betaFeedback

WorksPhysics and Astronomy

Ultrafast Hole Transport and Electron Localization in Superatomic Re6_{6}Se8_{8}Cl2_{2}

Abstract

Van der Waals superatomic crystals represent an intriguing class of materials, exhibiting unique ultrafast electronic and structural responses arising from their hierarchical crystal structure. These materials bridge the properties of semiconductors, molecular solids, and nanocrystal arrays, hosting a range of many-body interactions that can be controlled with atomically precise structures. To establish the structure-function correlation, it is necessary to understand element-specific electronic states arising from complex superatomic unit cells. Here, we use transient XUV spectroscopy to individually probe the dynamics of photoexcited electrons and holes in Re6_{6}Se8_{8}Cl2_{2} superatomic crystals. We simultaneously measure photodynamics at the Re O2_{2} and Se M4,5_{4,5} edges, enabling decomposition of excited-state dynamics into the contributions of electron, hole, band gap, and phonon dynamics occurring around Re and Se atoms. By correlating spectral features at the Re and Se edges, we infer the existence of flat conduction bands and identify hot hole transfer from Re to Se atoms. We observe phonon oscillations driven by hole-mediated displacive excitations. The measurements highlight that the photodynamics are dominated by charge transfer between metallic sites which drives phonon coupling in the nonequilibrium state. This work provides insights into the optoelectronic applications of superatomic crystals for future tuning of emergent material properties.

Introduction

The material properties and functions in hierarchical structures are an emerging field in materials science.[1, 2, 3] While structure-function relationships are well understood in many cases, they are challenging for materials with complex chemical compositions and hierarchical structures due to the correlated behavior of electrons and phonons across different length scales.[4] Two-dimensional van der Waals (vdW) superatomic crystals are emerging materials that have this challenge. Unlike conventional crystals, superatomic crystals use atomically precise clusters as building blocks that form long-range crystalline order. The materials are composed of discrete, modular clusters that mirror traditional atomic solids.[1, 2, 5, 6] The macroscopic properties of superatomic crystals can be widely tuned by modifications of the chemical composition [7, 8, 9] or atomic coordination of cluster building blocks,[10] leading to complex many-body interactions and rich dynamics.

Meanwhile, superatomic crystals have shown promising applications in optoelectronics,[11] thermoelectricity,[12, 13, 14] and catalysis.15 For example, superatomic materials have demonstrated significant thermoelectric performance without systematic material optimization.16 Recent optical studies reveal efficient quasi-ballistic exciton transport in superatomic crystals up to room temperature, with exciton mobility exceeding Si, making them promising candidates for efficient optoelectronic devices.17 Other optical studies have provided insight into charge density waves, coherent exciton transport, carrier, and phonon dynamics in Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2}.[6] [17, 18, 19] Further, tuning of electron and phonon bandwidths[1] and superconductivity[9] [20, 21] have been demonstrated.

The observed material properties are largely due to the intrinsically strong electron–phonon interactions in superatomic crystals.22,23 Superatomic crystals have a complex unit cell and hierarchical lattice structure, enabling these unique phonon dynamics. Typical phonon spectra consist of both localized molecular vibrations and long-range coherent phonons expected for crystals.24,25 Meanwhile, electronic excitations create molecular-like excitons that exhibit large binding energies. Despite its fundamental importance, understanding excited-state carrier–phonon dynamics in superatomic crystals is challenging due to the interplay between the complex unit-cell structure, electronic band structures, and various vibrational modes.

In this work, we measure the carrier- and element-resolved ultrafast electronic dynamics in the vdW superatomic semiconductor Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2} (Figure 1a), using transient extreme ultraviolet (XUV) reflection–absorption spectroscopy. In our measurements, carrier- and element-specific relaxation processes following photoexcitation are separately resolved. We measure the thermalization and charge redistribution of holes from Re to Se sites and confirm the existence of flat conduction bands (CBs) through minimal electron thermalization dynamics. Hole-mediated coherent cluster-tilt displacive phonon motions are observed in the transient spectra, supporting carrier-phonon coupling within the two-dimensional layers of Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2}. The element-specific observation of flat CBs and the displacive excitation of coherent phonon motions provides insights into the nature of enhanced thermal transport properties within this class of superatomic crystals.17,25,26 Our observations provide a guide for designing chemical structures and atomically-precise spatial layouts in these hierarchical superatomic crystals, which can potentially enhance carrier transport from the nanoscale to mesoscopic levels.

Re₆Se₈Cl₂ crystal and electronic structure

Figure 1. Re6_{6}Se8_{8}Cl2_{2} crystal and electronic structure. a, Crystal structure of the Re6_{6}Se8_{8}Cl2_{2} single crystal projected along the aa-axis. The broadband XUV probe can selectively measure the transient dynamics of charge carriers at both the Re O2_{2} and Se M4,5_{4,5} edges following photoexcitation. b, Calculated projected electronic density of states (PDOS). The valence states are composed of Re 6d6d, Se 4p4p, and Cl 3p3p orbitals. Core electrons in Re 5p5p and Se 3d3d orbitals are excited to these states by the XUV probe pulses. The calculation includes the spin-orbit splitting of the core orbitals. EFE_{\mathrm{F}} is Fermi energy and valence states are scaled by a factor of 9 for better visibility. c, Calculated atom-projected band structure of Re6_{6}Se8_{8}Cl2_{2} along the high-symmetry kk-path. The atomic contributions to the electronic structure indicated by a color gradient from red to blue, representing Se to Re, respectively. The CBs are predominantly composed of Re 6d6d orbitals, while the VBs are hybridized with Re 6d6d and Se 4p4p orbitals.

