Experimental Phase Research Articles

Electron probe microanalysis of trace sulfur in experimental basaltic glasses and silicate minerals

Abstract Sulfur (S) in the mantle is conventionally assumed to be exclusively stored in accessory sulfide phases, but recent work shows that the major silicate minerals that comprise >99% of the mantle could be capable of hosting trace amounts of S. Assessing the incorporation of trace S in nominally S-free mantle minerals and determining equilibrium S partitioning between these minerals and basaltic melt requires analyzing small experimental phases with low S contents. Here, we develop a protocol for EPMA analysis of the trace levels of S in silicate phases. We use a suite of natural and experimental basaltic glass primary and secondary standards with S contents ranging from 44 ppm to 1.5 wt%. The effects of beam current and counting time are assessed by applying currents ranging from 50 to 200 nA and total counting times between 200 and 300 s at 15 kV accelerating voltage. We find that the combination of 200 nA beam current with a 200 s counting time (80 s peak, 60 s each for upper and lower background, respectively) achieves precise yet cost-effective measurements of S down to a calculated detection limit of ~5 ppm and a blank-derived, effective detection limit of ~17 ppm. Close monitoring of the S peak intensity and position throughout the duration of each spot also shows that high currents and extended dwell times do not compromise the accuracy of measurements, and even low S contents of 44 ppm can be reproduced to within one standard deviation. Using our developed recipe, we analyzed a small suite of experimental clinopyroxenes (Cpx) and garnets (Gt) from assemblages of silicate partial melt + Cpx ± Gt ± sulfide, generated at 1.5 to 3.0 GPa and 1200 to 1300 °C. We find S contents of up to 71 ± 35 ppm in Cpx and 63 ± 28 ppm in Gt and calculate mineral-melt partition coefficients (Dsmin/melt) of up to 0.095 ± 0.064 and 0.110 ± 0.064 for DsCpx/melt and DsGt/melt, respectively. The sulfur capacity and mineral-partitioning for Cpx are in good agreement with SXRF measurements in a prior study by Callegaro et al. (2020), serving as an independent validation of our EPMA analytical protocol.

Optimization of the Disinfection Process in Potabilization Systems in Cuenca Alto Atoyac, Mexico

Access to drinking water in Cuenca Alto Atoyac (CAA), serving over 3.5 million people, is challenged by contamination and inadequate disinfection processes. This study investigates the use of oxidation–reduction potential (ORP) as a tool to optimize disinfection in CAA’s water systems. The research was conducted in three stages: an initial pilot phase with 10 sampled sources, expanding to 14 supply sources in the experimental phase, followed by analyses defining optimal chlorine dosing and contact times across two distinct hydrogeological zones. Principal component analysis (PCA) identified key factors, explaining 53.52% of water quality variability, while ORP measurements revealed significant differences in water quality among sources, with “Ameca” and “Pozo #2 Santa María Atlihuetzían” representing the extremes. Adjusted chlorine doses and contact times were tailored to local conditions, achieving target ranges and ensuring improved microbial control. In the final stage, significant correlations were established between ORP levels, chlorine doses, and total coliform concentration, supporting the effectiveness of ORP in guiding disinfection processes. Overall, chlorine doses of 0.3–1.5 mg/L with contact times around 15 min yielded ORP values (594–687 mV) optimal for CAA water safety.

Long Short-Term Memory Networks for the Automated Identification of the Stationary Phase in Tribological Experiments

This study outlines the development and optimization of a Long Short-Term Memory (LSTM) network designed to analyze and classify time-series data from tribological experiments, with a particular emphasis on identifying stationary phases. The process of fine-tuning key hyperparameters was systematically optimized through Bayesian optimization, coupled with K-fold cross-validation to minimize the inherent randomness associated with training neural networks. The refined LSTM network achieved a weighted average accuracy of 84%, demonstrating a high level of agreement between the network’s identified stationary phases and those manually determined by researchers. This result suggests that LSTM networks can reliably mimic manual identification processes in tribological data, providing a promising avenue for automating data analysis. The study underscores the potential of neural networks to transcend their traditional role in predictive modeling within tribology, opening up new possibilities for their application across a broader spectrum of tasks within the field.

