Spectroscopy-Guided Deep Learning Predicts Solid-Liquid Surface Adsorbate Properties in Unseen Solvents.

Accurately and rapidly acquiring the microscopic properties of a material is crucial for catalysis and electrochemistry. Characterization tools, such as spectroscopy, can be a valuable tool to infer these properties, and when combined with machine learning tools, they can theoretically achieve fast and accurate prediction results. However, on the path to practical applications, training a reliable machine learning model is faced with the challenge of uneven data distribution in a vast array of non-negligible solvent types. Herein, we employ a combination of the first-principles-based approach and data-driven model. Specifically, we utilize density functional theory (DFT) to calculate theoretical spectral data of CO-Ag adsorption in 23 different solvent systems as a data source. Subsequently, we propose a hierarchical knowledge extraction multiexpert neural network (HMNN) to bridge the knowledge gaps among different solvent systems. HMNN undergoes two training tiers: in tier I, it learns fundamental quantitative spectra-property relationships (QSPRs), and in tier II, it inherits the fundamental QSPR knowledge from previous steps through a dynamic integration of expert modules and subsequently captures the solvent differences. The results demonstrate HMNN's superiority in estimating a range of molecular adsorbate properties, with an error range of less than 0.008 eV for zero-shot predictions on unseen solvents. The findings underscore the usability, reliability, and convenience of HMNN and could pave the way for real-time access to microscopic properties by exploiting QSPR.

Blocking mitochondrial Zn2+ accumulation after ischemia reduces mitochondrial dysfunction and neuronal injury.

Zn2+ is an important contributor to ischemic brain injury and recent studies support the hypothesis that mitochondria are key sites of its injurious effects. In murine hippocampal slices (both sexes) subjected to oxygen glucose deprivation (OGD), we found that Zn2+ accumulation and its entry into mitochondria precedes and contributes to the induction of acute neuronal death. In addition, if the ischemic episode is short (and sublethal), there is ongoing Zn2+ accumulation in CA1 mitochondria after OGD that may contribute to their delayed dysfunction. Using this slice model of sublethal OGD, we have now examined Zn2+ contributions to the progression of changes evoked by OGD and occurring over 4-5 hours. We detected progressive mitochondrial depolarization occurring from ∼ 2 hours after ischemia, a large increase in spontaneous synaptic activity between 2-3 hours, and mitochondrial swelling and fragmentation at 4 hours. Blockade of the primary route for Zn2+ entry, the mitochondrial Ca2+ uniporter (MCU; with ruthenium red, RR) or Zn2+ chelation shortly after OGD withdrawal substantially attenuated the mitochondrial depolarization and the changes in synaptic activity. RR also largely reversed the mitochondrial swelling. Finally, using an in vivo rat (male) asphyxial cardiac arrest (CA) model of transient global ischemia, we found that ∼8 min asphyxia induces considerable injury of CA1 neurons 4 hours later that is associated with strong Zn2+ accumulation within many damaged mitochondria. These effects were substantially attenuated by infusion of RR upon reperfusion. Our findings highlight mitochondrial Zn2+ accumulation after ischemia as a possible target for neuroprotective therapy.SIGNIFICANCE STATEMENT:Brain ischemia is a leading cause of mortality and long-term disability that still lacks effective treatment. After transient ischemia delayed death of neurons occurs in vulnerable brain regions. There is a critical need to understand mechanisms of this delayed neurodegeneration which can be targeted for neuroprotection. We found progressive and long-lasting mitochondrial Zn2+ accumulation to occur in highly vulnerable CA1 neurons after ischemia. Here we demonstrate that this Zn2+ accumulation contributes strongly to deleterious events occurring after ischemia including mitochondrial dysfunction, swelling and structural changes. We suggest that this mitochondrial Zn2+ entry may constitute a promising target for development of therapeutic interventions to be delivered after termination of an episode of transient global ischemia.

Electric Dipole Descriptor for Machine Learning Prediction of Catalyst Surface-Molecular Adsorbate Interactions.