Results and Discussion

The layered structure of Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2} is shown in Figure 1a. Each layer in this superatomic semiconductor is composed of covalently bonded Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2} clusters arranged in a pseudo-square lattice within the aa–bb plane. These two-dimensional layers are stacked along the cc–axis through vdW interactions, giving rise to a hierarchical crystal structure.[5] This hierarchical structure leads to unusual electron-phonon coupling,17,18 primarily due to the interplay between bound excitons, localized molecular dynamics, and the long-range collective response. The calculated projected density of states (PDOS) using density functional theory (DFT) predicts an indirect band gap (EgE_{\mathrm{g}}) of 1.2 eV as shown in Figure 1b, similar to measured and reported values (see the Supplementary Information for details of DFT calculations).[5] The upper valence bands (VBs) have strong hybridization between the Re 6d6d and Se 4p4p orbitals, with minor contributions from Cl 3p3p orbitals. The atomic orbital-resolved electronic band structure, shown in Figure 1c, shows that Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2} has an indirect band gap between the valence band maximum (VBM) at the Γ\Gamma point and the conduction band minimum (CBM) at the T point.24

Transient XUV reflection-absorption spectroscopy measures element-specific core-to-valence transitions and can separate electron and hole dynamics at individual elemental X-ray edges.27 In our experiment, a 3.1 eV (400 nm) excitation pulse generates electron-hole pairs across the band gap of single crystalline Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2}. The excitation fluence used in transient XUV experiments was ∼9\sim9 mJ/cm2^{2}, resulting in an excited carrier density of ∼4.1×1021\sim4.1 \times10^{21} cm−3^{-3}. The resulting carrier dynamics are probed by a broadband XUV pulse in a 10∘10^{\circ} grazing incidence reflection geometry. The XUV pulse simultaneously measures the Re O2\mathrm{O}_{2} and Se M4,5\mathrm{M}_{4,5} edges, revealing the local chemical environment around Re and Se sites, as shown in Figure 2 and Figure 3, respectively. More details of the experimental configuration can be found in the Supplementary Information.

Carrier relaxation dynamics at the Re O$_2$ edge

Figure 2. Carrier relaxation dynamics at the Re O2_2 edge. a, Transient XUV reflection-absorption spectra at the Re O2_2 edge. Three main spectral features for the valence bands (VB), band gap renormalization (BGR), and conduction bands (CB) are denoted as dotted lines. b, Experimental and BSE theory calculated transient differential absorption spectra, shown as solid and dotted lines, respectively, at two time delays (0.07 and 1.7 ps) following photoexcitation. The BSE calculated spectra model an initial charge transfer state with electrons filled to the same energy (2.0 eV) above the VBM in both time points. In comparison, holes are filled up to 3.1 eV below the CBM (0.07 ps, purple dashed line) and 1.5 eV below the CBM (1.7 ps, orange dashed line), suggesting strong hole thermalization in the VBs with very minimal electron thermalization. c, BSE theory calculated Re O2_2 edge spectra for holes only (top) and electrons only (bottom). The BSE modeling of holes and electrons uses a state-filling model in which holes are modeled from the VBM down to energies as low as 3.1 eV below the CBM, to consider the possible allowed transitions following the 3.1 eV pump. Similarly, electrons are modeled starting at the CBM and up to 3.1 eV above the VBM. The modeled hole energies in c correspond to hole cutoff energies of 3.1 eV (purple), 2.5 eV (yellow), 2.0 eV (orange), and 1.5 eV (blue). The modeled electron energies correspond to electron cutoff energies of 3.1 eV (blue), 2.5 eV (orange), 2.0 eV (yellow), and 1.5 eV (purple). d, Kinetic trace of hot hole thermalization at the Re O2_2 edge averaged across ∼\sim44–45 eV. The solid line indicates a fitted curve. e, Re projected band structure of Re6_6Se8_8Cl2_2. The colored arrows (right) depict the corresponding state-filling electron and hole energies that are modeled in c and are color coded accordingly.

