Closed-loop transfer enables artificial intelligence to yield chemical knowledge.

Artificial intelligence-guided closed-loop experimentation has emerged as a promising method for optimization of objective functions1,2, but the substantial potential of this traditionally black-box approach to uncovering new chemical knowledge has remained largely untapped. Here we report the integration of closed-loop experiments with physics-based feature selection and supervised learning, denoted as closed-loop transfer (CLT), to yield chemical insights in parallel with optimization of objective functions. CLT was used to examine the factors dictating the photostability in solution of light-harvesting donor-acceptor molecules used in a variety of organic electronics applications, and showed fundamental insights including the importance of high-energy regions of the triplet state manifold. This was possible following automated modular synthesis and experimental characterization of only around 1.5% of the theoretical chemical space. This physics-informed model for photostability was strengthened using multiple experimental test sets and validated by tuning the triplet excited-state energy of the solvent to break out of the observed plateau in the closed-loop photostability optimization process. Further applications of CLT to additional materials systems support the generalizability of this strategy for augmenting closed-loop strategies. Broadly, these findings show that combining interpretable supervised learning models and physics-based features with closed-loop discovery processes can rapidly provide fundamental chemical insights.

Secondary structure determines electron transport in peptides.

Proteins play a key role in biological electron transport, but the structure-function relationships governing the electronic properties of peptides are not fully understood. Despite recent progress, understanding the link between peptide conformational flexibility, hierarchical structures, and electron transport pathways has been challenging. Here, we use single-molecule experiments, molecular dynamics (MD) simulations, nonequilibrium Green's function-density functional theory (NEGF-DFT), and unsupervised machine learning to understand the role of secondary structure on electron transport in peptides. Our results reveal a two-state molecular conductance behavior for peptides across several different amino acid sequences. MD simulations and Gaussian mixture modeling are used to show that this two-state molecular conductance behavior arises due to the conformational flexibility of peptide backbones, with a high-conductance state arising due to a more defined secondary structure (beta turn or 310 helices) and a low-conductance state occurring for extended peptide structures. These results highlight the importance of helical conformations on electron transport in peptides. Conformer selection for the peptide structures is rationalized using principal component analysis of intramolecular hydrogen bonding distances along peptide backbones. Molecular conformations from MD simulations are used to model charge transport in NEGF-DFT calculations, and the results are in reasonable qualitative agreement with experiments. Projected density of states calculations and molecular orbital visualizations are further used to understand the role of amino acid side chains on transport. Overall, our results show that secondary structure plays a key role in electron transport in peptides, which provides broad avenues for understanding the electronic properties of proteins.

Assessing molecular doping efficiency in organic semiconductors with reactive Monte Carlo.

The addition of molecular dopants into organic semiconductors (OSCs) is a ubiquitous augmentation strategy to enhance the electrical conductivity of OSCs. Although the importance of optimizing OSC-dopant interactions is well-recognized, chemically generalizable structure-function relationships are difficult to extract due to the sensitivity and dependence of doping efficiency on chemistry, processing conditions, and morphology. Computational modeling for an integrated OSC-dopant design is an attractive approach to systematically isolate fundamental relationships, but requires the challenging simultaneous treatment of molecular reactivity and morphology evolution. We present the first computational study to couple molecular reactivity with morphology evolution in a molecularly doped OSC. Reactive Monte Carlo is employed to examine the evolution of OSC-dopant morphologies and doping efficiency with respect to dielectric, the thermodynamic driving for the doping reaction, and dopant aggregation. We observe that for well-mixed systems with experimentally relevant dielectric constants, doping efficiency is near unity with a very weak dependence on the ionization potential and electron affinity of OSC and dopant, respectively. At experimental dielectric constants, reaction-induced aggregation is observed, corresponding to the well-known insolubility of solution-doped materials. Simulations are qualitatively consistent with a number of experimental studies showing a decrease of doping efficiency with increasing dopant concentration. Finally, we observe that the aggregation of dopants lowers doping efficiency and thus presents a rational design strategy for maximizing doping efficiency in molecularly doped OSCs. This work represents an important first step toward the systematic integration of molecular reactivity and morphology evolution into the characterization of multi-scale structure-function relationships in molecularly doped OSCs.

Distilling coarse-grained representations of molecular electronic structure with continuously gated message passing.