The challenge of evaluating catalyst surface-molecular adsorbate interactions holds the key for rational design of catalysts. Finding an experimentally measurable and theoretically computable descriptor for evaluating surface-adsorbate interactions is a significant step toward achieving this goal. Here we show that the electric dipole moment can serve as a convenient yet accurate descriptor for establishing structure-property relationships for molecular adsorbates on metal catalyst surfaces. By training a machine learning neural network with a large data set of first-principles calculations, we achieve quick and accurate predictions of molecular adsorption energy and transferred charge. The training model using NO/CO@Au(111) can be extended to study additional substrates such as Au(001) or Ag(111), thus exhibiting extraordinary transferability. These findings validate the effectiveness of the electric dipole descriptor, providing an efficient modality for future catalyst design.

Realizing a Not-Strong-Not-Weak Polarization Electric Field in Single-Atom Catalysts Sandwiched by Boron Nitride and Graphene Sheets for Efficient Nitrogen Fixation.

Developing efficient single-atom catalysts (SACs) for nitrogen fixation is of great importance while remaining a great challenge. The lack of an effective strategy to control the polarization electric field of SACs limits their activity and selectivity. Here, using first-principles calculations, we report that a single transition metal (TM) atom sandwiched between hexagonal boron nitride (h-BN) and graphene sheets (namely, BN/TM/G) acts as an efficient SAC for the electrochemical nitrogen reduction reaction (NRR). These sandwich structures realize stable and tunable interfacial polarization fields that enable the TM atom to donate electrons to a neighboring B atom as the active site. As a result, the partially occupied pz orbital of a B atom can form B-to-N π-back bonding with the antibonding state of N2, thus weakening the N≡N bond. The not-strong-not-weak electric field on the h-BN surface further promotes N2 adsorption and activation. The NRR catalytic activity of the BN/TM/G system is highly correlated with the degree of positively polarized charges on the TM atom. In particular, BN/Ti/G and BN/V/G are identified as promising NRR catalysts with high stability, offering excellent energy efficiency and suppression of the competing hydrogen evolution reaction.

Using Machine Learning to Predict the Dissociation Energy of Organic Carbonyls.

Bond dissociation energy (BDE), an indicator of the strength of chemical bonds, exhibits great potential for evaluating and screening high-performance materials and catalysts, which are of critical importance in industrial applications. However, the measurement or computation of BDE via conventional experimental or theoretical methods is usually costly and involved, substantially preventing the BDE from being applied to large-scale and high-throughput studies. Therefore, a potentially more efficient approach for estimating BDE is highly desirable. To this end, we combined first-principles calculations and machine learning techniques, including neural networks and random forest, to explore the inner relationships between carbonyl structure and its BDE. Results show that machine learning can not only effectively reproduce the computed BDEs of carbonyls but also in turn serve as guidance for the rational design of carbonyl structure aimed at optimizing performance.

Structure-dependent luminescence of tetra-(4-pyridylphenyl)ethylene: a first-principles study.

Luminescent materials with tunable wavelengths have been attracting much attention due to their many promising applications. In the present work, the recently reported supramolecular coordination complexes of tetra-(4-pyridylphenyl)ethylene (TPPE) that produce variable-wavelength light emissions were investigated by time-dependent density functional theory (TDDFT) calculations. We discovered that variations in the luminescent wavelength of TPPE mainly depend on the structural deviations from molecular planarity, which affect the molecular orbital wavefunction distribution. An interesting trigonometric-functional relationship between the emission wavelength (the emission energy) and the dihedral angle defining deviations from molecular planarity was uncovered. The solvent effect was also considered to reveal the mechanism of the solvent-dependent fluorescence color. These findings may be helpful to rational molecular design for high-performance luminescent materials with tunable color.

Label-Free Data Mining of Scientific Literature by Unsupervised Syntactic Distance Analysis.

Label-free data mining can efficiently feed large amounts of data from the vast scientific literature into artificial intelligence (AI) processing systems. Here, we demonstrate an unsupervised syntactic distance analysis (SDA) approach that is capable of mining chemical substances, functions, properties, and operations without annotation. This SDA approach was evaluated in several areas of research from the physical sciences and achieved performance in information mining comparable to that of supervised learning, as shown by its satisfactory scores of 0.62-0.72, 0.60-0.82, and 0.86-0.95 in precision, recall, and accuracy, respectively. We also showcase how our approach can assist robotic chemists programmed to perform research focused on double-perovskite colloidal nanocrystals, gold colloidal nanocrystals, oxygen evolution reaction catalysts, and enzyme-like catalysts by designing materials, formulations, and synthesis parameters based on data mined from 1.1 million literature references.

Machine Learning Descriptors for Data-Driven Catalysis Study.

