Portail HAL UHA (Université de Haute-Alsace)
Not a member yet
17715 research outputs found
Sort by
Thermal Behavior, Local-Scale Morphology, and Phase Composition of Spherulites in Melt-Crystallized Poly(Vinylidene Fluoride) Films
International audienceSynchrotron microbeam X-ray diffraction was employed to investigate the local-scale structure and solid-state phase transformation within individual spherulites of poly(vinylidene fluoride) (PVDF). In thin, non-oriented films, PVDF crystallizes into α and γ-phases, forming distinct spherulitic morphologies: large, banded α-spherulites and smaller, irregular “mixed” spherulites dominated by the γ-phase. For samples crystallized at high undercooling (160 °C), the mixed spherulites primarily consisted of the γ-phase, with only a minor fraction of α-lamellae localized at the spherulite boundaries. At higher crystallization temperatures (165 °C), the α-phase was entirely absent from the mixed spherulites. High-temperature annealing induced a phase transformation from the α-phase to the γ-phase, initiating at the interface between α- and γ-spherulites. The transformation propagated radially along the b-axis of the α-spherulite, while its characteristic banded morphology remained intact. Radial scanning with an X-ray microbeam provided spatially resolved mapping of the structural transition within the α-spherulite at the micrometer scale, offering detailed insights into the transformation mechanism and its impact on the spherulitic structure. The fast crystal growth direction remained unaltered during the transition, suggesting minimal material transport and maintaining structural coherence
Il “mistilinguismo” di Alice Ceresa: le lingue e l’identità
International audienceThe article explores Alice Ceresa’s reflections on the relationship between language and identity, starting from her personal experience as an Italian-speaking Swiss who grew up in a German-speaking environment and later lived in France and Italy. Through her diaries and writings, the article identifies three key moments: code-switching in her early texts, the formulation of the term “mistilinguismo” in 1985, and her mature re-reading of a bilingual childhood. Ceresa challenges the commonly positive view of multilingualism, presenting it instead as a source of identity tension. The concept of “mistilinguismo”, which she defines as the disharmonious coexistence of multiple mother tongues, becomes a metaphor for personal and cultural fragmentation. The article highlights how Ceresa, while rejecting theoretical frameworks, develops a profound awareness of the connection between language, belonging, and writing, offering an original perspective on the complexity of linguistic and cultural identity within and beyond the Swiss context.L’articolo esplora la riflessione di Alice Ceresa sul rapporto tra lingua e identità, a partire dalla sua esperienza personale di svizzera italofona cresciuta in un contesto germanofono e, in seguito, in Francia e in Italia. Attraverso i suoi diari e scritti, l’articolo individua tre momenti chiave: il code-switching nei testi giovanili, la definizione di “mistilinguismo” nel 1985 e la rilettura in età matura della propria infanzia bilingue. Ceresa mette in discussione la visione positiva del plurilinguismo, considerandolo piuttosto una fonte di tensione identitaria. Il “mistilinguismo”, da lei definito come convivenza non armonica di più lingue materne, diventa metafora di una frammentazione personale e culturale. L’articolo evidenzia come Ceresa, pur rifiutando sistematizzazioni teoriche, sviluppi una profonda consapevolezza del legame tra lingua, appartenenza e scrittura, offrendo uno sguardo originale sulla complessità dell’identità linguistica e culturale nel contesto svizzero e oltre
How deep is your love? Explore user attachment to digital sharing platforms
International audienceWhile sharing platforms have shown their continued rise and growing importance in recent years, they now operate in competitive market conditions and a changing environment: scam issues (Vinted platform), legal evolution (Airbnb issue). Today, it is becoming essential for digital sharing platforms to develop strong and durable relationships with users. In this context, platform attachment is the key to success (Li et al., 2022). This study explores a concept widely used in marketing literature, brand attachment, within the context of sharing platforms. A qualitative study was conducted among users of the HomeExchange platform to identify andbetter understand the platform attachment construct. Our exploratory study reveals five user profiles and suggests that while user experiences impact satisfaction, which leads to platform attachment, user compatibility remains the key factor. This study contributes to the marketing literature by investigating the role of user attachment to sharing platforms and provides a holistic understanding of how platform attachment develops
Efficient photocatalytic removal of ciprofloxacin antibiotic by dual Z-scheme Ag<sub>3</sub>PO<sub>4</sub>/g-C<sub>3</sub>N<sub>4</sub>/Bi<sub>2</sub>WO<sub>6</sub> under solar irradiation
International audienceThis study reports the development of a novel Ag3PO4/g-C3N4/Bi2WO6 photocatalyst featuring a ternary heterostructure and a dual Z-scheme pathway mechanism. SEM and TEM analyzes confirmed the intimate contact between Ag3PO4, Bi2WO6, and g-C3N4, and hence the successful formation of the ternary heterojunction. Further, under natural solar light irradiation, the Ag3PO4/g-C3N4/Bi2WO6 photocatalyst, in comparison to single Ag3PO4, g-C3N4 and Bi2WO6 component, achieved an improved and remarkable CIP antibiotic photodegradation efficiency of 94 % within 90 min. Reactive Oxygen Species (ROS) •OH and •O2− played the dominant roles in the photocatalytic reaction. The overall data suggest that the dual Z-scheme carrier transfer pathways are responsible for the superior photocatalytic activity as resulting from the efficient separation of photogenerated charge carriers while maintaining their strong redox capabilities. The Ag3PO4/g-C3N4/Bi2WO6 photocatalyst demonstrates promising potential for wastewater treatment applications due to its high efficiency, efficient charge separation, and generation of potent oxidizing species. Additionally, Monte Carlo simulation was used to calculate the interactions of Ag3PO4, Bi2WO6, and CIP, occurring on, respectively, g-C3N4, Ag3PO4/g-C3N4, and Ag3PO4/g-C3N4/Bi2WO6, in order to investigate their configurations and adsorption energies. The most stable geometries and pollutant remove were assessed from the negative adsorption energies (Ead) value