Bottom-up methods for coarse-grained (CG) molecular modeling are critically needed to establish rigorous links between atomistic reference data and reduced molecular representations. For a target molecule, the ideal reduced CG representation is a function of both the conformational ensemble of the system and the target physical observable(s) to be reproduced at the CG resolution. However, there is an absence of algorithms for selecting CG representations of molecules from which complex properties, including molecular electronic structure, can be accurately modeled. We introduce continuously gated message passing (CGMP), a graph neural network (GNN) method for atomically decomposing molecular electronic structure sampled over conformational ensembles. CGMP integrates 3D-invariant GNNs and a novel gated message passing system to continuously reduce the atomic degrees of freedom accessible for electronic predictions, resulting in a one-shot importance ranking of atoms contributing to a target molecular property. Moreover, CGMP provides the first approach by which to quantify the degeneracy of "good" CG representations conditioned on specific prediction targets, facilitating the development of more transferable CG representations. We further show how CGMP can be used to highlight multiatom correlations, illuminating a path to developing CG electronic Hamiltonians in terms of interpretable collective variables for arbitrarily complex molecules.

Excited State Dynamics of a Conformationally Fluxional Copper Coordination Complex.

The conversion of solar energy into chemical fuel represents a capstone goal of the 21st century and has the potential to supply terawatts of power in a globally distributed manner. However, the disparate time scales of photodriven charge separation (∼fs) and steps in chemical reactions (∼µs) represent an inherent bottleneck in solar-to-fuels technology. To address this discrepancy, we are developing earth-abundant coordination complexes that undergo light-induced conformational rearrangements such that charge separation (CS) is hastened, while charge recombination (CR) is slowed. To these ends, we report the preparation and characterization of a new series of conformationally fluxional copper coordination complexes that contain a twisted intramolecular charge transfer (TICT) fluorophore as part of their ligand scaffold. Structural and spectroscopic characterization of the Cu(I) and Cu(II) complexes formed with these ligands in their ground states establish oxidation state-dependent conformational dynamicity, while time-resolved emission and transient absorption spectroscopies define the photophysical parameters of photo-induced excited states. Building on initial reports with a related set of molecules, the improved ligand design presented here greatly simplifies the observed photophysics, effectively shutting down unwanted ligand-centered excited states previously observed. Time-dependent density functional theory (TDDFT) analyses reveal an unusual metal-to-TICT electronic transition only reported once before, and though the formation of a CS state is not observed directly through experiments, TDDFT geometry optimizations in the excited states support the formation of transient Cu(II) CS species, lending credence to the potential success of our approach. These studies establish a clear model for the excited state dynamics at play in proof-of-concept systems and clarify key design parameters for future optimizations toward achieving long-lived CS via photoinduced conformational gating.

AL4GAP: Active learning workflow for generating DFT-SCAN accurate machine-learning potentials for combinatorial molten salt mixtures.

Machine learning interatomic potentials have emerged as a powerful tool for bypassing the spatiotemporal limitations of ab initio simulations, but major challenges remain in their efficient parameterization. We present AL4GAP, an ensemble active learning software workflow for generating multicomposition Gaussian approximation potentials (GAP) for arbitrary molten salt mixtures. The workflow capabilities include: (1) setting up user-defined combinatorial chemical spaces of charge neutral mixtures of arbitrary molten mixtures spanning 11 cations (Li, Na, K, Rb, Cs, Mg, Ca, Sr, Ba and two heavy species, Nd, and Th) and 4 anions (F, Cl, Br, and I), (2) configurational sampling using low-cost empirical parameterizations, (3) active learning for down-selecting configurational samples for single point density functional theory calculations at the level of Strongly Constrained and Appropriately Normed (SCAN) exchange-correlation functional, and (4) Bayesian optimization for hyperparameter tuning of two-body and many-body GAP models. We apply the AL4GAP workflow to showcase high throughput generation of five independent GAP models for multicomposition binary-mixture melts, each of increasing complexity with respect to charge valency and electronic structure, namely: LiCl-KCl, NaCl-CaCl2, KCl-NdCl3, CaCl2-NdCl3, and KCl-ThCl4. Our results indicate that GAP models can accurately predict structure for diverse molten salt mixture with density functional theory (DFT)-SCAN accuracy, capturing the intermediate range ordering characteristic of the multivalent cationic melts.

