The Southern Ocean, wintertime cities, the upper troposphere, the Arctic and Antarctica, and alpine mountains are places where atmospheric chemistry impacts human health, air quality, climate, or geochemical cycles and that are characterized by low temperatures where ice or snow can be present. The atmospheric impact is evident from the role of polar biogenic sulphur emissions on aerosol formation, multiphase nitrogen and sulphur chemistry on wintertime haze, and industrial emissions in snow-covered areas on the ozone budget. The Cryosphere and ATmospheric CHemistry community (CATCH) addresses the environmental processes within these coupled cryosphere-atmosphere systems, and here we present open research questions specific to the cold environments, focusing on the unique interplay of chemistry and physics. These research needs call for interdisciplinary approaches to address atmospheric science in a warming climate with changing human impact in Earth's cold regions.

Active Thermochemical Tables (ATcT) were successfully used to resolve the existing inconsistencies related to the thermochemistry of glycine, based on statistically analyzing and solving a thermochemical network that includes >3350 chemical species interconnected by nearly 35 000 thermochemically-relevant determinations from experiment and high-level theory. The current ATcT results for the 298.15 K enthalpies of formation are -394.70 ± 0.55 kJ mol-1 for gas phase glycine, -528.37 ± 0.20 kJ mol-1 for solid α-glycine, -528.05 ± 0.22 kJ mol-1 for β-glycine, -528.64 ± 0.23 kJ mol-1 for γ-glycine, -514.22 ± 0.20 kJ mol-1 for aqueous undissociated glycine, and -470.09 ± 0.20 kJ mol-1 for fully dissociated aqueous glycine at infinite dilution. In addition, a new set of thermophysical properties of gas phase glycine was obtained from a fully corrected nonrigid rotor anharmonic oscillator (NRRAO) partition function, which includes all conformers. Corresponding sets of thermophysical properties of α-, β-, and γ-glycine are also presented.

We benchmark the rSCAN and r2SCAN exchange-correlation functionals by comparing the Nuclear Magnetic Resonance (NMR) magnetic shieldings predicted by Density Functional Theory (DFT) to experimentally observed chemical shifts of halide and oxide inorganic compounds. Significant improvement in accuracy is achieved compared to the Generalised Gradient Approximation (GGA) at a marginally higher computational cost. When using rSCAN or r2SCAN, the correlation coefficient between computationally predicted and experimental values approaches the theoretically expected value of -1 while reducing the deviation, allowing more accurate and reliable spectrum assignments of complex compounds in experimental investigations.

We consider membranes as fluid deformable surfaces and allow for higher order geometric terms in the bending energy related to the Gaussian curvature squared and the mean curvature minus the spontaneous curvature to the fourth power. The evolution equations are derived and numerically solved using surface finite elements. The two higher order geometric terms have different effects. While the Gaussian curvature squared term has a tendency to stabilize tubes and enhance the evolution towards equilibrium shapes, thereby facilitating rapid shape changes, the mean curvature minus the spontaneous curvature to the fourth power destabilizes tubes and leads to qualitatively different equilibrium shapes but also enhances the evolution. This is demonstrated in axisymmetric settings and fully three-dimensional simulations. We therefore postulate that not only surface viscosity but also higher order geometric terms in the bending energy contribute to rapid shape changes which are relevant for morphological changes of cells.

Elucidating how protein sequence determines the properties of disordered proteins and their phase-separated condensates is a great challenge in computational chemistry, biology, and biophysics. Quantitative molecular dynamics simulations and derived free energy values can in principle capture how a sequence encodes the chemical and biological properties of a protein. These calculations are, however, computationally demanding, even after reducing the representation by coarse-graining; exploring the large spaces of potentially relevant sequences remains a formidable task. We employ an "active learning" scheme introduced by Yang et al. (bioRxiv, 2022, https://doi.org/10.1101/2022.08.05.502972) to reduce the number of labelled examples needed from simulations, where a neural network-based model suggests the most useful examples for the next training cycle. Applying this Bayesian optimisation framework, we determine properties of protein sequences with coarse-grained molecular dynamics, which enables the network to establish sequence-property relationships for disordered proteins and their self-interactions and their interactions in phase-separated condensates. We show how iterative training with second virial coefficients derived from the simulations of disordered protein sequences leads to a rapid improvement in predicting peptide self-interactions. We employ this Bayesian approach to efficiently search for new sequences that bind to condensates of the disordered C-terminal domain (CTD) of RNA Polymerase II, by simulating molecular recognition of peptides to phase-separated condensates in coarse-grained molecular dynamics. By searching for protein sequences which prefer to self-interact rather than interact with another protein sequence we are able to shape the morphology of protein condensates and design multiphasic protein condensates.