Carrier relaxation dynamics in Se

Figure 3. Carrier relaxation dynamics in Se. a, Transient XUV reflection-absorption spectra at the Se M4,5_{4,5} edge. b, Experimental and BSE calculated transient differential absorption spectra, shown as solid and dashed lines, respectively, at two time delays (Δt\Delta t = 0.07 and 2.3 ps) following photoexcitation. The Se 3d5/2_{5/2} and 3d3/2_{3/2} transitions are depicted as vertical dashed black lines, with one pair denoting transitions to the VBs and the other pair denoting CB transitions. c, As in Figure 2c, BSE theory calculated spectra

We first focus on understanding the photodynamics occurring at the Re sites. Figure 2a shows the transient XUV reflection-absorption at the Re O2_2 edge. The reflection-absorption is defined as ΔOD=−log⁡10(Ipump onIpump off)\Delta\mathrm{OD}=-\log_{10}\left(\frac{I_{\mathrm{pump\ on}}}{I_{\mathrm{pump\ off}}}\right). In this definition, a positive signal (Figure 2a, yellow) represents a transient decrease in XUV reflectivity, which corresponds to an increase in XUV absorption. The two transient spectral lineouts at Δt=0.07\Delta t=0.07 and 1.7 ps show a clear difference between the initial and final excited states (Figure 2b). The positive signal at ∼\sim45 eV shifts to higher energies (blueshift) with concomitant decreases in amplitude. The increase in absorption around ∼\sim46.5 eV also blueshifts but with a lesser magnitude than the peak at ∼\sim45 eV. There is a decrease in absorption (Figure 2a, blue) that is centered around ∼\sim47.8 eV and does not shift spectrally or experience a decrease in amplitude.

To understand the origin of the spectral features at the Re O2\mathrm{O}_2 and Se M4,5\mathrm{M}_{4,5} edges of Re6Se8Cl2\mathrm{Re}_6\mathrm{Se}_8\mathrm{Cl}_2, we use a density functional theory (DFT) and Bethe–Salpeter equation (BSE) approach to model the excited state X-ray edge dynamics. The method uses Quantum ESPRESSO28,29 and an excited state modification to the OCEAN (Obtaining Core Excitations from the electronic structure and the NIST BSE solver) code.30–33 The theoretical methods applied here are discussed in more detail in the Supplementary Note S4. Briefly, to model carrier thermalization, we define two state-filling cutoff energies, EhE_\mathrm{h} and EeE_\mathrm{e}. Where EhE_\mathrm{h} is the maximum hole depth in the VBs referenced from the CMB to encompass all possible transitions in momentum space by the pump and EeE_\mathrm{e} is the maximum electron height in the CBs referenced from the VBM. In the case of our 3.1 eV excitation employed in the transient XUV experiments, this means that EeE_\mathrm{e}, Eh≤3.1E_\mathrm{h} \le3.1 eV, and smaller values correspond to carrier thermalization toward the respective band edges of the carriers. EhE_\mathrm{h} and EeE_\mathrm{e} are phenomenological state-filling cutoff energies that parameterize the modeled spectral sensitivity to hole and electron occupations. While these cutoffs describe the bounds of the pump-accessible excitations, they do not imply that a single photoexcited electron–hole pair simultaneously occupies both cutoffs. Additionally, the BSE state-filling model considers dynamically blocking (electrons) and allowing (holes) core-level transitions and does not consider carrier density. More details on these cutoff energies can be found in the Supplementary Note S4.B.

Figure 2b shows experimental and BSE modeled lineouts for two different time delays in the transient spectra. We model the state immediately following photoexcitation (Figure 2b, purple dashed line) based upon an electron cutoff energy of 2.0 eV and a hole cutoff energy of 3.1 eV. This suggests that the increased absorption between ~44–46.5 eV is primarily due to the photoexcited hole population in the VBs. The increased absorption is due to this state-filling effect, where photoinjected holes increase the available density of states in the VBs, leading to enhanced XUV absorption.34 Similarly, the decrease in absorption from ~46.5–49 eV is well described by photoexcited electrons occupying CBs and blocking XUV transitions, leading to decreased XUV absorption. This theoretical assignment agrees well with experimental spectra immediately following photoexcitation, suggesting that transitions are being probed from deep VBs into CBs near the CBM.

The orange dashed line in Figure 2b well describes the transient spectra at later times (1.7 ps). This lineout was modeled with a state-blocking EeE_\mathrm{e} and EhE_\mathrm{h} of 2.0 eV. This suggests that there is significant thermalization of holes at the Re O2O_{2} edge following photoexcitation, but minimal thermalization of electrons. Further, the blueshift observed experimentally at ∼45\sim45 eV in Figure 2a is well described by BSE theory modeling hole thermalization at the Re O2O_{2} edge, with no electron thermalization modeled. The spectral shift associated with hole thermalization in the VBs occurs on a timescale of 0.8 ±\pm 0.2 ps, with an energy shift of ∼0.4\sim0.4 eV (Figure 2d).

The photoinduced changes at the Re O2O_{2} edge can be understood as a mixed contribution from photoexcited electrons, holes, and band gap dynamics. To quantify their relative contribution, we apply a singular value decomposition (SVD), described in more detail in Supplementary Note S3. The first component is best described by hole thermalization in the VBs, as demonstrated by our BSE theory in Figure 2c (top). The second and third components are well described by band gap renormalization (BGR) and absorption by electrons in the CBs. In addition to photoexcited electron and hole dynamics, SVD reveals that BGR is observed around ∼46.5\sim46.5 eV. Microscopically, photoexcitation modifies the Coulomb screening, reducing the band gap.35 The reduction of the minimum band gap enables XUV transition into the otherwise forbidden gap region, causing a red shift of the XUV transition to the CB immediately following photoexcitation.