Traditional trial-and-error experiments and theoretical simulations have difficulty optimizing catalytic processes and developing new, better-performing catalysts. Machine learning (ML) provides a promising approach for accelerating catalysis research due to its powerful learning and predictive abilities. The selection of appropriate input features (descriptors) plays a decisive role in improving the predictive accuracy of ML models and uncovering the key factors that influence catalytic activity and selectivity. This review introduces tactics for the utilization and extraction of catalytic descriptors in ML-assisted experimental and theoretical research. In addition to the effectiveness and advantages of various descriptors, their limitations are also discussed. Highlighted are both 1) newly developed spectral descriptors for catalytic performance prediction and 2) a novel research paradigm combining computational and experimental ML models through suitable intermediate descriptors. Current challenges and future perspectives on the application of descriptors and ML techniques to catalysis are also presented.

Catalytic Descriptor Exploration for Ru-Based Fischer-Tropsch Catalysts: Effect of Chlorine and Sulfur Addition.

As an important factor in the design of catalysts, catalytic descriptor exploration has emerged as a novel frontier in heterogeneous catalysis. Here, the underlying structure-activity relationships of Ru-based catalysts are theoretically studied to shed light on this area. Calculations of different competing reaction paths suggest that the HCO*-mediated path─because of two synergistic active sites─is more favorable than others. In addition, compared to unadulterated Ru catalysts, the presence of Cl enhances the hydrocarbon production, whereas the presence of S decreases it. After a systematic examination of a series of structure-activity relationships (42 in total), we found that both charge transfer and average charge difference of active Ru atoms are good descriptors for the binding stability of reactants. However, for reactivity the Gibbs free energy of the reactants performs better. More interestingly, due to the quite different catalytic processes of the dissociation and hydrogenation steps, their correlations have opposite slopes.

Mitochondrial dysfunction in the liver of type 2 diabetic Goto-Kakizaki rats: improvement by a combination of nutrients.

Treatment with a combination of four nutrients, i.e. R-α-lipoic acid, acetyl-l-carnitine, nicotinamide and biotin, just as with pioglitazone, significantly improves glucose tolerance, insulin release, plasma NEFA, skeletal muscle mitochondrial biogenesis and oxidative stress in Goto-Kakizaki (GK) rats. However, it is not known whether treatment with these nutrients can improve mitochondrial function and reduce oxidative stress in GK rats. The effects of a combination of these four nutrients on mitochondrial function, oxidative stress and apoptosis in GK rat liver were investigated. Livers of untreated GK rats showed (1) abnormal changes in the activities of mitochondrial complexes (decreases in I, III and IV and increases in II and V), (2) increases in protein oxidation, (3) decreases in antioxidant enzymes (superoxide dismutase, glutathione S-transferase, NADH-quinone oxidoreductase-1), (4) a decrease in total antioxidant capacity but increases in reduced glutathione level and glyceraldehyde 3-phosphate dehydrogenase expression and (5) significant increases in apoptosis biomarkers, including expression of p21 and p53. A 3-month treatment with the four nutrients significantly improved most of these abnormalities in GK rats, and the effects of the nutrient combination were greater than those of pioglitazone for most of these indices. These results suggest that dietary supplementation with nutrients that are thought to influence mitochondrial function may be an effective strategy for improving liver dysfunction in GK diabetic rats.

Enabling Efficient Charge Separation for Optoelectronic Conversion via an Energy-Dependent Z-Scheme n-Semiconductor-Metal-p-Semiconductor Schottky Heterojunction.

Achieving good charge separation while maintaining energetic electronic states in heterostructures is a challenge in designing efficient photocatalyst materials. Using first-principles calculations, we propose a Z-scheme Sn-m-Sp (n-semiconductor-metal-p-semiconductor) heterojunction as a viable avenue for achieving broad-spectrum sunlight absorption and, importantly, energy-dependent charge separation. As a proof-of-concept investigation, we investigated two ternary heterostructures, CdS-Au-PdO and SnO2-W-Ag2O, in which the electronic Fermi levels line up by virtue of the presence of an intermediate metal layer. A cascade of work functions in the relative order Wn < Wm < Wp drives electrons flowing from Sn to m and from m to Sp. The inner electric fields established at the Sn-m and m-Sp Schottky junctions selectively guide low-energy photoexcited electrons from Sn (CdS/SnO2) and low-energy holes from Sp (PdO/Ag2O) to the interposing Au or W metal, respectively. Importantly, relatively low Schottky barriers enforce charge separation by constraining high-energy photogenerated charges to the individual semiconductor layers. Operating together, these two mechanisms enable the achievement of highly efficient optoelectronic conversion.