Deep Learning for Automated 3D Floor Plan Generation

Abstract: This paper presents an optimized method for reconstructing 3D floor plans using userdefined boundaries and constraints for residential structures. This approach allows users to providearchitectural constraints such as room types and quantities, as well as manual sketches or existingimages of the house boundaries. Advanced deep learning algorithms are used to automaticallypartition the house boundaries and create a customized, optimized interior layout based on the users’architectural constraints. In the experiment phase, we integrated the Graph2Plan-based deep learningmodule, which converts the user-provided boundary data and architectural constraints into astructured 2D floor plan, automatically allocating and refining the rooms to ensure a harmoniousspatial arrangement. The evaluation of the deep learning model’s performance shows that this is auseful and time-saving solution for designers. Then, we utilized graphics and image processingtechniques to generate the 3D floor plans. Based on this solution, we have developed a 3D floor plangeneration application that provides a flexible and adaptive solution for individual home planningwithin defined boundaries. The application has been thoroughly tested to demonstrate its features,including the ability to meet users’ architectural constraints, provide rapid response times, and offera convenient user interaction experience.Keywords: 3DFloorplan, Floorplan generation, layout graph, RPLAN dataset, house plan.

A comprehensive study on methylene blue removal via polymer and protein nanoparticle adsorbents

Water pollution, particularly from industrial contaminants such as dyes, is a significant global concern. Various technologies, including nanoscale materials, are employed for water and wastewater treatment. Among these, adsorption process as an effective method due to its simplicity, cost-effectiveness, and reliability. This study comprised both theoretical and experimental phases. Initially, computer simulations were utilized to evaluate the interaction between methylene blue and three selected nanoparticles, ultimately choosing Bovine Serum Albumin protein nanoadsorbent based on energy considerations. Subsequently, adsorption experiments were conducted using this nanosorbent. The results indicated a maximum dye removal efficiency of 69% under the conditions of pH 11, an initial dye concentration of 100 mg/L, an adsorbent dose of 0.5 g/L, a contact time of 60 min, and an optimal temperature of 25 °C. The maximum adsorption capacity under optimal conditions was found to be 38.52 mg/g. Additionally, the adsorption isotherm followed the Langmuir equation, and the kinetics adhered to the pseudo-second-order model.

A comprehensive construction of deep neural network‐based encoder–decoder framework for automatic image captioning systems

AbstractThis study introduces a novel encoder–decoder framework based on deep neural networks and provides a thorough investigation into the field of automatic picture captioning systems. The suggested model uses a “long short‐term memory” decoder for word prediction and sentence construction, and a “convolutional neural network” as an encoder that is skilled at object recognition and spatial information retention. The long short‐term memory network functions as a sequence processor, generating a fixed‐length output vector for final predictions, while the VGG‐19 model is utilized as an image feature extractor. For both training and testing, the study uses a variety of photos from open‐access datasets, such as Flickr8k, Flickr30k, and MS COCO. The Python platform is used for implementation, with Keras and TensorFlow as backends. The experimental findings, which were assessed using the “bilingual evaluation understudy” metric, demonstrate the effectiveness of the suggested methodology in automatically captioning images. By addressing spatial relationships in images and producing logical, contextually relevant captions, the paper advances image captioning technology. Insightful ideas for future study directions are generated by the discussion of the difficulties faced during the experimentation phase. By establishing a strong neural network architecture for automatic picture captioning, this study creates opportunities for future advancement and improvement in the area.

Influence of the new wavy teat liner “Stimulor StressLess” on milk yield performance and its quality in dairy cows: Results of a field study

The ideal milking system meets the physiological needs of dairy cows to increase milk yield, achieve better milk quality and maintain healthy udders. Therefore, the settings of the milking machine and the properties of the teat cup liners are very important on dairy farms. The aim of the present study was to test a new teat cup liner “Stimulor StressLess” (SSL) in two commercial dairy farms and to investigate its influence on daily milk production and quality having different experimental settings. For this purpose, 40 dairy cows of different breeds in Tirol, Austria (farm 1) were investigated for 6 months, where 3 months represent the control phase (Gr 1) and milked with conventional teat liners, and the second 3 months phase (Gr 2) was the experimental phase and milked with SSL teat cup liners. On the second farm 90 dairy cows of Simmental breed in Baden-Württemberg, Germany (farm 2) were examined for one year equally divided in the first 6 months of control phase and second 6 months of treatment phase. All cows on both farms had the same stage of lactation and lactation number. During the study period, the daily milk production of each cow was recorded and milk samples were collected to determine the ingredients. The results showed that higher daily milk production and better milk quality were observed after using the new SSL teat cup liner in the existing milking machines. However, the use of the new teat cup liner SSL was more efficient in the longer treatment in farm 2 than in farm 1. In addition, after the use of the new SSL teat cup liner, the udders remained healthy throughout the study period, showing lower somatic cell counts (SCC). It can be concluded that high milk yield and better milk quality can be achieved by using SSL teat cup liners, as they are adapted to all teat shapes and dimensions.