for holes (top) and electrons (bottom) uses a state-filling model in which Eh≤3.1E_{\mathrm{h}} \le3.1 eV, and Ee≤3.1E_{\mathrm{e}} \le3.1 eV. The modeled holes correspond to holes filling 3.1 eV (purple), 2.5 eV (yellow), 2.0 eV (orange), and 1.5 eV (blue) below the CBM. The modeled electrons correspond to electrons filling to 3.1 eV (blue), 2.5 eV (orange), 2.0 eV (yellow), and 1.5 eV (purple) above the VBM. d, Kinetic profiles for thermalization of holes and electrons at Se sites. The feature at ∼56.5\sim56.5 eV attributed to a Se 3d5/23d_{5/2} transition to the VB has a time constant of 0.9±0.20.9 \pm0.2 ps (red). The feature at ∼57.5\sim57.5 eV associated with a Se 3d3/23d_{3/2} transition to the VB increases in intensity with a time constant of 0.7±0.30.7 \pm0.3 ps (yellow). The electrons in the CB at ∼59.3\sim59.3 eV have a thermalization time constant of 0.8±0.40.8 \pm0.4. e, Atom-projected band structure for Se atoms where the color scale of the band structure indicates the PDOS of Se and the colored arrows (right) depict the corresponding state-filling electron and hole energies that are modeled in c.

We performed similar transient XUV reflectance measurements at the Se M4,5_{4,5} edges around ∼57\sim57 eV (Figure 3a) and find that the spectral response at the Se M4,5_{4,5} edge is qualitatively different from the Re O2_2 edge. Immediately after photoexcitation, there is a transient increase in XUV absorption around ∼56.5\sim56.5 eV, and two transient decreases in absorption are present at ∼58.1\sim58.1 eV and ∼59.3\sim59.3 eV, respectively (Figure 3b, yellow solid line). At longer pump-probe time delays, the reduced absorption at ∼58.1\sim58.1 eV disappears, and an increased absorption feature appears at ∼57.5\sim57.5 eV (Figure 3b, blue solid line).

The transient XUV spectra at the Se M4,5_{4,5} edges reflect local carrier densities around Se atoms and can be understood through photoexcited state-filling effects.33,34 The Se M4,5_{4,5} edge consists of core-to-valence transitions from two spin-orbit separated Se 3 core levels, 3d3/23d_{3/2} and 3d5/23d_{5/2}, with ∼1\sim1 eV energy spacing.36 We assign the transient reflectance changes at ∼57.5\sim57.5 eV and ∼59.3\sim59.3 eV to an XUV transition from the Se 3d3/23d_{3/2} core-level to the VBs and CBs, respectively. Similarly, spectral changes at ∼56.5\sim56.5 eV and ∼58.1\sim58.1 eV are assigned to XUV transitions from the Se 3d5/23d_{5/2} core level. The ∼1\sim1 eV spacing of the spin-orbit transitions is consistent with our measured and BSE calculated ground-state core-level absorption spectra shown in Figure S3 in the Supplementary Information.

Ground state reflectivity at the Re O2 and Se M4,5 edges

Figure S3. Ground state reflectivity at the (a) Re O2_2 and (b) the Se M4,5_{4,5} edges. BSE theory of the edges (blue) is compared to experimental static XUV reflectivity (orange) for good agreement.

We further deconvolve the transient dynamics of Re6Se8Cl2\mathrm{Re}_6\mathrm{Se}_8\mathrm{Cl}_2 at the Se M4,5_{4,5} edges through BSE modeling of photoexcited thermalization dynamics. Figure 3b compares BSE modeled transient dynamics and experimental lineouts immediately after photoexcitation (0.07 ps) and following thermalization (2.3 ps). The modeled lineout immediately following photoexcitation is based upon state-filling of electrons up to 3.1 eV above the VBM. In comparison, a hole cutoff energy of 2.5 eV best agrees with the experimental lineout at 0.07 ps (Figure 3b, yellow dashed line). For lineouts following carrier thermalization (Figure 3b, blue dashed line), BSE state-filling of an electron cutoff energy of 3.1 eV and a hole cutoff energy of 1.5 eV best agree with experiment. This suggests that, like carriers at the Re O2_2 edge, there is little to no electron thermalization, while the holes do thermalize. The lack of a clear Se M4,5_{4,5} edge thermalization signal and our BSE modeling indicates that electronic thermalization is not readily resolved at the edge. This may arise from electron redistribution within relatively flat CBs of similar Se character that can only be weakly distinguished at the Se edge. Further, this suggests the Se M4,5_{4,5} edge’s is sensitive to different CBs than the Re edge, as BSE state-filling of electrons at the Re edge agree better with spectra at lower electron cutoff energies (Figure 3c). We fit the thermalization of holes at the Se M5_5 edge, an increase in hole signal at the Se M4_4 edge, and a decrease in electron signal at the M4_4 edge in Figure 3d.