Efficient Intermolecular Charge Transport in π-Stacked Pyridinium Dimers Using Cucurbit[8]uril Supramolecular Complexes.

Intermolecular charge transport through π-conjugated molecules plays an essential role in biochemical redox processes and energy storage applications. In this work, we observe highly efficient intermolecular charge transport upon dimerization of pyridinium molecules in the cavity of a synthetic host (cucurbit[8]uril, CB[8]). Stable, homoternary complexes are formed between pyridinium molecules and CB[8] with high binding affinity, resulting in an offset stacked geometry of two pyridiniums inside the host cavity. The charge transport properties of free and dimerized pyridiniums are characterized using a scanning tunneling microscope-break junction (STM-BJ) technique. Our results show that π-stacked pyridinium dimers exhibit comparable molecular conductance to isolated, single pyridinium molecules, despite a longer transport pathway and a switch from intra- to intermolecular charge transport. Control experiments using a CB[8] homologue (cucurbit[7]uril, CB[7]) show that the synthetic host primarily serves to facilitate dimer formation and plays a minimal role on molecular conductance. Molecular modeling using density functional theory (DFT) reveals that pyridinium molecules are planarized upon dimerization inside the host cavity, which facilitates charge transport. In addition, the π-stacked pyridinium dimers possess large intermolecular LUMO-LUMO couplings, leading to enhanced intermolecular charge transport. Overall, this work demonstrates that supramolecular assembly can be used to control intermolecular charge transport in π-stacked molecules.

Bypassing backmapping: Coarse-grained electronic property distributions using heteroscedastic Gaussian processes.

We employ deep kernel learning electronic coarse-graining (DKL-ECG) with approximate Gaussian processes as a flexible and scalable framework for learning heteroscedastic electronic property distributions as a smooth function of coarse-grained (CG) configuration. The appropriateness of the Gaussian prior on predictive CG property distributions is justified as a function of CG model resolution by examining the statistics of target distributions. The certainties of predictive CG distributions are shown to be limited by CG model resolution with DKL-ECG predictive noise converging to the intrinsic physical noise induced by the CG mapping operator for multiple chemistries. Further analysis of the resolution dependence of learned CG property distributions allows for the identification of CG mapping operators that capture CG degrees of freedom with strong electron-phonon coupling. We further demonstrate the ability to construct the exact quantum chemical valence electronic density of states (EDOS), including behavior in the tails of the EDOS, from an entirely CG model by combining iterative Boltzmann inversion and DKL-ECG. DKL-ECG provides a means of learning CG distributions of all-atom properties that are traditionally "lost" in CG model development, introducing a promising methodological alternative to backmapping algorithms commonly employed to recover all-atom property distributions from CG simulations.

Charge transport network dynamics in molecular aggregates.

Due to the nonperiodic nature of charge transport in disordered systems, generating insight into static charge transport networks, as well as analyzing the network dynamics, can be challenging. Here, we apply time-dependent network analysis to scrutinize the charge transport networks of two representative molecular semiconductors: a rigid n-type molecule, perylenediimide, and a flexible p-type molecule, [Formula: see text] Simulations reveal the relevant timescale for local transfer integral decorrelation to be [Formula: see text]100 fs, which is shown to be faster than that of a crystalline morphology of the same molecule. Using a simple graph metric, global network changes are observed over timescales competitive with charge carrier lifetimes. These insights demonstrate that static charge transport networks are qualitatively inadequate, whereas average networks often overestimate network connectivity. Finally, a simple methodology for tracking dynamic charge transport properties is proposed.

Layered nested Markov chain Monte Carlo.

A configurational sampling algorithm based on nested layerings of Markov chains (Layered Nested Markov Chain Monte Carlo or L-NMCMC) is presented for simulations of systems characterized by rugged free energy landscapes. The layerings are generated using a set of auxiliary potential energy surfaces. The implementation of the method is demonstrated in the context of a rugged, two-dimensional potential energy surface. The versatility of the algorithm is next demonstrated on a simple, many-body system, namely, a canonical Lennard-Jones fluid in the liquid state. In that example, different layering schemes and auxiliary potentials are used, including variable cutoff distances and excluded-volume tempering. In addition to calculating a variety of properties of the system, it is also shown that L-NMCMC, when combined with a free-energy perturbation formalism, provides a straightforward means to construct approximate free-energy surfaces at no additional computational cost using the sampling distributions of each auxiliary Markov chain. The proposed L-NMCMC scheme is general in that it could be complementary to any number of methods that rely on sampling from a target distribution or methods that exploit a hierarchy of time scales and/or length scales through decomposition of the potential energy.