Influence of the new wavy teat liner “Stimulor StressLess” on milk yield performance and its quality in dairy cows: Results of a field study

The ideal milking system meets the physiological needs of dairy cows to increase milk yield, achieve better milk quality and maintain healthy udders. Therefore, the settings of the milking machine and the properties of the teat cup liners are very important on dairy farms. The aim of the present study was to test a new teat cup liner “Stimulor StressLess” (SSL) in two commercial dairy farms and to investigate its influence on daily milk production and quality having different experimental settings. For this purpose, 40 dairy cows of different breeds in Tirol, Austria (farm 1) were investigated for 6 months, where 3 months represent the control phase (Gr 1) and milked with conventional teat liners, and the second 3 months phase (Gr 2) was the experimental phase and milked with SSL teat cup liners. On the second farm 90 dairy cows of Simmental breed in Baden-Württemberg, Germany (farm 2) were examined for one year equally divided in the first 6 months of control phase and second 6 months of treatment phase. All cows on both farms had the same stage of lactation and lactation number. During the study period, the daily milk production of each cow was recorded and milk samples were collected to determine the ingredients. The results showed that higher daily milk production and better milk quality were observed after using the new SSL teat cup liner in the existing milking machines. However, the use of the new teat cup liner SSL was more efficient in the longer treatment in farm 2 than in farm 1. In addition, after the use of the new SSL teat cup liner, the udders remained healthy throughout the study period, showing lower somatic cell counts (SCC). It can be concluded that high milk yield and better milk quality can be achieved by using SSL teat cup liners, as they are adapted to all teat shapes and dimensions.

Developing Semantic Textual Similarity for Guragigna Language Using Deep Learning Approach

Semantic Similarity is one of the highest levels of NLP. STS has significant advantages in NLP applications like information retrieval, information extraction, text summarization, data mining, machine translation, and other tasks. This research aims to present a deep learning approach for capturing semantic textual similarity (STS) in the Guragigna. The methodology involves collecting a Guragigna language corpus and preprocessing the text data and text representation is done using the Universal Sentence Encoder (USE), along with word embedding techniques including Word2Vec and GloVe and mean Square Error (MSE) is used to measure the performance. In the experimentation phase, models like LSTM, Bidirectional RNN, GRU, and Stacked RNN are trained and evaluated using different embedding techniques. The results demonstrate the efficacy of the developed models in capturing semantic textual similarity in the Guragigna language. Across different embedding techniques, including Word2Vec, GloVe, and USE, the Bidirectional RNN model with USE embedding achieves the lowest MSE of 0.0950 and the highest accuracy of 0.9244. GloVe and Word2Vec embedding also show competitive performance with slightly higher MSE and lower accuracy. The Universal Sentence Encoder consistently emerges as the top-performing embedding across all RNN architectures. The research results demonstrate the effectiveness of LSTM, GRU, Bi RNN, and Stacked RNN models in measuring semantic textual similarity in the Guragigna language.

Experimental electronic phase diagram in a diamond-lattice antiferromagnetic system

Experimental electronic phase diagram in a diamond-lattice antiferromagnetic system

Stress recognition identifying relevant facial action units through explainable artificial intelligence and machine learning

Background and objective: Facial cues and expressions constitute a component of bodily responses that provide useful information about one’s stress levels. According to the Facial Action Coding System, they can be modelled consistently in terms of fundamental facial muscle movements, called facial Action Units (AUs). This article investigates automatic acute stress recognition based on AUs using conventional Machine and Deep Learning techniques. Methods: We created a new experimental dataset containing 58 participants performing 4 experimental phases and 11 stress and non-stress tasks in which the proposed system performs automatic facial AUs recognition. A computational feature selection method was employed to select a robust relevant AU combinations subset, which integrated with conventional Machine Learning and Deep Learning methods using the Layer-Wise Relevance Propagation algorithm to assess and model the implication of AUs under acute stress conditions. Ordinal modelling was used following the pairwise transformation to establish a common reference based on the personalized values of each participant. Results: The results indicate that, under acute stress conditions, participants’ faces presented significantly more AUs and with greater intensity compared to neutral conditions. The most relevant combination of AUs to each stress type was computationally identified, ranked and selected. The mean yielded classification accuracy of stress condition versus neutral achieved across all experimental tasks was greater than 93%. Conclusions: There are specific combinations of AUs that are relevant to the stress conditions of each experimental phase leading in each case to better neutral and stress separability.