The element-specific dynamics at Re and Se atoms can be further used to infer the distribution of photoexcited carriers in the excited state. As shown in Figure 2 and Figure 3, photoexcited electrons are distributed on both Re and Se atoms. Meanwhile, based on the atomic projected band structure (Figure 1c), transient spectra (Figure 2a and Figure 3a), and BSE calculations (Figure 2b and Figure 3b), we observe that the hole states have strong Re orbital contributions, particularly near the top of the valence bands. This observation is consistent with our experimental findings that hot holes occur predominantly at the Re sites after photoexcitation and our BSE theory that requires 0.6 eV higher hole cutoff energies on Re sites than Se sites to agree with experiment. As hot holes thermalize at the Re edge, a spectral crossing appears at the Se edge from ∼57\sim57–5858 eV (Figure 3a), where the initial electron spectral weight (blue) after photoexcitation is transferred to holes (yellow) at later times.

This spectral crossing at Se edge could be intuitively understood as a transient increase in hole concentration around the Se atoms during carrier diffusion. Schematically, during the hole thermalization, hot holes migrate from the initial photoexcited states towards the Γ\Gamma point at the top of the VB, which has increasing contributions from Se atoms (Figure 1c). The net effect of hot hole thermalization is equivalent to hot hole transfer from Re to Se atoms in the unit cell. Given that the spectral crossing timescale at Se sites is comparable to the hole thermalization time at the Re edge (Figure 2d and Figure 3d), this suggests that the hole transfer from Re to Se atoms slows down overall hole diffusion at Se sites, which leads to the observed spectral crossing at the Se edge.

We measure phonon oscillations in our transient XUV absorption spectra at the Re O2_2 edge (Figure 2a). The kinetic profile and a fit curve display a time-dependent modulation of intensity with a periodicity of 270 fs (3.6 THz) (Figure 4a). The periodicity aligns with a coherent phonon motion associated with the twisting of superatomic clusters in an Re6_6Se8_8Cl2_2 layer, primarily driven by the displacement of the corner Se atoms (Figure 4b and Figure 4c).18 These two-dimensional phonon modes, fall within the frequency range of 1.69–5.25 THz and can be linked to the electronic modulation of free charge carriers. The oscillation has a cosine character at t0 that suggests the mechanism of this oscillation is due to displacive excitation of coherent phonons (DECP) rather than impulsive stimulated Raman scattering (ISRS).18,37,38 Specifically, this coherent oscillation is only present in the positive absorption signal at the Re O2_2 edge, suggesting that DECP is hole-mediated.39

Ultrafast hole thermalization, diffusion, and phonon oscillations in Re6Se8Cl2

Figure 4. Ultrafast hole thermalization, diffusion, and phonon oscillations in Re6_6Se8_8Cl2_2. a, Kinetic profile of intensity oscillation averaged across ∼45.9 – 46.3 eV, fitted with a sine function (dashed red line). The oscillation period is measured to be 270 fs, which corresponds to 3.6 THz as shown in panel b. b, Fast Fourier transform (FFT) of the kinetic profile in panel a. The highest amplitude oscillation frequency is indicated with an asterisk. c, Overall schematics for the photoinduced electronic dynamics in Re6_6Se8_8Cl2_2. The atomic view of the time-evolution of the photoexcited states (top), two-dimensional view of the clusters (middle), and charge carrier dynamics depicted onto a simplified band structure (bottom). The element- and carrier-sensitive XUV probe enables elucidation of the initial charge carrier distribution in the VB and CB, and their relaxation dynamics with coupling to displacive phonons. The cluster tilt phonon mode accompanies vibrations of the Se atoms at the corner, leading to the in-plane distortion of a cluster.

Mechanistically, this can be understood as photoexcited holes transporting through metallic sites from Re to Se atoms in Re6_6Se8_8Cl2_2 single crystals and the thermalization and the associated intravalley charge redistribution dynamics across different momenta mediate coherent phonon oscillations, facilitating energy dissipation.

Notably, hole thermalization is observed at both the Re O2_2 and Se M4,5_{4,5} edges in the more dispersive VBs, while electron thermalization is minimal in the flat CBs. Our consensus is that the electrons experience very little relative thermalization in energy compared to holes in the Re6_6Se8_8Cl2_2 superatomic crystal, suggesting that the VBs are much more dispersive, with more delocalized electronic states than the CBs. We further find that hole transport from Re to Se atoms displacively excites coherent cluster twisting phonons, which facilitate energy transfer towards adjacent interclusters within 2D metal layers.

In summary, we report transient XUV reflection–absorption spectroscopy on the photoexcited dynamics in Re6_6Se8_8Cl2_2 superatomic crystals. Photoinduced electron, hole, phonon, and band gap dynamics are observed and separated at both Re and Se sites. By correlating the ultrafast dynamics at Re and Se atoms, we infer the existence of flat conduction bands and identify an ultrafast hole transfer from Re to Se atoms. Our results provide a detailed picture of element- and carrier-specific dynamics in superatomic crystals, as well as new insights into the real-space charge transfer processes in Re6_6Se8_8Cl2_2, paving the way for a deeper understanding of more complex many-body interactions in superatomic crystals. These observations underscore the impact of nonequilibrium many-body effects and enhance the understanding of photoexcited electron and hole dynamics across different elements in superatomic crystals. These insights are vital for advancing the application of novel photophysics and for driving the development of exotic phases, such as doping-induced superconductivity,[9] in superatomic crystals.

Methods

Details about sample preparation and characterization, the transient XUV spectrometer, spectral analysis, and theoretical modeling of excited state spectra using DFT and the BSE can be found in the Supplementary Information.