Exciton Absorption Spectra by Linear Response Methods: Application to Conjugated Polymers.

The theoretical description of the time-evolution of excitons requires, as an initial step, the calculation of their spectra, which has been inaccessible to most users due to the high computational scaling of conventional algorithms and accuracy issues caused by common density functionals. Previously (J. Chem. Phys. 2016, 144, 204105), we developed a simple method that resolves these issues. Our scheme is based on a two-step calculation in which a linear-response TDDFT calculation is used to generate orbitals perturbed by the excitonic state, and then a second linear-response TDDFT calculation is used to determine the spectrum of excitations relative to the excitonic state. Herein, we apply this theory to study near-infrared absorption spectra of excitons in oligomers of the ubiquitous conjugated polymers poly(3-hexylthiophene) (P3HT), poly(2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene) (MEH-PPV), and poly(benzodithiophene-thieno[3,4-b]thiophene) (PTB7). For P3HT and MEH-PPV oligomers, the calculated intense absorption bands converge at the longest wavelengths for 10 monomer units, and show strong consistency with experimental measurements. The calculations confirm that the exciton spectral features in MEH-PPV overlap with those of the bipolaron formation. In addition, our calculations identify the exciton absorption bands in transient absorption spectra measured by our group for oligomers (1, 2, and 3 units) of PTB7. For all of the cases studied, we report the dominant orbital excitations contributing to the optically active excited state-excited state transitions, and suggest a simple rule to identify absorption peaks at the longest wavelengths. We suggest our methodology could be considered for further developments in theoretical transient spectroscopy to include nonadiabatic effects, coherences, and to describe the formation of species such as charge-transfer states and polaron pairs.

Mesoscale molecular network formation in amorphous organic materials.

High-performance solution-processed organic semiconductors maintain macroscopic functionality even in the presence of microscopic disorder. Here we show that the functional robustness of certain organic materials arises from the ability of molecules to create connected mesoscopic electrical networks, even in the absence of periodic order. The hierarchical network structures of two families of important organic photovoltaic acceptors, functionalized fullerenes and perylene diimides, are analyzed using a newly developed graph methodology. The results establish a connection between network robustness and molecular topology, and also demonstrate that solubilizing moieties play a large role in disrupting the molecular networks responsible for charge transport. A clear link is established between the success of mono and bis functionalized fullerene acceptors in organic photovoltaics and their ability to construct mesoscopically connected electrical networks over length scales of 10 nm.

Ultrafast Excited State Relaxation of a Metalloporphyrin Revealed by Femtosecond X-ray Absorption Spectroscopy.

Photoexcited Nickel(II) tetramesitylporphyrin (NiTMP), like many open-shell metalloporphyrins, relaxes rapidly through multiple electronic states following an initial porphyrin-based excitation, some involving metal centered electronic configuration changes that could be harnessed catalytically before excited state relaxation. While a NiTMP excited state present at 100 ps was previously identified by X-ray transient absorption (XTA) spectroscopy at a synchrotron source as a relaxed (d,d) state, the lowest energy excited state (J. Am. Chem. Soc., 2007, 129, 9616 and Chem. Sci., 2010, 1, 642), structural dynamics before thermalization were not resolved due to the ∼100 ps duration of the available X-ray probe pulse. Using the femtosecond (fs) X-ray pulses of the Linac Coherent Light Source (LCLS), the Ni center electronic configuration from the initial excited state to the relaxed (d,d) state has been obtained via ultrafast Ni K-edge XANES (X-ray absorption near edge structure) on a time scale from hundreds of femtoseconds to 100 ps. This enabled the identification of a short-lived Ni(I) species aided by time-dependent density functional theory (TDDFT) methods. Computed electronic and nuclear structure for critical excited electronic states in the relaxation pathway characterize the dependence of the complex's geometry on the electron occupation of the 3d orbitals. Calculated XANES transitions for these excited states assign a short-lived transient signal to the spectroscopic signature of the Ni(I) species, resulting from intramolecular charge transfer on a time scale that has eluded previous synchrotron studies. These combined results enable us to examine the excited state structural dynamics of NiTMP prior to thermal relaxation and to capture intermediates of potential photocatalytic significance.

