Session Y22: First-Principles Modeling of Excited-State Phenomena in Materials VII: X-Ray Spectroscopy

Live

Many interesting properties of functional materials depend on their excited state properties, such as dynamic response and thermodynamic behavior. Often this behavior depends on details of excitations in the system such as phonons and plasmons, which lead to inelastic losses and damping effects. These excitations can be probed by photoemission spectra, where they show up as satellites beyond the quasi-particle peak [1]. These many-body effects are neither amenable to density functional theory nor extensions based on quasi-particle approximations. Here we discuss an approach based on the cumulant Green’s function, which provides a unified treatment of such dynamic correlation effects [2]. The approach is illustrated with several applications. Remarkably, a cumulant calculated in linear response within a quasi-boson approximation, is adequate to explain the loss spectra and charge-transfer excitations in many systems. Finite-temperature exchange-correlation potentials, thermodynamic properties, and the finite-temperatureTDDFT kernel, and can also be obtained [3,4]. Finally, some extensions are briefly discussed, including nonlinear contributions to the cumulant [5].
[1] Jianqiang Sky Zhou et al., PNAS, https://doi.org/10.1073/pnas.2012625117.
[2] L. Hedin, J. Phys.: Condens. Matter 11, R489 (1999).
[3] J. J. Kas, et al., Phys. Rev. B 100, 195144 (2019).
[4] John J. Rehr and Joshua J. Kas, Eur. Phys. J. B 91, 153 (2018).
[5] M. Tzavala et al., Phys. Rev. Research 2, 033147 (2020).   

Live

Velocity-gauge real-time TDDFT within the generalized Kohn-Sham (GKS) formulation provides a convenient framework for accessing both linear and nonlinear optical response in solids across near-infrared to soft X-ray energy ranges [1, 2]. In this work the GKS TDDFT description of low-order nonlinear response properties such as second harmonic generation and two-photon absorption is investigated in materials like hexagonal BN where excitonic effects are important. Additionally the relevance of this approach to the emerging field of nonlinear X-ray spectroscopy is discussed. Results from TDDFT are compared to benchmark many-body perturbation theory results or experimental data where available.

[1] C. D. Pemmaraju, Computational Condensed Matter 18, e00348 (2019)
[2] C. D. Pemmaraju, New J. Phys. 22 083063 (2020)   

Live

Resonant inelastic x-ray scattering (RIXS) is a powerful spectroscopic technique that offers an elemental- and orbital-selective probe of the electronic excitations over a huge energy range. In this talk, we present a novel many-body approach to determine RIXS spectra in solids, yielding an intuitive expression for the RIXS cross section in terms of pathways between intermediate many-body states containing a core hole, and final many-body states containing a valence hole. Explicit excited many-body states are obtained from the diagonalization of the Bethe-Salpeter equation in an all-electron framework. For the paradigmatic examples of the fluorine K edge of LiF and the carbon K edge in diamond, we show how the excitation pathways determine the spectral shape of the emission, and demonstrate the nontrivial role of electron-hole correlation in the RIXS spectra. Finally, we discuss for the example of Ga₂O₃ how RIXS can be employed to determine the nature of bound valence excitons and probes the valence hole distribution of the bound excitons in this material.   

Live

Oxygen K-edge x-ray absorption spectroscopy (XAS) provides an important local probe of the H-bond network in liquid water. Theoretical calculations of XAS spectra demand accurate modeling of both the molecular structure and the electron-hole excitation process. We generate the water structure from path-integral DeePMD calculations, whose neural network potential is trained on state-of-the-art DFT data at the SCAN0 hybrid functional level. Based on the above equilibrated trajectory, the XAS spectra of liquid water are computed by solving the Bethe-Salpeter equation (BSE) as implemented in the BerkeleyGW package. Our calculated XAS spectra of water agree well with experiments. Our results indicate that self-consistent GW quasiparticle wavefunctions and local-fields effects in electronic screening are crucial in the GW-BSE approach in order to yield XAS spectra that are in quantitative agreement with experiments.   