The effect of questioning gender stereotype threat on girl’s standing long jump performance

BackgroundOver the past few years, the sport psychology literature has established that gender stereotype threat (ST) is one of the factors that can impair girls’ performance. However, few studies have attempted to annihilate these negative effects. The purpose of the current study was to investigate whether questioning gender ST can mitigate the classical decline in girls’ standing long jump (SLJ) performance.MethodsThe participants were 120 girls (Mage = 10.74 ± 0.85 years), selected through convenience sampling and randomly assigned to three groups: the gender ST group (n = 40), the questioning group (n = 40), and the control group (n = 40). For all groups, baseline performance (i.e., SLJ) was measured by a female researcher following a warm-up period. In the experimental phase, the control group repeated the baseline conditions; the gender ST group completed the same test but was evaluated by a male examiner (i.e., implicit stereotype induction), while participants in the questioning group were assessed after receiving questioning statements while performing the task in front of a male examiner.ResultsThe results of the present study showed that the induction of a gender ST leads to a decrease in SLJ in girls. Additionally, if these inducing conditions of gender ST are accompanied by a questioning condition, the negative effects of gender ST can be reduced, and SLJ in girls does not decline.ConclusionBased on our findings, this intervention is recommended as a simple, inexpensive, and quick solution for mitigating the negative effects of gender ST on girl’s motor performance.

The Interaction of Microalgae Dietary Inclusion and Forage-to-Concentrate Ratio on the Lipid Metabolism-Related Gene Expression in Subcutaneous Adipose Tissue of Dairy Goats.

Long-chain polyunsaturated fatty acids (PUFAs) such as docosahexaenoic acid (DHA) have been shown to be effective in enhancing the quality of ruminant products, including meat and milk. However, under these dietary conditions, the de novo lipogenesis could be influenced, too. On the other hand, even if the forage-to-concentrate ratio (F:C) is also a key factor affecting lipid metabolism in small ruminants, there is scarce information about its interaction with dietary PUFA. This study investigates the potential of the F:C ratio as a lever to manipulate lipid metabolism in dairy goats under high dietary PUFA supplementation. For this purpose, twenty-two crossbred dairy goats [Alpine × Local (Greek) breeds] (BW = 50.6 ± 6.1 kg) at early lactation (70 ± 10 days in milk) during the age of 3-4 years old, were separated into two homogeneous subgroups (n = 11). In the first phase, each goat was fed 20 g Schizochytrium spp./day followed by either a high-forage (20 HF) or a high-grain (20 HG) diet, while in the second phase, each goat was fed 40 g Schizochytrium spp./day followed once again either a high-forage (40 HF) or a high-grain (40 HG) diet. The F:C ratio of a high-forage and high-grain diet was 60:40 and 40:60, respectively. Tail fat tissue samples were collected by biopsy on the 42nd day of each experimental phase (last day). Significant decreases (p < 0.05) in the gene expression of ACACA, CBR2, COX4I1, ELOVL5, ELOVL7, LEP, and SCD were presented in goats fed 40 g compared to those fed 20 g Schizochytrium spp., while the gene expression of ACACA, AGPAT2, AGPAT3, ELOVL5, ELOVL6, EPHX2, FASN, and SCD was decreased in high grain compared to high-forage diets. This study also indicated that with the aim to enrich goat products with PUFA by increasing their levels in the diet, lipid metabolism is negatively affected. However, a diet with higher forage inclusion can partially attenuate this condition.

CFD Two-Phase Blood Model Predicting the Hematocrit Heterogeneity Inside Fiber Bundles of Blood Oxygenators.