Imaging ultrafast excited state pathways in transition metal complexes by X-ray transient absorption and scattering using X-ray free electron laser source.

This report will describe our recent studies of transition metal complex structural dynamics on the fs and ps time scales using an X-ray free electron laser source, Linac Coherent Light Source (LCLS). Ultrafast XANES spectra at the Ni K-edge of nickel(ii) tetramesitylporphyrin (NiTMP) were measured for optically excited states at a timescale from 100 fs to 50 ps, providing insight into its sub-ps electronic and structural relaxation processes. Importantly, a transient reduced state Ni(i) (π, 3dx2-y2) electronic state is captured through the interpretation of a short-lived excited state absorption on the low-energy shoulder of the edge, which is aided by the computation of X-ray transitions for postulated excited electronic states. The observed and computed inner shell to valence orbital transition energies demonstrate and quantify the influence of the electronic configuration on specific metal orbital energies. A strong influence of the valence orbital occupation on the inner shell orbital energies indicates that one should not use the transition energy from 1s to other orbitals to draw conclusions about the d-orbital energies. For photocatalysis, a transient electronic configuration could influence d-orbital energies up to a few eV and any attempt to steer the reaction pathway should account for this to ensure that external energies can be used optimally in driving desirable processes. NiTMP structural evolution and the influence of the porphyrin macrocycle conformation on relaxation kinetics can be likewise inferred from this study.

Boron Subphthalocyanine Based Molecular Triad Systems for the Capture of Solar Energy.

In this study a number of chromophores based on boron subphthalocyanines are investigated for use in the future design of organic photovoltaic devices based on molecular triad systems. The computational study is performed at the TD-DFT CAM-B3LYP/6-311G(d) level of theory. The absorption spectra of these chromophores are simulated using TD-DFT and compared to experimental results. All investigated chromophores absorb light in the visible range and thus are suitable for absorption of sunlight in solar cell applications. On the basis of energy-level alignments, suitable combinations of moieties for a molecular triad system are proposed. The molecular triads will be used in future work as the functional part of organic photovoltaic devices, where the chromophore will be used both to absorb the incoming solar radiation and to increase the distance between the separated charges on donor and acceptor units to increase the lifetime of the charge-separated state.

A n-vector model for charge transport in molecular semiconductors.

We develop a lattice model utilizing coarse-grained molecular sites to study charge transport in molecular semiconducting materials. The model bridges atomistic descriptions and structureless lattice models by mapping molecular structure onto sets of spatial vectors isomorphic with spin vectors in a classical n-vector Heisenberg model. Specifically, this model incorporates molecular topology-dependent orientational and intermolecular coupling preferences, including the direct inclusion of spatially correlated transfer integrals and site energy disorder. This model contains the essential physics required to explicitly simulate the interplay of molecular topology and correlated structural disorder, and their effect on charge transport. As a demonstration of its utility, we apply this model to analyze the effects of long-range orientational correlations, molecular topology, and intermolecular interaction strength on charge motion in bulk molecular semiconductors.

Conformational order in aggregates of conjugated polymers.

With the abundant variety and increasing chemical complexity of conjugated polymers proliferating the field of organic semiconductors, it has become increasingly important to correlate the polymer molecular structure with its mesoscale conformational and morphological attributes. For instance, it is unknown which combinations of chemical moieties and periodicities predictably produce mesoscale ordering. Interestingly, not all ordered morphologies result in efficient devices. In this work we have parametrized accurate classical force-fields and used these to compute the conformational and aggregation characteristics of single strands of common conjugated polymers. Molecular dynamics trajectories are shown to reproduce experimentally observed polymeric ordering, concluding that efficient organic photovoltaic devices span a range of polymer conformational classes, and suggesting that the solution-phase morphologies have far-reaching effects. Encouragingly, these simulations indicate that despite the wide-range of conformational classes present in successful devices, local molecular ordering, and not long-range crystallinity, appears to be the necessary requirement for efficient devices. Finally, we examine what makes a "good" solvent for conjugated polymers, concluding that dispersive π-electron solvent-polymer interactions, and not the electrostatic potential of the backbone interacting with the solvent, are what primarily determine a polymer's solubility in a particular solvent, and consequently its morphological characteristics.