The development of models that allow access to atomic-scale information is crucial for expanding our understanding of interfacial processes that enable the formation and growth of biominerals in both abiotic and biogeochemical environments. In this work, a recently developed potential model for the simulation of aspartic acid species in biomineralization environments has been combined with the most recent calcium carbonate potential model. Force field parameters have been refined and validated, demonstrating high accuracy in predicting ion pair formation dynamics and free energy. Two different options to model the interaction between aspartic acid species and carbonate anions have been discussed, and preliminary simulations of aspartic acid on calcium carbonate mineral surfaces have been performed. The steps to predict accurate binding free energies of molecules on mineral surfaces have been discussed using acetate and the {020} surface of monoclinic vaterite as a model system.

Adarsh V. Kalikadien opened a general discussion of the paper by Jiayun Pang: I was wondering about your broader view on the interpretability of ML models. In your paper you say: “FlanT5 seems slightly better at identifying key reagents involved in the reactions compared to ByT5” (https://doi.o

Crystallization is a fundamental process in biomineralization, yet its complexity increases significantly in non-classical pathways of nucleation and growth, where numerous intermediates exist between free ions in solution and the final crystalline phase. Characterization of such intermediates can be delicate, as some processes can proceed very fast, inducing a short temporal window for their characterization. The problem is particularly true for the so-called prenucleation species, for which this difficulty is made worse due to their solubility, nanometric size and highly dynamic nature. In this communication, we introduce an innovative approach based on the "trapping" of reactive intermediates in a vitreous matrix, inspired by cryo-TEM sample preparation but adapted specifically to solid-state NMR characterization. This approach enables time-resolved analysis, since the aging time of the mineralization reaction is controlled on the millisecond time-scale using a dedicated stopped-flow device. The cryo-fixation is achieved by spraying the reacting solution into cold liquid isopentane at -145 °C, and the NMR rotors are filled with specifically designed packing tools, ensuring the control of the low temperature during the whole process. Finally, the cryo-fixed solution can be studied via low-temperature solid-state NMR by acquiring time-resolved NMR spectra as snapshots of the ongoing reaction. First, we show that phosphate solutions can be efficiently vitrified using this protocol and studied via low-temperature 31P solid-state NMR at -120 °C. We demonstrate that the 31P chemical-shift interaction of cryo-quenched solutions varies with the pH, allowing us to distinguish the different phosphate species coexisting in solution (PO43-, HPO42-, H2PO4-and H3PO4) based on their chemical shift anisotropy patterns, which are characteristic of their protonation degree. The determination of the proportion of each species at varying pH levels enables us to construct a speciation diagram under our experimental conditions. We observe that the "apparent pKa" values, i.e., the pH values for each chemical equilibrium, are slightly influenced by the low temperature, and possibly by the preparation conditions. Finally, we demonstrate that our method can be applied to study fast calcium phosphate crystallization, revealing the early stages of amorphous calcium phosphate (ACP) nucleation within just 20 ms of reaction time, as shown by 31P ssNMR. Importantly, the described methodology is the first step towards studying fast out-of-equilibrium solutions through trapping and then studying transient intermediate species via solid-state NMR. Indeed, proper freeze-quenching prevents the transient species from transforming and preserves their native environment, such as hydration, pH, or ionic strength. We demonstrated applications for biomineralization-relevant reactions, but in principle, any aqueous reaction can be studied.