nelin, C. R. Brundle, B. V. Crist, N. Lahiri, and K. M. Rosso, Phys. Chem. Chem. Phys., 2022, 24, 4562-4575.
6. P. S. Bagus, C. J. Nelin, C. R. Brundle, B. V. Crist, E. S. Ilton, N. Lahiri, and K. M. Rosso, Inorganic Chemistry in press.

Zinc, cadmium, and their related compounds have a myriad of applications, including, but not limited to, automotive manufacturing, batteries, photovoltaic devices, corrosion protection, and pigments. From a research perspective, the ability to accurately speciate different compounds is critical to understanding how surface chemistry can influence material properties. This ability allows for the advancement of our mechanistic understanding. It has been repeatedly demonstrated that X-ray photoelectron spectroscopy (XPS) is a powerful analytical tool for the study of changes in surface chemistry.Typically, speciation is achieved through the careful analysis of photoelectrons, and occasionally, Auger electrons. Since the main photoelectron binding energies of many zinc and cadmium compounds overlap, the modified Auger parameter must also be utilized to reliably detect chemical shifts. While the binding energies and Auger parameter data can be found in the literature for some zinc and cadmium compounds, in most cases, the available data lacks consistency, certainty, and reproducibility. Notably, Auger parameter data is often underreported in the existing literature, despite its proven utility. Additionally, many of these sources do not provide the variety of compounds needed to perform accurate chemical state analyses.

In this presentation, recent work to extend upon previously published XPS reference data and curve-fitting procedures will be detailed for a wide range of high-purity zinc- and cadmium-containing compounds. This will include a summary of literature data, carefully excluding those which contain issues related to reliability. Lastly, the applicability of peak fitting Auger lines will be explored for the analyses of mixed species systems.