Mesoscopic features of charge generation in organic semiconductors.

CONSPECTUS: In the past two decades, organic materials have been extensively investigated by numerous research groups worldwide for implementation in organic photovoltaic (OPV) devices. The interest in organic semiconductors is spurred by their potential low cost and facile tunability, making OPV devices a potentially disruptive technology. To study OPV operating mechanisms is also to explore a knowledge gap in our general understanding of materials, because both the time scales (femtosecond to microsecond) and length scales (nanometer to micrometer) relevant to OPV functionality occupy a challenging and fascinating space between the traditional regimes of quantum chemistry and solid-state physics. New theoretical frameworks and computational tools are needed to bridge the aforementioned length and time scales, and they must satisfy the criteria of computational tractability for systems involving 10(4)-10(6) atoms, while also maintaining predictive utility. While this challenge is far from solved, advances in density functional theory (DFT) have allowed researchers to investigate the ground- and excited-state properties of many intermediate sized systems (10(2)-10(3) atoms) that provide the outlines of the larger problem. Results on these smaller systems are already sufficient to predict optical gaps and trends in valence band energies, correct erroneous interpretations of experimental data, and develop models for charge generation and transport in OPV devices. The active films of high-efficiency OPV devices are comprised of mesoscopic mixtures of electron donor (D) and electron acceptor (A) species, a "bulk-heterojunction" (BHJ) device, subject to variable degrees of structural disorder. Depending on the degree of intermolecular electronic coupling and energy level alignment, the spatial delocalization of photoexcitations and charge carriers can affect the dynamics of the solar cell. In this Account, we provide an overview of three pivotal characteristics of solar cells that possess strong delocalization dependence: (1) the exciton binding energy, (2) charge transfer at the D-A heterojunction, and (3) the energy landscape in the vicinity of the D-A heterojunction. In each case, the length scale dependence can be assessed through DFT calculations on reference systems, with a view to establishing general trends. Throughout the discussion, we draw from the experimental and theoretical literature to provide a consistent view of what is known about these properties in actual BHJ blends. A consistent interpretation of the results to date affords the following view: transient delocalization effects and resonant charge transfer at the heterojunction are capable of funneling excitations away from trap states and mediating exciton dissociation; these factors alone are capable of explaining the remarkably good charge generation currently achieved in OPV devices. The exciton binding energy likely plays a minimal role in modern OPV devices, since the presence of the heterojunction serves to bypass the costly exciton-to-free-charge transition state.

Ultrafast structural dynamics of Cu(I)-bicinchoninic acid and their implications for solar energy applications.

In this study, ultrafast optical transient absorption and X-ray transient absorption (XTA) spectroscopy are used to probe the excited-state dynamics and structural evolution of copper(I) bicinchoninic acid ([Cu(I)(BCA)2](+)), which has similar but less frequently studied biquinoline-based ligands compared to phenanthroline-based complexes. The optical transient absorption measurements performed on the complex in a series of polar protic solvents demonstrate a strong solvent dependency for the excited lifetime, which ranges from approximately 40 ps in water to over 300 ps in 2-methoxyethanol. The XTA experiments showed a reduction of the prominent 1s → 4pz edge peak in the excited-state X-ray absorption near-edge structure (XANES) spectrum, which is indicative of an interaction with a fifth ligand, most likely the solvent. Analysis of the extended X-ray absorption fine structure (EXAFS) spectrum shows a shortening of the metal-ligand bond in the excited state and an increase in the coordination number for the Cu(II) metal center. A flattened structure is supported by DFT calculations that show that the system relaxes into a flattened geometry with a lowest-energy triplet state that has a dipole-forbidden transition to the ground state. While the short excited-state lifetime relative to previously studied Cu(I) diimine complexes could be attributed to this dark triplet state, the strong solvent dependency and the reduction of the 1s → 4pz peak in the XTA data suggest that solvent interaction could also play a role. This detailed study of the dynamics in different solvents provides guidance for modulating excited-state pathways and lifetimes through structural factors such as solvent accessibility to fulfill the excited-state property requirements for efficient light harvesting and electron injection.