Faraday Discussions最新文献

Minimum energy conical intersections (MECIs) are widely used to rationalize nonadiabatic relaxation pathways, yet it remains unclear to what extent they provide a representative description of the intersection seam effectively explored during dynamics. In this work, we combine large ensembles of seam points from nonadiabatic simulations with dimensionality reduction and density-based analysis to systematically investigate the structure of the accessed seam regions and its relationship to MECIs. We first establish a robust, physically-meaningful protocol for representing seam morphology across systems. After accounting for rigid-body alignment and permutation symmetry, Cartesian coordinates are found to best preserve structural relationships, while densMAP provides an optimal low-dimensional embedding that captures both geometric variation and sampling density. Applying this framework to ethylene, butadiene, and benzene, reveals that the intersection seam is organized into dynamically accessible basins that are not uniformly represented by discrete MECIs. While MECIs associated with highly populated basins are located near density maxima and capture the dominant relaxation pathways, many other MECIs correspond to peripheral or weakly sampled regions. Moreover, neither geometric proximity nor energetic accessibility alone reliably predicts dynamical relevance: geometrically closest MECIs do not always correspond to optimization outcomes, and low-energy MECIs may remain largely unvisited. These results demonstrate that MECIs should be viewed as discrete markers within a broader, continuous, and dynamically weighted seam landscape. A comprehensive understanding of photochemical relaxation therefore requires not only the identification of MECIs, but also an explicit characterization of the regions of the seam that are actually sampled and the pathways connecting them.

Accurate simulations of excited-state photodynamics of polyenes remain challenging due to the high computational cost of multireference electronic-structure methods required to describe the coexistence of ionic and covalent states. In this work, active learning together with machine learning interatomic potentials (MLIP) trained on multireference configuration interaction with single and double excitations (MR-CISD) level energies and gradients, is used to enable efficient and accurate nonadiabatic molecular dynamics for all-trans-hexatriene. A compact yet representative training set is constructed by reducing an initial pool of ∼850 k configurations to ∼3 k geometries while preserving the regions of configurational space relevant for the dynamics. The resulting model reproduces the topology of the excited-state potential energy surfaces and yields a physically consistent relaxation mechanism, characterized by a sequential S₂ → S₁ → S₀ pathway and time constants in good agreement with experimental data. Analysis of representative trajectories provides mechanistic insight into the structural factors governing nonadiabatic transitions, and identifies a notably broader and more diffuse structural distribution associated with the S₁ → S₀ transition in comparison to the more localized S₂ → S₁ decay pathway. Analysis of the torsional angles around the CC double bonds reveals preferential isomerization at the central C₃C₄ bond, providing semi-quantitative evidence for the initiation of cis-trans photoisomerization in the hot ground state. The MLIP simulations trained at high-level MR-CISD are extremely efficient, enabling 200 trajectories running up to 800 fs in short time, whereas direct on-the-fly multireference dynamics simulations would be practically impossible at this MR level.

In quantum chemistry calculations, molecular systems containing heavy elements require additional considerations compared with molecules comprising only light elements. In particular, accurate descriptions of both relativistic effects and static correlation are essential to obtain reliable wave functions and energies. Such calculations are computationally demanding due to the high cost of methods that properly incorporate relativistic effects and static correlation. This computational cost becomes even steeper when geometry optimisations or dynamical simulations are undertaken, as nuclear derivatives must be evaluated repeatedly. In this work, we combine the analytical gradients for the effective core potential (ECP), implemented in the LIBECPINT library, with the spin-free one-electron exact two-component (X2C) approach, implemented in our in-house version of BAGEL, for density-fitted MRPT2 methods. This enables scalar-relativistic geometry optimisations of molecules containing heavy elements, for both ground and electronically excited states, at a cost comparable to conventional (non-relativistic) MRPT2 calculations. We demonstrate the utility of the current computer program for optimising the geometries of Pd(IV) terminal oxo complexes and Fe(II) porphyrin.

The dynamics of photo-excited 4-bromophenolate anions in aqueous solution are investigated using femtosecond extreme ultraviolet (XUV) time-resolved photoelectron spectroscopy (TRPES) in a liquid flat jet. Excitation to the first singlet S₁ (1¹ππ*) excited state at 290 nm yields lifetimes of ∼160 fs and 1.1 ps that are attributed to the timescale for the transition from the initially excited S₁ state to the optically-dark ¹πσ* state and the subsequent decay of the ¹πσ* state, respectively. Excitation at 240 nm to the second singlet excited state (S₂, 2¹ππ*) reveals an additional ultrafast component assigned to S₂ → S₁ decay with a lifetime of 60 fs. Subsequent decay components show lifetimes of 130 fs and 1.4 ps, corresponding to the transition from S₁ to the ¹πσ* state and its subsequent decay. These dynamics differ from the phenolate anion, which undergoes S₁ → S₀ internal conversion and hydrated electron formation within ∼20 ps upon S₁ excitation. Instead, photoexcited 4-bromophenolate decays primarily through the ¹πσ* state, which is repulsive along the C-Br bond length and correlates to ˙C₆H₄O˙ (biradical fragment) + Br^-. Quantum chemical calculations suggest that the S₁-¹πσ* state crossing involves a small potential energy barrier that promotes a facile non-adiabatic transition. This study highlights the competing dynamics between photo-oxidation and non-adiabatic relaxation within the framework of anions in aqueous solution.

