Research

Overview

Figure 1: Primary scientific targets
The Vura-Weis lab measures electron transfer and reactivity in systems containing multiple heavy atoms at femtosecond to nanosecond timescales. The central thesis of our group is that by combining the ultrafast time resolution of femtosecond lasers with the element specificity of core-to-valence spectroscopy, we can measure short-lived states that are obscured using traditional spectroscopic techniques. Using a process called high-harmonic generation, we convert near-infrared laser pulses into extreme ultraviolet (XUV) pulses in the 30-100 eV energy range. This technique has primarily been developed in the physics and chemical physics communities, and it is the goal of our group to extend its use to mainstream problems in physical and inorganic chemistry. Our major scientific targets are shown in Figure 1: molecular photocatalysis with bimetallic metal complexes, photomagnetism and photoinduced spin crossover, and organohalide perovskite photovoltaics. These systems are an order of magnitude more complex thanthe usual targets of XUV spectroscopy, so this work requires innovations in both instrumentation and interpretation

What we learn with XUV spectroscopy

1) Electronic structure and dynamics of transition metal complexes

Figure 2: M-edge XANES spectra of first-row transition metal complexes, showing oxidation state, ligand field, spin state, and elemental specificity.

Transition metal clusters that catalyze multielectron reactions accumulate up to eight redox electrons or holes over a series of single electron transfer events. Much of the effort in understanding the mechanism of these clusters centers around determining the oxidation and spin state, or more generally the electronic structure, of each metal during and after each charge transfer step. We have shown that M-edge XANES spectroscopy, which probes 3p→3d transitions, is a sensitive probe of the electronic structure of first-row transition metals. As shown in Figure 2, a particular electronic structure (i.e. “octahedral high-spin CoII”) has a distinct spectral shape and position. We adapted a computational framework called the Ligand Field Multiplet (LFM) model, originally designed for L-edge spectra (2p→3d absorption), to simulate the M-edge spectra with excellent qualitative accuracy. This semiempirical method allows us to interpret the peaks in the absorption spectra using concepts from traditional inorganic spectroscopy. In collaboration with Prof. Frank de Groot of Utrecht University, we have developed a software tool that can project these spectra into a more intuitive orbital basis. For example, a particular feature might be identified as a transition into a state that is 90% t2g orbital-based, and the height of this peak in a range of similar molecules would indicate the electron occupation of that orbital set and therefore reflect the covalency of the metal-ligand bond.

Figure 3: Experimental and simulated XUV spectra of low-spin FeII complexes, showing a clear difference between strong-field 1Fe(CNC)2 and moderate-field 1Fe(Tren(Py)3

We continue to explore the sensitivity of M-edge spectra to minor changes in ligand field strength and symmetry. Figure 3 shows the experimental and simulated spectra of two representative low-spin FeII complexes. There is a significant edge shift and change in peak intensity ratios between the strong-field complex 1Fe(CNC)2 and the moderate-field complex 1Fe(Tren(Py)3). These trends are replicated in the semiempirical ligand field multiplet simulations, highlighting the predictive power of this relatively simple calculation.

Figure 4: Excited-state relaxation of FeTPPCl measured with transient M-edge XANES. Lifetimes in purple were measured with XUV spectropy, while ground-state vibrational relaxation (green) was measured with transient UV/Vis

Using this electronic structure sensitivity and the femtosecond time resolution of our instrument, we can now measure the ultrafast dynamics of transition metal complexes. In our first such study, we measured the excited-state relaxation of iron tetraphenyl porphyrin (FeTPPCl). Iron porphyrins are the active sites of many natural and artificial catalysts, and their photoinduced dynamics have been described as either relaxation into a vibrationally hot ground state or as a cascade through metal-centered states. Using our tabletop XUV probe, combined with semiempirical ligand field multiplet calculations, we distinguished between metal-centered and ligand-centered excited states and resolved competing accounts of Fe(III) porphyrin relaxation. This work (currently under review) introduced tabletop M-edge XANES as a valuable tool for measuring femtosecond dynamics of molecular transition metal complexes in the condensed phase.

