Portail HAL ENSCP
Not a member yet
9647 research outputs found
Sort by
From lab to market with sustainable sodium-ion batteries
No full textInternational audienceSodium-ion batteries (NIBs) have emerged as a promising alternative to lithium-ion batteries in many areas, including the mobility and grid-level storage sectors. They are now explicitly included in many technology-strategy roadmaps and are the subject of active research and development efforts around the globe. Current NIBs are enabled by three distinct chemical compositions, each of which has its own specific characteristics and, consequently, performance and economic considerations. This Review highlights the remaining scientific challenges for researchers and provides consumers with contemporary factual indicators to enable them to select the appropriate NIB for their needs
Corrigendum to “Sodium hydrosulfide hydrate as sodium precursor for low-cost synthesis of Na3SbS4 ionic conductor” [Solid State Ionics 427 (2025) 116892]
No full textInternational audienc
Atomistic Investigation of Crown Ether Complexes with Lithium Iodide as Solid-State Electrolytes
No full textInternational audienceWe investigated the surprising ionic conductivity trend of a series of hybrid complexes based on lithium iodide (LiI) associated with different size of crown ethers. Through a combination of experimental and theoretical methods, LiI hybrid complexes were characterized to rationalize the unexpected tendency of ionic conductivity. Our findings revealed that the solvation structure of the lithium iodide (LiI) complex of crown ethers is influenced by the size of the crown ether. Consequently, the solvation structure of the complex affects the transport properties. Notably, we observed a unique sandwich-like complex is formed between two 12C4 molecules and the Li cation during classical NVT molecular dynamics (MD) simulation, which is rarely observed with larger crown ethers. Based on this observation, we postulate that the idiosyncratic structure of [Li ⊂ 12C4][I] increases the ratio of SSIP solvation structure, ultimately resulting in higher ionic conductivity compared to [Li ⊂ 15C5][I], despite 15C5 is a larger crown ether than 12C4. We also propose a hypothesis, based on the MD simulation, that adjusting the size of crown ether molecules can potentially enhance the transference number
Boosting Cancer Chemoimmunotherapy with Mitochondria‐Targeting Iridium(III)‐Based Immunogenic Oncosis Inducers
No full textInternational audienceAbstract Most clinically used chemotherapeutic agents act by inducing apoptosis. However, their clinical effectiveness is often limited by poor therapeutic efficacy and the rapid development of drug resistance. In contrast, oncosis, as an inflammatory form of cell death independent of adenosine triphosphate (ATP) and apoptotic pathways, exhibits unique advantages in overcoming tumor drug resistance and regulating anti‐tumor immune responses. Herein, we present the first iridium(III)‐based immunogenic oncosis inducers designed to concurrently induce oncosis and activate the cGAS–STING pathway, thereby bridging chemotherapy with immunotherapy. Through a bioisosteric design strategy, we identified benzoselenazole and benzothiazole derivatives as key pharmacophores for triggering oncosis. These iridium(III)‐based oncosis‐inducers rapidly disrupt mitochondrial architecture, induce oxidative stress, and promote Ca(II) release, which subsequently activate calpain and porimin to initiate oncosis in multidrug‐resistant cancer cells. Transcriptomic profiling further revealed their ability to regulate actin cytoskeleton organization, modulate ABC transporter activity, and affect glycolysis/gluconeogenesis. Notably, the metal complexes induce mitochondrial swelling and mt‐DNA damage, leading to robust activation of the cGAS–STING innate immune pathway and eliciting a strong anticancer immune response. Based on these multimodal mechanisms, the Ir(III)‐based immunogenic oncosis inducers were able to effectively kill drug‐resistant cancer cells and enhance the anticancer immune response in tumor mouse models
Prediction of Propene Hydroformylation using Machine Learning