Tuning Phase Transitions in Metal Oxides by Hydrogen Doping: A First-Principles Study.

Optimizing physical or chemical properties via electronic phase transitions is important for developing high-performance semiconductor materials. Using first-principles calculations, we established hydrogen doping (H-doping) as an alternative strategy for modulating phase transitions in metal oxide semiconductors. We found that H-doping can induce an insulator-to-metal phase transition in rutile TiO2 or wurtzite ZnO, a metal-to-insulator phase transition in rutile VO2, and sequential insulator-metal-insulator phase transitions in SnO2. H-doping not only creates defect-induced states inside the original bandgap but also induces new occupancy patterns in conduction band edge states. A linear relationship between occupation of the conduction band edge and H-doping concentration was found, offering the possibility of precisely tuning the material's free-carrier concentration, electrical conductance, and photoabsorption ability. We envision that this work will provide a promising way to tune electrical phase transitions in metal oxides.

Carbon Monoxide Oxidation Promoted by Surface Polarization Charges in a CuO/Ag Hybrid Catalyst.

Composite structures have been widely utilized to improve material performance. Here we report a semiconductor-metal hybrid structure (CuO/Ag) for CO oxidation that possesses very promising activity. Our first-principles calculations demonstrate that the significant improvement in this system's catalytic performance mainly comes from the polarized charge injection that results from the Schottky barrier formed at the CuO/Ag interface due to the work function differential there. Moreover, we propose a synergistic mechanism underlying the recovery process of this catalyst, which could significantly promote the recovery of oxygen vacancy created via the M-vK mechanism. These findings provide a new strategy for designing high performance heterogeneous catalysts.

Synergistic Effect of Surface-Terminated Oxygen Vacancy and Single-Atom Catalysts on Defective MXenes for Efficient Nitrogen Fixation.

The production of ammonia (NH3) from molecular dinitrogen (N2) under ambient conditions is of great significance but remains as a great challenge. Using first-principles calculations, we have investigated the potential of using a transition metal (TM) atom embedded on defective MXene nanosheets (Ti3-xC2Oy and Ti2-xCOy with a Ti vacancy) as a single-atom electrocatalyst (SAC) for the nitrogen reduction reaction (NRR). The Ti3-xC2Oy nanosheet with Mo and W embedded, and the Ti2-xC2Oy nanosheet with Cr, Mo, and W embedded, can significantly promote the NRR while suppressing the competitive hydrogen evolution reaction, with the low limiting potential of -0.11 V for W/Ti2-xC2Oy. The outstanding performance is attributed to the synergistic effect of the exposed Ti atom and the TM atom around an extra oxygen vacancy. The polarization charges of the active center are reasonably tuned by the embedded TM atoms, which can optimize the binding strength of key intermediate *N2H. The good feasibility of preparing such TM SACs on defective MXenes and the high NRR selectivity with regard to the competitive HER suggest new opportunities for driving NH3 production by MXene-based SAC electrocatalysts under ambient conditions.

A Hydrogenated Metal Oxide with Full Solar Spectrum Absorption for Highly Efficient Photothermal Water Evaporation.

Searching for cost-effective photothermal material that can harvest the full solar spectrum is critically important for solar-driven water evaporation. Metal oxides are cheap materials but cannot cover the full solar spectrum. Here we prepared a hydrogenated metal oxide (H1.68MoO3) material, in which H-doping causes the insulator-to-metal phase transition of the originally semiconductive MoO3. It offers a blackbody-like solar absorption of ≥95% over the entire visible-to-near-infrared solar spectrum, owing to its unusual quasi-metallic energy band, and high solar-to-heat conversion rate due to quick relaxation of excited electrons. Using a self-floating H1.68MoO3/airlaid paper photothermal film, we achieved a stable and high water vapor generation rate of 1.37 kg m-2 h-1, a superb solar-to-vapor efficiency of 84.8% under 1 sun illumination, and daily production of 12.4 L of sanitary water/m2 from seawater under natural sunlight. This thus opens a new avenue of designing cost-effective photothermal materials based on metal oxides.

Mitochondrial nutrients improve immune dysfunction in the type 2 diabetic Goto-Kakizaki rats.