In collaboration with the Rauchfuss and Girolami groups, we are now synthesizing multimetallic synthetic analogues of natural reaction centers and measuring the dynamics of photoinduced electron transfer to each metal (Figure 1). The elemental and oxidation state specificity of x-ray absorption will allow us to determine whether the electron tunnels directly to a localized state on a particular metal or passes through one or more intermediate states. These model complexes are active for H2 evolution, and we will measure the metal electronic structure in real time as H-H bonds are formed and the product is released. Our photomagnetic subgroup is measuring the interplay between electron transfer and spin crossover in heterobimetallic Prussian Blue analogues.

2) Carrier- and element-specific semiconductor photophysics

Figure 5:(A) Band offsets in a TiO2/mixed halide perovskite/CuSCN photovoltaic. (B) Simulated visible-pump/XUV probe transient spectra of this multilayer. The XUV light measures transitions from atomic core orbitals to the valence and conduction bands and allows the electron and hole to be separately tracked from material to material

Organohalide perovskite photovoltaics such as MAPbI3 (MA=methylammonium) are emerging as attractive alternatives to traditional semiconductor solar cells, due to their high efficiency (>22%) and solution processability. However, their photophysics are still poorly understood due to spectral overlap in the visible region, especially in devices with electron and hole collection layers such as TiO2 and CuSCN. We are using XUV spectroscopy to map the spatial separation of the electron and hole in each layer of a device onto an energetic separation in a transient spectrum. Unique signals are observed for the electron and the hole, enabling us to measure carrier-specific dynamics. The basic idea is illustrated in Figure 5. Photoexcitation of a mixed-halide perovskite above the band gap will initially create electrons and holes in a Br-based valence band. This will be observed as an positive transient signal around 70 eV representing induced absorption from the Br 3d orbitals to the valence band hole and a negative signal at 72 eV due to state-filling in the conduction band. Hole relaxation to an I-based valence band and then to the CuSCN collection layer will be observed as growth of positive features at 50 eV (I 4d→conduction band) and 80 eV (Cu 3p→3d), with the electron transfer to TiO2 observed as the growth of a negative signal at 40 eV (Ti 3p→3d).

Figure 6:(A) Simplified valence and conduction bands of PbI2, with Fermi levels and hole/electron populations after photoexcitation. (B) Transient spectrum expected after bandgap renormalization (BGR) and band-filling (BF). (C) Simulation compared to experimental transient spectrum.

We recently showed that the ground-state spectrum of the perovskite precursor PbI2 measures the conduction band density of states convolved with the core-hole spin-orbit splitting. Photoexcitation at 400 nm produces clearly distinguishable signals for electrons and holes. As shown in Figure 6, valence band holes open up a new absorption channel and cause a positive transient signal at 47.4 eV, approximately 2.5 eV below the ground-state absorption onset (this difference corresponds to the band gap). The electron signal is caused by a combination of bandgap renormalization and band-filling, resulting in an asymmetric derivative-shaped feature at the edge of the ground-state spectrum. Using values of band-filling and bandgap renormalization calculated from the excitation fluence, the transient spectrum can be simulated nearly quantitiatively. Rapid 2nd-order nonradiative recombination converts the initial electronic excitation into lattice heat, which is observed as a shrinking of the band-gap and broadening of the conduction band density of states. This relaxation pathway was confirmed using ultrafast electron diffraction.

We work with the Schleife group in MatSE to simulate the ground- and excited-state spectra of these systems. While the simple model shown in Figure 6 captures about about 80% of the photophysics, further analysis requires a careful treatment of the exact band structure and spin-orbit dependent transition dipole moments.

The Lab

X-ray absorption spectroscopy is an immensely powerful technique because it is element-specific, oxidation state-specific, and spin state-specific. We generate sub-30 femtosecond x-ray pulses in our own laboratory using the technique of High-Harmonic Generation. A powerful near-infrared laser pulse is focused into Neon gas, where the intense electric field ionizes the gas atoms, accelerates the free electrons, then drives them back into the gas nuclei, releasing an XUV pulse in the 30-100 eV range.

Figure 7: Tabletop transient x-ray apparatus. A laser pulse is focused into Ne or Ar gas, where the intense electric field ionizes the gas atoms, accelerates the free electrons, then drives them back into the gas nuclei, releasing XUV photons. The coherent XUV beam is focused onto the sample then dispersed onto an array CCD. Samples are photoexcited at 400 nm or by a tunable 500-1000 nm NOPA

top

Home * Research * Group Members * Publications * Funding * Links * Teaching
Copyright 2013 The Vura-Weis Research Group