Live

Core level X-ray Photoelectron Spectroscopy (XPS) is one of the most widely used experimental techniques in the study of materials and molecules. The key quantity in the analysis of experimental spectra is the core electron binding energy. Measured binding energies allow the identification of the elemental composition of the probed region, and small shifts in binding energies can also be used to make inferences about the local chemical environments of the atoms. However, this information can be difficult to interpret, especially for complex systems.

First principles calculations provide an alternative way to link the core electron binding energy of an atom to its chemical environment. Previously, we have shown how Δ-Self-Consistent-Field (Δ-SCF) calculations can predict absolute core electron binding energies in small molecules with extremely hight accuracy (errors < 0.2 eV). We have also shown how this method can be extended to solids and surfaces via the use of cluster models. [Kahk and Lischner, Phys. Rev. Materials 3, 100801(R) (2019)]

In this talk, the prediction of core electron binding energies in periodic solids will be discussed further, and the finite cluster approach will be constrasted to the periodic supercell approach with representative examples.   

Live

We propose to use pump-probe x-ray absorption spectroscopy (XAS) as an ultrafast thermometer of the electronic subsystem, when it is driven out of equilibrium. The idea is inspired by the fact that the shape of the so-called satellite peak in the x-ray photoemission spectra (XPS) depends strongly on temperature in equilibrium. This satellite peak also demonstrates a strong dependence on time in time-resolved spectroscopy, and it is connected to the total energy of the system. Further, by comparing the energy gain during the nonequilibrium process to the equilibrium energy, allows one to measure in situ the effective temperature of the thermalized electron state. XAS is closely connected to XPS and might be an effective thermometer of the nonequilibrium system and it can be performed in many experiments where ultrafast x-rays are available.
We solve the XAS in the spinless Falicov-Kimball model within the dynamical mean-field theory, which gives an exact result. We examine metallic and Mott-insulator phases of the model.   

Live

Reliable methods to calculate Resonant Inelastic X-Ray Scattering (RIXS) spectra in strongly correlated many-body systems have been constrained to relatively small problems, as they rely on evaluating a response function (Kramers-Heisenberg formula) that is obtained from a time-dependent perturbative analysis of the scattering problem. This requires the knowledge of all eigenstates of the many-body Hamiltonian to account for intermediate processes involving excitations of core electrons. We introduce a novel approach recasting such calculation as an effective time-dependent problem in which incoming photons scatter with the system and outgoing photons are captured by a detector. The spectral density is obtained by solving the time-dependent Schroedinger equation explicitly involving the photon degree of freedom. We demonstrate the formalism with an application to Mott insulating Hubbard chains using the time-dependent density matrix renormalization group method. Our approach can readily be applied to systems out of equilibrium without modification and generalized to other spectroscopies.   

Live

Nonlinear X-ray spectroscopies such as resonant inelastic X-ray scattering (RIXS) are powerful probes to investigate the structure of matter and dynamics with unprecendented resolution, especially when augmented with reliable theoretical tools. However, modeling nonlinear X-ray spectra is challenging due to the difficulties in describing the physics of underlying core-electron processes. In particular, the presence of the valence ionization continuum in which core-level states are embedded renders standard theoretical methods, designed for modeling valence-electron processes, inadequate. The ionization continuum is also responsible for the non-convergent behavior of the response states in standard ab initio calculations for modeling nonlinear spectra. We will present a novel equation-of-motion coupled-cluster framework based on the core--valence separation scheme that provides convergent calculations of X-ray response calculations and facilitates robust modeling of RIXS spectra for both closed- and open-shell species. We will illustrate the capabilities of our novel approach by measuring its performance in modeling the RIXS spectrum of the transient aqueous OH radical formed in the ionization of water against experiments.   

Live

A recent study of energy transfer due to laser excitation in a Fe/MgO heterostructure showed that the dynamic behavior is dominated by phonon-mediated processes [1]. We present an analysis of the spectral changes in the transient x-ray absorption spectra (XAS) using a finite-temperature real-space Green’s function approach. We approximate the phonon contribution via a harmonic dynamical matrix sampling method. Our theory accounts for the spectral differences shortly after laser excitation and at long times.   