Blood is commonly treated as single-phase homogeneous fluid in numerical simulations of blood flow within fiber bundles of blood oxygenators. However, microfluidics tests revealed the presence of hematocrit heterogeneity in blood flowing across such geometries. Given the significant role of red blood cells (RBCs) in the oxygenation process, this study aims to propose a multiphase blood model able to correctly describe the experimental evidence and computationally investigate hematocrit heterogeneities inside fiber bundles. The experimental results of microfluidics tests performed in a previous study were processed and based on quantitative data of image intensity, a two-phase blood model following the Eulerian-Eulerian approach was calibrated and evaluated in its predictive ability against the experimental data. The two-phase model was then used to study the RBCs distribution inside different fiber bundles at average hematocrit values of 25% and 35%, representative of hemodilution in extracorporeal blood circulation. The numerical model proved to be able to describe and predict the experimental phase separation between plasma and RBCs within the microchannel geometry at different test conditions. Moreover, blood flow simulation in commercial fiber bundles revealed the presence of specific patterns in hematocrit distribution and their dependence on variations in bundle microstructure. The two-phase blood model proposed in this study provides a tool for advanced evaluation of local fluid dynamics and identification of optimal bundle microstructure allowing further gas transfer simulations to account for a reliable heterogeneous distribution of RBCs around the oxygenating fibers.

Multi-agent model-based intensification-driven tabu search for solving the dual-resource constrained flexible job shop scheduling

ABSTRACT In today's rapidly evolving manufacturing landscape, the efficient allocation of human resources is crucial for optimizing production systems. This paper addresses the dual resource-constrained flexible job shop scheduling problem (DRCFJSP) by implementing a Multi-Start Tabu Search-based Multi-Agent Model (MuSTAM). MuSTAM considers a set of initial solutions that are run in parallel using an intensification technique. The main objective of MuSTAM is to reduce the completion time (makespan), which is an important metric in workshops. The proposed model is composed of two classes of agents: a supervisor agent named MainAgent and a set of tabu search agents called TabuAgents. The MainAgent launches the system, collects inputs, generates the initial population, creates TabuAgents based on the number of solutions in the initial population (PopSize), and displays the best solution. Each TabuAgent receives a solution from the initial population and then conducts a concentrated search around its neighbourhood space by applying the intensification technique of tabu search. TabuAgents cooperate and communicate with each other to enhance search quality. Numerical tests are performed during the experimental phase to compare MuSTAM with ITS based on well-known benchmark instances of the DRCFJSP. The results demonstrate that MuSTAM is effective in terms of makespan and CPU time.

Effectiveness of Three Front-of-Pack Food Labels in Guiding Consumer Identification of Nutrients of Concern and Purchase Intentions in Kenya: A Randomized Controlled Trial.

Background: Front-of-pack-labels (FOPLs) on packaged foods provide essential information to help consumers make informed dietary choices. However, evidence on their effectiveness, particularly in low- and middle-income countries like Kenya, is limited. Objective: This study assessed the effectiveness of three FOPLs in helping consumers identify nutrients of concern in packaged food products and influencing their purchase intention in Kenya. Methods: A total of 2198 shoppers from supermarkets in Nairobi, Mombasa, Kisumu, and Garissa were randomized into three groups: Red and Green Octagon label (RG), Red and Green Octagon with icons (RGI), and Black Octagon Warning label (WL). In the control phase, participants were shown unlabeled images of packaged foods, followed by questions. In the experimental phase, the same images were presented with one assigned FOPL, and participants responded again to the same set of questions. Differences in correct identification of nutrients of concern and changes in purchase intention were analyzed using frequency tables and Chi-Square tests, while modified Poisson regression assessed FOPL effectiveness. Results: FOPLs significantly improved correct nutrient identification and reduced the intention to purchase unhealthy foods, with the WL proving most effective. Conclusions: These findings highlight the potential of FOPLs, particularly the WL, as an effective regulatory tool for promoting healthier food choices in Kenya.