Acknowledgements

Work on superatomic materials at Columbia University is supported by the Air Force Office of Scientific Research (AFOSR) under MURI grant number FA9550-25-1-0288. The computations presented here were conducted in the Resnick High Performance Computing Center, a facility supported by the Resnick Sustainability Institute at the California Institute of Technology. This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 using NERSC award BES-ERCAP0024109. Y.-J.K. acknowledges support by the Liquid Sunlight Alliance (the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, Fuels from Sunlight Hub under award number DE-SC0021266). J.L.M. acknowledges support by the National Science Foundation Graduate Research Fellowship Program under grant no. 1745301. H.L. acknowledges support from the National Science Foundation under award number 2325212.

*Correspondence and requests for materials should be addressed to H. L. and S.K.C.

(email: [email protected], [email protected])

Re6_6Se8_8Cl2_2 sample preparation

Re6_6Se8_8Cl2_2 crystals were synthesized via chemical vapor transport adapted from previous reports.[1, 2] Re (330330 mg, 1.771.77 mmol), Se (190190 mg, 2.412.41 mmol) and ReCl5_5 (200200 mg, 0.550.55 mmol) were ground and mixed using an agate mortar inside the glove box, and subsequently pressed into a pellet. The pellet was sealed in a 2020 cm quartz tube under vacuum and heated in a tube furnace at a rate of 1 ∘C1\,^\circ\mathrm{C}/min to 1100 ∘C1100\,^\circ\mathrm{C} for 3 days. To achieve large single crystals, the tube was then cooled over six hours to a temperature gradient of 11001100–1030 ∘C1030\,^\circ\mathrm{C} and held for 270 hours. The tube was further cooled to room temperature over 12 hours, and the furnace was then shut off. To extract pristine Re6_6Se8_8Cl2_2 single crystals, a gradient of 300300–25 ∘C25\,^\circ\mathrm{C} was used to separate any excess volatile components, yielding ∼1\sim1–33 mm single crystals at the top section of the tube. Typical bulk samples have a few millimeters in lateral dimensions. Figure S1 shows a picture of a typical sample piece used in the transient XUV experiments.

A typical mm-size single crystalline bulk Re6Se8Cl2 sample

Figure S1. A typical mm-size single crystalline bulk Re6_6Se8_8Cl2_2 sample used in our transient XUV measurements.

Transient extreme ultraviolet spectrometer

The transient extreme ultraviolet (XUV) reflection-absorption spectrometer is described previously.[3, 4, 5] Briefly, a Legend Elite Duo Ti:Sapphire system (Coherent Inc.) with 1 kHz, 35 fs, 13 mJ pulses centered around 800 nm is split by a 75:25 beam splitter. ∼3\sim3 mJ generates a few-cycle white light beam (<6<6 fs, 550−$950-\$950) for high harmonic generation (HHG) and ∼10\sim10 mJ is used for the pump path. The pump path employs the 400 nm frequency doubled output of a BBO crystal with p-polarization, pumped with the 800 nm output from the Ti:Sapphire laser for an above band gap photoexcitation of the Re6_6Se8_8Cl2_2 single crystals. The optical excitation fluence used in these experiments was ∼9\sim9 mJ/cm2^2, resulting in an excited carrier density of ∼4.1×1021\sim4.1 \times10^{21} cm−3^{-3}.[6]

Transient reflection is measured at a 10∘10^\circ grazing incidence (80∘80^\circ from normal incidence) geometry and measures the varying delay times between the pump and probe pulses using an optomechanical delay stage. The photoexcited dynamics were probed with an XUV pulse produced through HHG in argon with an s-polarized few-cycle white light pulse driving HHG. The fundamental white light beam is removed with a 100 nm thick Al filter (Luxel). The generated XUV continuum is used to probe the Re O2_2 and Se M4,5_{4,5} absorption edges between 40–60 eV. An edge-pixel referencing scheme uses signal free spectral regions to denoise intensity fluctuations in the spectra.[7]

Spectral analysis

The transient XUV spectra at the Re O2_2 edge consists of contributions from electrons, holes, and band shifts. We apply singular value decomposition (SVD) to all pump-probe delays within the transient spectrum to extract the weights of relevant dynamics and reconstruct the transient spectrum at the Re O2_2 edge using the fewest number of singular values. The SVD analysis factorizes the two-dimensional matrix of the transient spectrum by rotating (U), scaling (S), and then again rotating (V) the matrix. The scaling matrix contains the singular values along its diagonal, which we refer to as the number of components. The number of singular values or components in the scaling matrix is selected to minimize the total number of singular values, where at the Re O2_2 edge three singular values are needed.

Following SVD, we ascribe component 1 to holes, component 2 to electrons, and component 3 to band gap renormalization (BGR). Figure S2 shows the components of the SVD analysis. Component 1 well describes excited state hole effects, attributed to an increase in absorption at the edge. Due to the minimal band gap renormalization shift, and minimal thermalization in the electron component with respect to time, components 2 and 3 are combined in a weighted average to visualize not only the spectral shift associated with BGR but also minimal electron thermalization. It is worth noting that the lineouts in component 1 are an order of magnitude more intense than those in the weighted average of components 2 and 3. The reconstructed spectra in Figure S2 is well described by three components, where the band gap renormalization shift is not observed with less than three components.