No full textHere, we develop machine learning (ML) models to predict catalytic performance in rhodium-catalyzed propene hydroformylation and identify new ligands and reaction conditions for enhanced yield and selectivity. A curated dataset of 215 experimentally obtained data points involving [Rh(acac)(CO)₂] precatalyst and 49 phosphine ligands was used. Molecular, electronic, and reaction descriptors were constructed using cheminformatic representations and dimensionality reduction. Among several algorithms tested, XGBoost exhibited the best performance, achieving a root-mean-square error (RMSE) of 9.85% for iso-selectivity under leave-one-ligand-out cross-validation. Feature importance analysis revealed that ligand and solvent descriptors most strongly influence selectivity, whereas turnover number (TON) predictions were more challenging. A bootstrapped ensemble of XGBoost models integrated with a genetic algorithm enabled the exploration of vast ligand–condition space, yielding 58 candidate ligands predicted to achieve TON ≥ 330 and Iso(%) ≥ 70. This study demonstrates that interpretable ML models can complement mechanistic understanding and accelerate catalyst design for small-alkene hydroformylation
Characterizing drug activity with sensitive interactomes in human living cells
No full textA number of human diseases results from abnormal protein-protein interactions (PPIs) involving key regulatory proteins. Therefore, an important strategy in therapeutics consists in developing inhibitory molecules that should ideally be specific for the aberrant PPI. In this context, it is critical to evaluate the number of PPIs that could be affected by the candidate molecule and to analyze the inhibitory potential before and after the formation of the PPI. Surprisingly, these two molecular aspects are rarely considered, due to a lack of appropriate methodological approaches. In this study, we present a novel methodology that captures drug-sensitive PPIs by considering drug-induced cellular functions in live cell conditions. As a proof-of-concept, we identified interactions of the human core signaling protein ERK1 that are specifically affected by two different inhibitory molecules. In addition, we used a complementary set of innovative tools that allowed visualizing the inhibitory effect on ERK1/cofactor protein complexes after their assembly in living cells. Overall, our work establishes a unique methodological approach for deciphering drug activity for potentially any target bait protein of interest
Fast-Tracking Solar Cell Materials Discovery Using Automation, Physics Modeling, and Bayesian Machine Learning
No full textInternational audienceCombating climate change hinges on our ability to accelerate scientific breakthroughs, particularly in the development of future energy solutions. Luckily, artificial intelligence (AI) is making this possible by transforming every stage of materials development from conceptual design and synthesis to characterization and performance evaluation. In photovoltaics, AI is improving the reliability of mature technologies such as silicon and driving innovation in emerging solutions like perovskites, opening the door to higher efficiencies and broader deployment scenarios. In this work, we demonstrate how combining Bayesian machine learning with physics-based modeling not only accelerates scientific discovery but also improves the interpretability of black-box optimization in automated fabrication platforms. First, we showcase how our probabilistic Bayesian framework (Chakar et al., 2024, 10.1016/j.solener.2024.112595), combined with drift-diffusion modeling, offers a physically grounded alternative to conventional techniques by linking performance and stability metrics to interpretable parameters such as charge carrier mobility and defect densities. We highlight how this approach – which has been successfully validated for identifying photovoltaic degradation pathways, linking indoor and outdoor aging, and assessing passivation effects – can be extended to provide actionable feedback for automated platforms for the synthesis and characterization of novel perovskite solar cells. Specifically, we show how connecting batch-to-batch variability and performance improvements to fundamental cell properties helps pinpoint the automation steps that need further refinement. The ability to capture prediction uncertainty makes this approach particularly well-suited for multi-solution problems such as identifying performance bottlenecks and optimizing processes, where multiple routes can yield comparable outcomes. However, as this requires extensive simulations, Bayesian optimization (BO) emerges as a practical solution for navigating the search space more intelligently. Nonetheless, BO has its own set of challenges, particularly due to the curse of dimensionality and high computational cost when exploring large parameter spaces. We therefore introduce SCORE (Chakar, 2024, 10.48550/arXiv.2406.12661), a technique that can overcome these limitations and tackle key problems in solar energy research and beyond, providing a resource-efficient and accessible solution for a broad range of researchers. Finally, we describe how these tools will be integrated into a high-throughput platform being established at IPVF to support automated perovskite thin-film fabrication, characterization, and stability testing. Bringing together diverse research efforts, the platform will enable precise control over processing and measurement workflows, generating reproducible, high-quality data to accelerate joint research efforts across the perovskite community
Rationally designed universal passivator for high-performance single-junction and tandem perovskite solar cells
No full textInternational audienceInterfacial trap-assisted nonradiative recombination hampers the development of metal halide perovskite solar cells (PSCs). Herein, we report a rationally designed universal passivator to realize highly efficient and stable single junction and tandem PSCs. Multiple defects are simultaneously passivated by the synergistic effect of anion and cation. Moreover, the defect healing effect is precisely modulated by carefully controlling the number of hydrogen atoms on cations and steric hindrance. Due to minimized interfacial energy loss, L-valine benzyl ester p-toluenesulfonate (VBETS) modified inverted PSCs deliver a power conversion efficiency (PCE) of 26.28% using vacuum flash processing technology. Moreover, by suppressing carrier recombination, the large-area modules with an aperture area of 32.144 cm 2 and perovskite/Si tandem solar cells coupled with VBETS passivation deliver a PCE of 21.00% and 30.98%, respectively. This work highlights the critical role of the number of hydrogen atoms and steric hindrance in designing molecular modulators to advance the PCE and stability of PSCs
Conception et synthèse de dérivés du ganglioside GM3 en tant que candidats vaccins anticancéreux
No full textGlycosylation is a prevalent post-translational modification of cell-surface proteins and lipids, essential for cellular recognition, signaling, and immune modulation. In cancer, glycosylation patterns are often dysregulated, leading to the overexpression of tumor-associated carbohydrate antigens (TACAs). These aberrant structures provide targets for synthetic vaccine development, aiming to induce immune memory and improve tumor recognition. Advances in glycan synthesis have significantly enhanced the design and immunogenicity of carbohydrate-based cancer vaccines. Ganglioside GM3 (Neu5Ac-α2,3-Gal-β1,4-Glc-β1-ceramide) is widely distributed in neuronal and epithelial membranes and regulates cell growth, adhesion, apoptosis, and signaling. Notably, GM3 is overexpressed in various tumors, including melanoma, breast cancer, and lung cancer. It is also one of the nine TACAs that have been progressed into clinical trials by the FDA. In our previous studies, a series of GM3 analogues were synthesized by replacing the glycosidic units excluding sialic acid. These non-natural glycan structures demonstrated promising anti-tumor activities by in vitro assays. In this study, a series of allyl glycosides (glucose, galactose, mannose, and lactose) were synthesized. Selective protection strategies were applied to expose either the C-3 or C-6 hydroxyl groups to serve as acceptors for sialylation. The key glycosylation step employed sialyl xanthate as glycosylation donor to afford six products in good yields (41~65%). Side reactions arising from allyl electrophilic addition were minimized by careful control of promoter addition. All new compounds were purified by chromatography and fully characterized using high-resolution mass spectrometry (HRMS) and nuclear magnetic resonance (NMR) spectroscopy. Subsequently, the allyl glycosides were converted to their corresponding glycolaldehydes via ozonolysis and conjugated to carrier protein BSA and KLH. The conjugation degree was determined using MALDI-TOF-MS or specific colorimetric assays. Notably, even among structural isomers, significant differences in conjugation efficiency were observed. To further investigate this phenomenon, α- and β-glycolaldehyde glycosides of five biologically relevant monosaccharides (Man, Glc, Gal, GlcNAc, and GalNAc) were synthesized and conjugated to proteins at a fixed sugar-to-protein molar ratio. Continuous kinetic monitoring revealed that anomeric configuration markedly influences both reaction rates and final sugar loading: α-Man > β-Man; β-Gal/GalNAc > their α-counterparts; Glc and GlcNAc showed no significant anomeric differences. NMR analysis identified cyclic intermediates that reduce the effective concentration of reactive aldehyde, while DFT calculations provided complementary insight into anomer-specific electrophilic character. This work presents a systematic study on the synthesis of GM3 derivatives and the conjugation of glycolaldehyde to protein, providing valuable information for design, synthesis, and conjugation of GM3 analogues and their vaccine application.La glycosylation est une modification post-traductionnelle fréquente des protéines et lipides membranaires, essentielle à la reconnaissance cellulaire, à la signalisation et à la modulation immunitaire. Dans les cancers, les profils de glycosylation sont souvent altérés, conduisant à la surexpression d'antigènes tumoraux glucidiques (TACAs). Ces structures aberrantes constituent des cibles pour le développement de vaccins synthétiques, visant à induire une mémoire immunitaire et à améliorer la reconnaissance tumorale. Les avancées en synthèse glycannique ont considérablement renforcé la conception et l'immunogénicité des vaccins anticancéreux à base de glucides. Le ganglioside GM3 (Neu5Ac-α2,3-Gal-β1,4-Glc-β1-céramide) est largement distribué dans les membranes neuronales et épithéliales, où il régule la croissance cellulaire, l'adhésion, l'apoptose et la signalisation. Il est notoirement surexprimé dans divers types de tumeurs, notamment le mélanome, le cancer du sein et le cancer du poumon. GM3 fait également partie des neuf TACAs ayant atteint les essais cliniques approuvés par la FDA. Dans nos travaux précédents, une série d'analogues de GM3 a été synthétisée par remplacement des unités glycosidiques autres que l'acide sialique. Ces structures glycosidiques non naturelles ont démontré une activité antitumorale prometteuse lors de tests in vitro. Dans cette étude, une série de glycosides allyliques (glucose, galactose, mannose et lactose) a été synthétisée. Des stratégies de protection sélective ont été appliquées pour exposer sélectivement les groupes hydroxyles en C-3 ou C-6 servant de récepteurs à la sialylation. L'étape clé de glycosylation a utilisé un donneur sialyl xanthate permettant d'obtenir six produits avec de bons rendements (41-65%). Les réactions secondaires impliquant l'addition électrophile sur le groupement allyle ont été minimisées par un contrôle rigoureux de l'addition du promoteur. Tous les nouveaux composés ont été purifiés par chromatographie et pleinement caractérisés par spectrométrie de masse à haute résolution (HRMS) et spectroscopie RMN. Les glycosides allyliques ont ensuite été convertis en leurs dérivés glycolaldéhydiques par ozonolyse, puis conjugués aux protéines porteuses BSA et KLH. Le degré de conjugaison a été déterminé par MALDI-TOF-MS ou par des dosages colorimétriques spécifiques. Il est à noter que, même entre isomères de structure, des différences significatives d'efficacité de conjugaison ont été observées. Afin d'explorer plus en détail ce phénomène, des glycosides glycolaldéhydiques α et β de cinq monosaccharides d'importance biologique (Man, Glc, Gal, GlcNAc et GalNAc) ont été synthétisés et conjugués à des protéines selon un rapport molaire sucre/protéine fixe. Un suivi cinétique en continu a révélé que la configuration anomérique influençait fortement à la fois la vitesse de réaction et la charge finale en sucres : α-Man > β-Man ; β-Gal/GalNAc > leurs α-contreparties; Glc et GlcNAc ne montraient pas de différence anomérique significative. L'analyse RMN a mis en évidence des intermédiaires cycliques réduisant la concentration effective en aldéhyde réactif, tandis que des calculs DFT ont apporté un éclairage complémentaire sur le caractère électrophile spécifique à chaque anomère. Ce travail présente une étude systématique sur la synthèse de dérivés de GM3 et la conjugaison de glycolaldéhydes à des protéines, apportant des informations utiles pour la conception, la synthèse et l'application vaccinale d'analogues de GM3