Les valeurs multizétas tridendriformes et dendriformes par les arbres de Schroeder
To build new generalisations of Multiple Zeta Values, we define new spaces of formal series and formal integrals. We show that they are tridendriform and dendriform algebras. This allows us to reinterpret the fact that Multiple Zeta Values are algebra morphisms for shuffles of words in terms of finer tridendriform and dendriform structures. Applying universal properties of Schroeder trees we obtain generalisations of Multiple Zeta Values that are algebra morphisms for associative products. Hence we find new properties of Arborified Zeta Values and state how this enables the computation of some Shintani Zeta Values.Nous construisons de nouvelles généralisations des valeurs multizétas, nous définissons des nouveaux espaces de séries formelles et d'intégrales formelles. Nous montrons qu'il s'agit d'algèbres tridendriformes et dendriformes. Le fait que les fonctions multizétas soient des morphismes d'algèbres pour le produit de mélange des mots peut être réinterprété dans des structures tridendriformes et dendriformes. En appliquant des propriétés universelles propres aux arbres de Schroeder, nous obtenons des généralisations des fonctions multizétas qui restent des morphismes d'algèbres pour des produits associatifs. Ainsi, nous trouvons de nouvelles propriétés des multizétas arborifiées et nous remarquons que cela permet d'effectuer des calculs pour des fonctions multizétas de Shintani
Apprentissage profond pour l'analyse des séries temporelles avec application au mouvement humain
Time series, composed of regularly spaced data, are essential in fields such as medicine, telecommunications, and energy. Their analysis addresses various tasks: movement classification, recovery prediction, detection of similar behaviors, and data enrichment for research. Leveraging temporal dependencies requires tailored algorithms. Recently, deep learning, already successful in natural language processing and image classification, has proven effective for time series analysis. This thesis explores several tasks using deep learning techniques. We enhance classification with advanced models and propose a novel architecture achieving state-of-the-art performance. The challenge of limited labeled data is tackled through self-supervised learning. Our contributions are demonstrated through real-world applications: movement analysis for rehabilitation, data generation for cinema and gaming, and prototyping to enrich costly datasets. Finally, we assess the limitations of current methodologies and propose a robust, unified, and fair evaluation framework to strengthen the impact of new methodologies in time series analysis.Les séries temporelles, composées de données régulières, sont essentielles en médecine, télécommunications et énergie. Leur analyse permet de résoudre des tâches variées : classification de mouvements, prédiction de récupération, détection de comportements similaires ou enrichissement de données pour la recherche.Exploiter leurs dépendances temporelles nécessite des algorithmes adaptés. Récemment, l’apprentissage profond, performant en traitement du langage et classification d’images, a prouvé son efficacité pour ces analyses. Cette thèse explore plusieurs tâches en utilisant des techniques profondes. Nous améliorons la classification avec des modèles avancés et une architecture novatrice atteignant des performances de pointe. Le manque de données annotées est abordé via l’apprentissage auto-supervisé. Nos contributions sont démontrées par des applications réelles : analyse de mouvements pour la rééducation, génération de données pour le cinéma et les jeux, et prototypage pour enrichir des ensembles coûteux. Enfin, nous évaluons les limites des méthodologies actuelles et proposons un cadre d’évaluation robuste et équitable, visant à renforcer l’impact des nouvelles méthodologies en analyse de séries temporelles
cis and trans isomers of phospho-tubulin-associated unit proteins differently affect amyloid aggregation
International audienceThe co-localization of phospho-tau proteins (p-tau) and amyloid beta (Aβ) or alpha-synuclein (α-Syn) aggregates and their combined pathological impacts have been recognized in the brains of both human and animal models of Alzheimer's disease (AD) and Parkinson's disease (PD). The fibrillation of amyloidogenic proteins is a dynamic process that can be affected by the presence of other proteins and therefore, it's essential to reveal how the crossinteraction between Aβ or α-Syn with p-tau affects their amyloidogenesis. Using simulation studies, we showed that cis and trans conformations of phosphorylated tau (phosphorylated at Thr231) show different affinities to Aβ and α-Syn. Trans p-tau showed more favorable and stronger interaction with Aβ and α-Syn and accelerated their aggregation. Consistently, the experimental approaches demonstrated that trans p-tau considerably enhances the fibrillation of both Aβ (from ~7 h to ~30 min) and α-Syn (from ~10 h to ~1 h) in a concentration-dependent manner, whereas cis p-tau, a neurotoxic isomer and early driver of tauopathy, exhibited a less pronounced acceleratory effect on amyloid aggregation (from 7 h to 5 h for Aβ and from 7 h to 5 h for α-Syn)
Étude Comparative des CNNs et des Vision-Language Models pour la Classification d'Images de Graphiques
International audienceÉtude Comparative des CNNs et des Vision-Language Models pour la Classification d'Images de Graphique