SVD components and reconstructed data at the Re O$_2$ edge

Figure S2. Component 1 (a) of the SVD analysis at the Re O2_2 edge. Components 2 and 3 (b) are plotted as a weighted average as they are an order of magnitude weaker than component 1. The SVD reconstruction (c) agrees with transient XUV spectra of the Re O2_2 edge shown in Figure 2a in the main text.

Ab initio modeling

Density functional theory

Geometry optimization and density functional theory (DFT) calculations were conducted using the Quantum ESPRESSO package with norm-conserving general gradient approximation (GGA), Perdew–Burke–Ernzerhof (PBE) pseudopotentials.[8, 9, 10] Geometry optimizations employed an input Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2} unit cell containing 16 atoms and were based upon the lattice parameters a=b=c=6.67a=b=c=6.67 Å. A 300 Rydberg kinetic energy cutoff and an 8×8×88 \times8 \times8 Monkhorst–Pack k-mesh sampling of the first Brillouin zone was used to sample 250 bands. The structures employed in the DFT and BSE calculations described below were fully relaxed and calculations were performed to a convergence of 10−810^{-8} eV/atom with the forces on ions under 10−310^{-3} eV/Å.

Bethe–Salpeter equation theory of core-level spectra

The fully relaxed parameters used in the DFT calculations discussed in Supplementary note 4.A were applied to the BSE calculations of the Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2} ground state reflectivity at the Re O2_2 and Se M4,5_{4,5} edges. The method uses Quantum ESPRESSO[9, 10] and an excited state modification to the OCEAN (Obtaining Core Excitations from the Ab initio electronic structure and the NIST BSE solver) code.[11, 12, 13] The BSE calculation employed a screening mesh of 2×2×22 \times2 \times2, a dielectric constant of 21.8 at 293 K, a 4.0 Bohr screening radius, and a 0.7 scaling factor for the Slater G parameter. A spectral broadening of 0.25 eV was applied for all calculated spectra.

The calculated ground state reflectivity of Re6Se8Cl2\mathrm{Re}_{6}\mathrm{Se}_{8}\mathrm{Cl}_{2} is plotted in Figure S3 and is compared to experimentally measured XUV ground states at either edge. As shown in Figure S3, the BSE calculated ground state agrees well with the measured ground state reflectivity of the sample.

Following ground state calculations of the XUV reflectivity as the Re O2_2 and Se M4,5_{4,5} edges, excited state calculations were performed using a modification to the OCEAN code discussed previously.[14] The modification to the OCEAN code enables selective excitation to valence and conduction bands at different points in momentum space, to simulate state-filling of electrons and holes following photoexcitation. To model above band gap photoexcitation, we populate the conduction band with electrons and similarly populate the valence band with holes to model the state-filling of carriers following photoexcitation. This population of electrons and holes effectively forbids or allows XUV transitions to specific points in kk-space but does not account for carrier density, therefore, the modeled peak position and relative intensity is used to compare experiment to theory. We model our transient XUV spectra by calculating the differential absorption between the calculated ground state and the calculated excited state spectra. Figure 2 and Figure 3 in the main text demonstrate agreement between BSE modeled excited states at the ReO2\mathrm{ReO}_{2} and Se M4,5\mathrm{Se}\ \mathrm{M}_{4,5} edges, respectively.

Specifically, to model carrier thermalization and excited-state state filling of electrons and holes, we employ two “cutoff” energies: EhE_h and EeE_e. Where EhE_h is the maximum hole depth possible following 3.1 eV excitation the valence bands and EeE_e is the maximum electron energy accessible following 3.1 eV excitation in the conduction bands. Here, EhE_h is defined as the energy separation between the CBM and the deepest depopulated valence state, while EeE_e is defined as the energy separation between the VBM and the highest populated conduction state. Equations (1) and (2) depict this relation:

Eh=ECBM−Eocch(1)E_h = E_{CBM} - E_{occ}^{h} \tag*{(1)}
Ee=Eocce−EVBM(2)E_e = E_{occ}^{e} - E_{VBM} \tag*{(2)}

Here, EocchE_{occ}^{h} and EocceE_{occ}^{e} are the energetic positions within the band structure of the electrons and holes that are being applied through the excited-state BSE modeling. Thus, EhE_h and EeE_e are employed to parameterize the pump-accessible excited carrier manifold rather than directly visualize carrier energies referenced to their own band edges. For the 3.1 eV pump employed in our XUV experiments, the largest allowed values are EeE_e, Eh≤3.1E_h \le3.1 eV, and smaller values correspond to carrier thermalization toward the respective band edges of the carriers.