We present the PhotoChem Explorer (PCE) prototype, a proof-of-concept web-based platform that represents computational photochemistry studies of gas-phase molecules in a structured and queryable form. Rather than treating excited-state dynamics simulations as isolated results reported in heterogeneous formats and across different sources, PCE organizes studies according to their molecular systems, computational workflows, methodological choices, and associated observables, facilitating that relationships and trends can be explored systematically across studies. In this way, the prototype illustrates how structured data representations could support improved findability, comparison, reproducibility, and reuse in computational photochemistry. At a fundamental level, it demonstrates how excited-state dynamics studies could become part of a more connected and cumulative body of knowledge, potentially enabling new modes of investigation. The prototype exposes practical challenges that must be addressed to realize this vision, including the development of shared representations, workflow descriptions, and their integration into scientific practice. More broadly, PCE is intended as a practice-driven experimentation platform to stimulate discussion and iterative community development of standards and representations for excited-state dynamics studies.

We introduce a new equation-of-motion coupled-cluster method based on a pair coupled-cluster doubles (pCCD) reference, termed frozen-pair EOM-CCSD (EOM-fpCCSD). This approach combines the computational efficiency of the pCCD ansatz with a dynamical correlation correction, enabling a reliable description of electronically excited states within the EOM framework. The method has been implemented in the open-source PyBEST software package. Its performance is systematically benchmarked against standard EOM-CCSD and its pair-tailored variant (EOM-ptCCSD), using both canonical Hartree-Fock and pCCD natural orbitals. For charge-transfer (CT) excitations taken from the QUEST database, EOM-fpCCSD yields excitation energies very close to those of EOM-CCSD, outperforming EOM-ptCCSD, as well as to the theoretical best estimates (TBEs). Working within the localized pCCD natural orbital basis allows us to determine the directed CT character, which quantifies the directed charge flow from one molecular domain to another. Numerical results show that EOM-fpCCSD, EOM-CCSD, and EOM-ptCCSD provide nearly identical descriptions of the directed CT character, despite changes in excitation energies. The true advantage of EOM-fpCCSD becomes evident for the challenging QUEST subset of doubly excited states. While EOM-ptCCSD performs similarly to standard EOM-CCSD, EOM-fpCCSD significantly outperforms both methods for these problematic states compared to TBEs. In addition to improving the accuracy of excitation energies, EOM-fpCCSD also converges for several states that standard EOM-CCSD and EOM-ptCCSD fail to converge. These results demonstrate that EOM-fpCCSD offers a promising and computationally efficient route toward a more accurate description of complex electronic excitations.

Active learning has become central to modern machine-learning workflows for trajectory surface hopping, enabling automatic construction of training sets for excited-state potentials. However, currently used protocols can struggle with convergence, requiring large datasets and extensive human intervention. Here, we combine active learning with Δ-learning - where the ML model predicts the difference between a target quantum chemical method and a computationally efficient baseline - into an end-to-end protocol for active delta-learning (ADL) of potential energy surfaces for trajectory surface hopping dynamics. The protocol is tested on a benchmark system, fulvene, where a twofold data efficiency is achieved with respect to fine-tuning from a pre-trained foundational model, OMNI-P2x. Next, we analyze the case of PSB4, a protonated Schiff base with multiple competing torsional deactivation channels. Here, pure ML active learning fails to converge despite extensive sampling of over 14 000 points, while ADL converges within just two iterations and 900 sampled points, yielding good agreement with the reference.

The electrocatalytic reduction of CO₂ to multicarbon (C₂₊) products on copper-based catalysts is dictated by the dynamic reconstruction of the active sites. While copper-aluminium (CuAl) layered double hydroxides (LDHs) have emerged as promising pre-catalysts, the role of the initial precursor phase, specifically the transition from highly crystalline LDH to oxide-rich hybrid phases, and its influence on catalytic activity in near-neutral electrolytes remain less understood. In this study, Cu_xAl pre-catalysts were systematically investigated across Cu : Al ratios from 5 : 1 to 1 : 2 in aqueous K₂SO₄, with the electrolyte concentration adjusted to probe its complex role in activity. Under optimised conditions, using an oxide-rich 5 : 1 pre-catalyst, the faradaic efficiency for C₂₊ reached approximately 55.3% at 250 mA cm^-2 in a near-neutral environment. The selectivity toward C₂₊ is dependent on the amount of Cu present and the oxide richness, with intermediate Cu amounts providing a balance between the availability of Cu active sites and the Al-mediated structural stabilisation. In contrast, Al-rich compositions offer less favourable product distributions. Increasing the K₂SO₄ concentration modulates the cell voltage and C₂₊ product formation. This suggests that combined interfacial electrolyte effects, including shifted local pH gradients and cation-mediated stabilisation, regulate C-C coupling pathways. Physicochemical characterisation, particularly in situ Raman spectroscopy, and electrochemical measurements indicate that the initial stoichiometry of the CuAl LDH and the presence of oxides/hydroxides influence the nature of the reconstructed Cu-based catalyst, yielding behaviour consistent with the partial stabilisation of Cu⁺ under operating conditions. These results demonstrate that the Cu content of CuAl pre-catalysts, together with the interfacial electrolyte environment, plays a key role in directing CO₂ selectivity. This systematic study offers guidance on rationally tuning the precursor phase, oxide richness, and reaction environment of Cu-based electrocatalysts for high-rate CO₂ reduction in non-alkaline media.

We present a mechanistic study on selective alcohol decomposition over model cobalt oxide-based catalysts: pristine Co₃O₄(111)/Au(111) and Co₃O₄(111) containing different amounts and forms of vanadium oxide. These two types of model catalysts were prepared and investigated with respect to their structural and catalytic properties following a rigorous surface science approach under well-defined ultra-high vacuum (UHV) conditions employing a combination of scanning-tunnelling microscopy (STM), molecular-beam techniques, infrared reflection absorption spectroscopy (IRAS) and temperature programmed desorption (TPD). Upon vanadium deposition, three types of V-containing structures were identified on Co₃O₄(111): V atoms embedded into the Co₃O₄ lattice, and triangular two-dimensional (2D) and three-dimensional (3D) truncated triangular VO_x islands, whose relative abundance strongly depends on the amount of deposited V. Employing CO as a sensitive probe for adsorption sites, V atoms in different oxidation states were identified, including V³⁺, V⁴⁺ and V⁵⁺ involved in vanadyl (V⁵⁺O) groups. The latter species were shown to participate in dissociation of 2-propanol to the propoxy reaction intermediate, likely serving as a H acceptor. The propoxy intermediate was found to dissociate to acetone on pristine Co₃O₄(111) and Co₃O₄(111) containing low V coverages, while increasing V loading leads to an alternative reaction pathway towards propene proceeding via C-O bond scission. Both the adsorbed reaction intermediate preceding propene formation and the gaseous product propene evolving above 430 K in TPD were detected for this reaction pathway. We discuss the possible reaction mechanisms of both competing reaction routes and connect them to the structure of different types of VO_x species.

Binding energies of adsorbed CO and H are key descriptors governing the activity and selectivity of the co-electrolysis of CO₂ and H₂O to produce syngas with desired CO/H₂ ratios. Palladium hydride (PdH), which forms in situ at negative overpotentials, has been identified as the active Pd phase for CO₂ reduction to syngas. Herein, binding energies of CO and H are determined using temperature programmed desorption (TPD) of CO and H₂ from Pd(111), PdH/Pd(111), and Cu/PdH/Pd(111) under ultra-high vacuum (UHV) conditions. TPD results reveal that desorption of H₂ from subsurface PdH occurs at 460 K, while desorption from surface PdH is more facile at 320 K. CO desorption temperatures shift 20 K lower on PdH/Pd(111) compared to on Pd(111). The presence of 0.7 ML Cu further increases the desorption temperature of H₂ by 30 K while simultaneously reducing CO desorption temperatures by 70 K. Density functional theory (DFT) calculations show that CO adsorption onto Pd sites is hindered on the 0.7 ML Cu/PdH/Pd(111) surface while the kinetic barrier for H₂ desorption is increased. The trends in the binding energies of CO and H on model surfaces are consistent with electrochemical measurements of CuPd powder catalysts in a membrane electrode assembly (MEA), where H₂ evolution is reduced while CO production is enhanced compared to unmodified Pd catalysts. Overall, the results from model surface studies (TPD and DFT) provide a prediction and explanation for the activity and CO/H₂ ratios observed in electrochemical experiments. This study also demonstrates that CuPd is a promising catalyst with reduced Pd-loading to produce CO-rich syngas.