High-Pressure Monosulfide Solid Solution FexNi1−xS Phases: X-Ray Diffraction Analysis and Raman Spectroscopy

High-pressure high-temperature (HPHT) crystalline monosulfide solid solution (Mss) phases (FexNi1−xS, x = 0.90, 0.75, 0.50, 0.25), troilite (FeS I), and α-NiS of the Fe-Ni-S system were synthesized at 7 GPa and 900–1550 °C. The structural parameters of the obtained phases were refined by XRD using the Rietveld method. Factor group analysis revealed the number of active Raman modes for FeS I and α-NiS. Raman spectra of troilite, α-NiS, and Mss phases were obtained. It was shown that the Raman spectra of Mss phases and α-NiS have a similar topology. The Raman spectra of the experimental phases in the Fe-Ni-S system were analyzed with non-negative matrix factorization, which provided a meaningful concentration dependence of the spectral patterns. The spectral components were assigned to the FeS I and α-NiS structures. The structural and spectroscopic studies show linear dependencies of unit cell parameters and spectral components on composition and confirm the existence of a series of monosulfide solid solution FexNi1−xS.

The effect of social anxiety on social attention in naturalistic situations

ABSTRACT Background and Objectives This multimodal two-phase study investigated the impact of trait social anxiety on exploration, social attention, and autonomic responses in a naturalistic setting. We expected higher avoidance of potentially crowded spaces, reduced visual attention on other people, and heightened physiological arousal in social situations for participants with higher social anxiety levels. Design and Methods Eighty-seven participants, pre-screened for high variance in trait social anxiety, first completed a half-hour walk on a freely chosen route and subsequently had a staged social interaction with a confederate consisting of a non-interactive waiting phase and a short conversation. Results While social anxiety did not modulate the choice of route during the walk phase, socially anxious participants avoided gazing at other individuals in non-interactive situations, i.e., during the walk and the waiting phase. In contrast, during actual interaction, they showed increased visual attention towards the confederate’s face. Across all experimental phases, highly socially anxious individuals showed elevated heart rates, but this effect was independent of the social context. Conclusions This study showed that social anxiety affects social exploration behavior not in a way of general avoidance, but rather in nuanced adaptations depending on the concrete situation, likelihood of interaction and associated socio-evaluative threat.

Reparameterization of the mW model to accurately predict the experimental phase diagram of methane hydrate.

Due to their high computational efficiency, the coarse-grained water models are of particular importance for practical molecular simulations of gas hydrates. In these models, the mW model is successfully used to study many thermodynamics and dynamics of methane hydrate. Yet, despite several decades of intense research, the mW model is still found to overestimate the melting temperature of methane hydrate. We here employ the minimum mean squared error estimation to revisit the key parameter of the mW model, which determines the strength of the tetrahedral angle of the water system. Relying on the free energy calculations, we first estimate the chemical potentials of water in the liquid phase for temperatures at which methane hydrate forms. We then turn to the mean squared error to describe the chemical potential deviation between the mW model and the TIP4P/ice model (the latter could reproduce the experimental phase diagram of methane hydrate). By minimizing the mean squared error, we finally have an optimized parameter for the mW model. In this part, we also discuss the pressure effect on such reparameterization procedure. Moreover, relying on the direct coexistence method, the melting temperature determined using the reparameterized mW model is found to be consistent with the experimental data. This strategy provides a means to improve the coarse-grained model to match the experimental observations for temperatures in the range of interest.

Understanding solid nitrogen through molecular dynamics simulations with a machine-learning potential

We construct a fast, transferable, general purpose, machine-learning interatomic potential suitable for large-scale simulations of N2. The potential is trained only on high quality quantum chemical molecule-molecule interactions; no condensed phase information is used. Although there are no explicit or implicit many-molecule interaction terms, the potential reproduces the experimental phase diagram including the melt curve and the molecular solid phases of nitrogen up to 10GPa. This demonstrates that many-molecule interactions are unnecessary to explain the condensed phases of N2. With increased pressure, transitions are observed from cubic (α), which optimizes quadrupole-quadrupole interactions, through tetragonal (γ), which allows more efficient packing, to monoclinic (λ), which packs still more efficiently. On heating, we obtain the hcp three-dimensional (3D) rotor phase (β) and, at pressure, the cubic δ phase which contains both 3D and 2D rotors, tetragonal δ* phase with 2D rotors, and the rhombohedral ε. Molecular dynamics demonstrates where these phases are indeed rotors, rather than frustrated order. The model supports the metastability of the complex ι phase, but not the reported existence of the wide range of bond lengths. The thermodynamic transitions involve both shifts of molecular centers and rotations of molecules: the onset of rotation is rapid, whereas motion of molecular centers is inhibited and we suggest that this is the cause of the experimentally observed sluggishness of transitions. Routine density functional theory calculations give a similar picture to the potential. Published by the American Physical Society 2024