This parameterization allows us to separably map the thermalization of electrons and holes. Figure S4 and Figure S5 show different energy holes (a) and electrons (b) to model thermalization effects immediately following excitation, all the way to thermalization at the band edge at the ReO2\mathrm{ReO}_{2} and Se M4,5\mathrm{Se}\ \mathrm{M}_{4,5} edges, respectively. Figure S4 presents the combination of spectra calculated in S4(a) and S4(b) where a fixed electron cutoff energy of 2.0 eV is applied to mimic little-to-no electron thermalization at the ReO2\mathrm{ReO}_{2} edge, and a varied hole cutoff energy is employed to model ultrafast hole cooling to the valence band edge. This scenario agrees best with experimental XUV spectra at the ReO2\mathrm{ReO}_{2} edge. Similarly, Figure S5 presents the combination of spectra calculated in S5(a) and S5(b) at the Se M4,5\mathrm{Se}\ \mathrm{M}_{4,5} edge, however, a fixed electron cutoff energy of 3.1 eV is applied to mimic little-to-no electron thermalization, and a varied hole cutoff energy is employed to model ultrafast hole cooling to the valence band edge. This scenario agrees best with experimental spectra at the Se M4,5\mathrm{Se}\ \mathrm{M}_{4,5} edge and suggests that higher energy electrons are being probed at the Se M4,5\mathrm{Se}\ \mathrm{M}_{4,5} edge due to the element-specific nature of the core-level probe. Different projected densities of states are probed by different core-valence transitions. However, in both the case of Re and Se atoms, very minimal electron thermalization is observed spectrally, as supported by BSE theory.

BSE modeled excited-state carrier dynamics at the Re O2 edge

Figure S4. BSE modeled excited-state carrier dynamics for holes (a), electrons (b), and holes and electrons together (c) at the Re O2_2 edge. The electron and hole cutoff energies correspond to the VBM and CBM of the Re6_6Se8_8Cl2_2 electronic structure respectively.

BSE modeled excited-state carrier dynamics at the Se M4,5 edge

Figure S5. BSE modeled excited-state carrier dynamics for holes (a), electrons (b), and holes and electrons together (c) at the Se M4,5_{4,5} edge. A higher electron energy is needed in (c) to agree with experiment compared to the Re O2_2 edge, suggesting higher energy electrons are selectively probed at the Se M4,5_{4,5} edge due to the element-sensitivity of the transient XUV technique.

References

  1. [1]Choi, B. et al. Two-Dimensional Hierarchical Semiconductor with Addressable Surfaces. J. Am. Chem. Soc. 140, 9369–9373 (2018).DOI
  2. [2]Leduc, L., Perrin, A. & Sergent, M. Structure du dichlorure et octaséléniure d’hexarhénium, Re₆Se₈Cl₂: composé bidimensionnel à clusters octaédriques Re₆. Acta Crystallogr C 39, 1503–1506 (1983).DOI
  3. [3]Liu, H. et al. Measuring Photoexcited Electron and Hole Dynamics in ZnTe and Modeling Excited State Core-Valence Effects in Transient Extreme Ultraviolet Reflection Spectroscopy. J. Phys. Chem. Lett. 2106–2111 (2023) doi:10.1021/acs.jpclett.2c03894.DOI
  4. [4]Kim, Y.-J. et al. Electron charge hopping suppresses photoexcited small polarons in ErFeO₃ by anadiabatic formation mechanism. Sci. Adv. 10, eadk4282 (2024).DOI
  5. [5]Mendes, J. L. et al. Dynamic Competition between Hubbard and Superexchange Interactions Selectively Localizes Electrons and Holes through Polarons. J. Am. Chem. Soc. https://doi.org/10.1021/jacs.4c16837 (2025) doi:10.1021/jacs.4c16837.
  6. [6]Zhong, X. et al. Superatomic Two-Dimensional Semiconductor. Nano Lett. 18, 1483–1488 (2018).DOI
  7. [7]Géneaux, R., Chang, H.-T., Schwartzberg, A. M. & Marroux, H. J. B. Source noise suppression in attosecond transient absorption spectroscopy by edge-pixel referencing. Opt. Express 29, 951 (2021).DOI
  8. [8]Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 77, 3865–3868 (1996).pubmed.ncbi.nlm.nih.gov/10062328
  9. [9]Giannozzi, P. et al. Advanced capabilities for materials modelling with Quantum ESPRESSO. J. Phys. Condens. Matter 29, 465901 (2017).
  10. [10]Giannozzi, P. et al. QUANTUM ESPRESSO: a modular and open-source software project for quantum simulations of materials. J. Phys. Condens. Matter 21, 395502 (2009).DOI
  11. [11]Vinson, J. Advances in the OCEAN-3 spectroscopy package. Phys. Chem. Chem. Phys. 24, 12787–12803 (2022).DOI
  12. [12]Vinson, J., Rehr, J. J., Kas, J. J. & Shirley, E. L. Bethe–Salpeter equation calculations of core excitation spectra. Phys. Rev. B 83, 115106 (2011).DOI
  13. [13]Gilmore, K. et al. Efficient implementation of core-excitation Bethe–Salpeter equation calculations. Comput. Phys. Commun. 197, 109–117 (2015).DOI
  14. [14]Klein, I. M., Krotz, A., Lee, W., Michelsen, J. M. & Cushing, S. K. Ab Initio Calculations of XUV Ground and Excited States for First-Row Transition Metal Oxides. J. Phys. Chem. C 127, 1077–1086 (2023).arxiv.org/abs/2209.06347

Paper details

vlnc:W_sf5mkpzf
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents