Chimie ParisTech

Based on SAFT equations of state (EoS), we propose a physics informed neural network that learns the residual Helmholtz free energy, yielding pressure, chemical potentials, entropy, and phase equilibria by automatic differentiation. Transport properties are connected through additional relations. The model can be pretrained on synthetic EoS data and fine-tuned with heterogeneous experimental data. The model operates on logarithmic scale and is designed to stabilize learning properties that span several orders of magnitude between phases, state points, and molecules. A differentiable transport model further allows viscosity and self-diffusion data to learn the temperature derivative of the Helmholtz energy during fine-tuning. As a proof of concept, we fine-tune the model to experimental temperature–pressure–density, viscosity, and self-diffusion data for water. Despite using only state point information and transport data for this single, particularly challenging fluid, the model predicts vapor pressure, vaporization enthalpy, density and entropy anomalies, and the critical temperature without direct exposure to these properties during fine-tuning. This demonstrates that jointly learning Helmholtz energies and transport coefficients is viable and beneficial. We argue that such Helmholtz energy centered, data driven deep learning models provide an alternative to the current proliferation of isolated, single property machine learning solutions. Generalization to unseen molecules and a fixed functional mode with optimized molecule representation parameters are also demonstrated.

The synthesis, characterization, and electronic structure of a mixed-valent triiron complex of redox-active tppz ligands [{Fe0(CO)2(μ-tppz•-)}2FeII] are discussed. This complex bears a low-spin FeII center, bound to two tppz radical anions, that displays no exchange coupling at room temperature between the unpaired electrons of the ligands. The assignment of this electronic structure is supported by Mössbauer, EPR, and IR spectroscopies, solid-state geometric parameters, SQUID magnetometry, multireference calculations, and cyclic voltammetry. The electronic structure of the title compound is compared to previously reported FeII complexes of polyazine radicals, which bear high-spin FeII centers with strong antiferromagnetic coupling with the ligand-centered unpaired electrons. This difference sparks a discussion on the relative importance of ligand-metal exchange coupling and ligand field strength in controlling the configuration of metal complexes of redox-active ligands.

Eyring's absolute rate theory relates fluid flow with the activation energy necessary for a molecule to go from one equilibrium position to another. Developed nearly a century ago, it remains a powerful approach for understanding transport processes in liquids. This work presents a revision of Eyring's theory by replacing the vaporization-based parts of the theory with a residual approach that conceptualizes transport as local (re)movement of molecules within the system rather than removal from it. In this new approach, the energy barrier corresponds to the difference between a molecule with and without intermolecular interactions, effectively treating it as the residual property related to an ideal gas reference state. Furthermore, we explore the physical connections between Eyring's absolute rate theory and Rosenfeld excess entropy scaling, revealing that both approaches describe complementary aspects of the same transport phenomena. It is tempting to link both theories. The activation parameters in Eyring's theory, particularly the energy of activation, are shown to relate to the residual entropy used in entropy scaling. This provides a pathway to establish entropy scaling on a more rigorous physical foundation while offering deeper insights into the molecular mechanisms governing viscous flow. In comparison to experimental data, the revised theory demonstrated significant predictive power for viscosity across a wide range of thermodynamic conditions and species, including associating liquids. Moreover, parameters are transferable to other properties, such as self-diffusion.

Since Cu(InxGa1 - x)Se2 (CIGS) absorbers are among the most efficient materials for photovoltaic energy conversion, it is crucial to understand their degradation mechanisms and the key factors driving them. To do so, we present aging studies up to 2000 h at 80% relative humidity and 20°C, first under illumination and then in the dark. Three complementary techniques are used to unveil the evolution of the absorber at different probing depths. Surface and subsurface chemical reactivities are investigated, respectively, thanks to XPS (x-ray Photoelectron Spectroscopy) and HaXPES (Hard x-ray Photoemission Spectroscopy) characterizations (probing from 3 up to 41 nm). In addition, EDS (Energy Dispersive Spectroscopy) measurements unveil the bulk chemical reactivity at 0.7 and 1.3 µm. The main result is a greater sensitivity of the absorber's surface when RH is combined with light rather than dark exposure, which triggers the CIGS network faster, leading to oxide phase growth, Se0 side phases appearance, and Na diffusion. This shows that the degradation is boosted in the presence of light. On the other hand, the CIGS bulk results intact around 1 µm under all the ambient conditions investigated.

The diiron(I) complexes [Fe2Cp2(CNR)(3){CN(Me) (R)}][CF3SO3] ([2a-b][CF3SO3], R = Xyl, Cy; Xyl = 2,6-C6H3Me2, Cy = cyclohexyl) and [Fe2Cp2(CNR)(2){CN(Me) (R)}(2)][CF3SO3](2) ([3a-c][CF3SO3](2), R = Xyl, Cy, Me) were synthesized in moderate-to-high yields from [Fe2Cp2(CO)(4)]. They were characterized by IR and NMR spectroscopy and single-crystal X-ray diffraction in five representative cases. The electrochemical behavior of [2a][CF3SO3] and [3a][CF3SO3](2) was investigated by cyclic voltammetry and IR-spectroelectrochemistry, assisted by DFT calculations. UV-vis and H-1 NMR studies showed >= 84% of each complex remaining intact after 72 h of incubation at 37 degrees C in aqueous media. Complexes [3a-c][CF3SO3](2) displayed significant aqueous solubility, and [2b][CF3SO3] and [3a-b][CF3SO3](2) exhibited amphiphilic character (-1 < Log P-ow < +1). The monocationic species [2a-b][CF3SO3] exerted strong cytotoxicity in A2780, A549, and U87 cancer cell lines, with [2a][CF3SO3] showing IC50 values in the nanomolar range. Among dicationic complexes, [3a][CF3SO3](2) exhibited significant activity despite limited cellular uptake, whereas [3b-c][CF3SO3](2) were inactive. Compounds [2a][CF3SO3], [2b][CF3SO3] and [3a][CF3SO3](2) impaired mitochondrial bioenergetics without detectable elevation of intracellular ROS levels in both A549 and U87 cells. No significant interaction with double-stranded DNA was detected by ethidium bromide displacement assays. Conversely, fluorescence titrations with bovine serum albumin indicated the formation of 1:1 adducts for [2a](+) and [3a](2+).