Not Participating

Materials based on the Zn-paratacamite family have generated intense research efforts because of their spin-½ frustrated magnetism that may exhibit a quantum spin liquid (QSL) ground state. Intrinsic disorder in a potential QSL candidate must be characterized as this can introduce unwanted interactions that destroy QSL physics. Using first-principles calculations combined with experimental X-ray absorption measurements, we focus on characterizing the local and long-range structures of two leading QSL candidates: herbertsmithite (Cu₃Zn_0.85Cu_0.15(OH)₆Cl₂) and Zn-substituted barlowite (Cu₃Zn_xCu_1-x(OH)₆FBr). Our results indicate that Zn-barlowite is more resistant to antisite disorder compared to herbertsmithite while also providing additional structural advantages important for QSL.   

Live

The cumulant approximation to the one electron Green's function has recently gained attention as an improved method for including the effects of many-body excitations in calculations of x-ray spectra [1]. Multiple plasmon satellites in the XPS spectra of solid-state systems are well reproduced by the cumulant, but not the GW approximation. Approximations of the cumulant are usually based on an expansion to linear order in the screened coulomb interaction, but this may become inadequate at higher correlation. Recently, a new approach for including non-linear contributions was presented, and a simple approximation was proposed based on real-time TDDFT [2]. Here we apply this new approximation and compare to the linear response form as well as to available experimental data. Extensions of the approximation are also briefly discussed.

[1] Jianqiang Sky Zhou et al., J. Chem. Phys. 143, 194109 (2015).
[2] M. Tzavala et al., Phys. Rev. Research 2, 033147 (2020).   

Live

GW has become the method of choice for the calculation of valence photoemission spectra [1]. To apply GW also to deep core excitations as measured by X-ray photoelectron spectroscopy, we recently advanced the GW methodology and our implementation by combining exact numeric algorithms in the real frequency domain [2] with partial self-consistency [3] and relativistic corrections [3,4]. We benchmarked our core-level GW implementation for 65 1s core excitations, for which we find that GW reproduces absolute molecular 1s excitations within 0.3 eV of experiment and relative binding energies with average deviations smaller than 0.2 eV [3]. We then combined GW with machine learning (ML). We computed 16,000 C1s and O1s excitations of organic molecules with GW and used them to train a Kernel Ridge Regression (KRR) ML model. The KRR-ML model predicts molecular core-level energies within 0.1 eV of the GW reference data. We used these models as starting point to develop GW-ML schemes for amorphous carbon materials, which show promise as electrode material for biomedical devices.

[1] D. Golze et. al. Front. Chem, 7 (2019), 377
[2] D. Golze et. al. JCTC, 14 (2018), 4856
[3] D. Golze et al. JPCL, 11 (2020), 1840
[4] L. Keller et al. JCP, 153 (2020), 114110   

Live

UV pump-XUV probe measurements have been successfully applied in the study of photo-induced chemical reactions. Although rich element-specific electronic structure information is accessible within XUV (inner-shell) absorption spectra, it can be difficult to interpret the chemistry directly from the spectrum. Due to the multi-reference character of the excited states of heavier alkyl halide molecules, we developed and applied a multi-reference method to completely simulate UV pump-XUV probe measurements to study photodissociation in alkyl halides molecules. Fewest switches surface hopping (FSSH) trajectories were used to explore the coupled electronic and ionic dynamics upon photoexcitation. Interpretation of previous measurements is provided by associated multi-reference, restricted active space, inner-shell spectral simulations. This combination of FSSH trajectories and XUV spectra provides an interpretation of experimental transient features and validates the branching ratio between different spin-orbit split states in the photodissociated products. This methodology should prove useful for interpretation of the increasing number of inner-shell probe studies of molecular excited states or for designing new experiments to control the direction of chemical reactions.