The development of type 2 diabetes is accompanied by decreased immune function and the mechanisms are unclear. We hypothesize that oxidative damage and mitochondrial dysfunction may play an important role in the immune dysfunction in diabetes. In the present study, we investigated this hypothesis in diabetic Goto-Kakizaki rats by treatment with a combination of four mitochondrial-targeting nutrients, namely, R-alpha-lipoic acid, acetyl-L-carnitine, nicotinamide and biotin. We first studied the effects of the combination of these four nutrients on immune function by examining cell proliferation in immune organs (spleen and thymus) and immunomodulating factors in the plasma. We then examined, in the plasma and thymus, oxidative damage biomarkers, including lipid peroxidation, protein oxidation, reactive oxygen species, calcium and antioxidant defence systems, mitochondrial potential and apoptosis-inducing factors (caspase 3, p53 and p21). We found that immune dysfunction in these animals is associated with increased oxidative damage and mitochondrial dysfunction and that the nutrient treatment effectively elevated immune function, decreased oxidative damage, enhanced mitochondrial function and inhibited the elevation of apoptosis factors. These effects are comparable to, or greater than, those of the anti-diabetic drug pioglitazone. These data suggest that a rational combination of mitochondrial-targeting nutrients may be effective in improving immune function in type 2 diabetes through enhancement of mitochondrial function, decreased oxidative damage, and delayed cell death in the immune organs and blood.

Role of Hydrogen Bonding in Green Fluorescent Protein-like Chromophore Emission.

The fluorescence emission from green fluorescent protein (GFP) is known to be heavily influenced by hydrogen bonding between the core fluorophore and the surrounding side chains or water molecules. Yet how to utilize this feature for modulating the fluorescence of GFP chromophore or GFP-like fluorophore still remains elusive. Here we present theoretical calculations to predict how hydrogen bonding could influence the excited states of the GFP-like fluorophores. These studies provide both a new perspective for understanding the photophysical properties of GFP as well as a solid basis for the rational design of GFP-based fluorophores.

Protective effects of R-alpha-lipoic acid and acetyl-L-carnitine in MIN6 and isolated rat islet cells chronically exposed to oleic acid.

Mitochondrial dysfunction due to oxidative stress and concomitant impaired beta-cell function may play a key role in type 2 diabetes. Preventing and/or ameliorating oxidative mitochondrial dysfunction with mitochondria-specific nutrients may have preventive or therapeutic potential. In the present study, the oxidative mechanism of mitochondrial dysfunction in pancreatic beta-cells exposed to sublethal levels of oleic acid (OA) and the protective effects of mitochondrial nutrients [R-alpha-lipoic acid (LA) and acetyl-L-carnitine (ALC)] were investigated. Chronic exposure (72 h) of insulinoma MIN6 cells to OA (0.2-0.8 mM) increased intracellular oxidant formation, decreased mitochondrial membrane potential (MMP), enhanced uncoupling protein-2 (UCP-2) mRNA and protein expression, and consequently, decreased glucose-induced ATP production and suppressed glucose-stimulated insulin secretion. Pretreatment with LA and/or ALC reduced oxidant formation, increased MMP, regulated UCP-2 mRNA and protein expression, increased glucose-induced ATP production, and restored glucose-stimulated insulin secretion. The key findings on ATP production and insulin secretion were verified with isolated rat islets. These results suggest that mitochondrial dysfunction is involved in OA-induced pancreatic beta-cell dysfunction and that pretreatment with mitochondrial protective nutrients could be an effective strategy to prevent beta-cell dysfunction.

Melatonin and the aging brain.

The events associated with brain aging are enumerated with emphasis on increased oxidative and inflammatory processes and on mitochondrial dysfunction. Several of these factors are further increased in a wide range of overt age-related neurological diseases. This generality has given impetus to concepts concerning similar therapeutic approaches common to a series of neurodegenerative disorders. Animal and cell culture models of several such disorders have benefited from the application of melatonin. The mechanisms underlying the neuroprotective properties of melatonin are likely to involve activation of specific melatonin receptors. This can lead to modulation of transcription factors and consequent altered gene expression, resulting in enhancement of antioxidant enzymes and downregulation of basal levels of inflammation. Melatonin has potential utility both in slowing normal brain aging and in treatment of neurodegenerative conditions. This is reinforced by the low cost of melatonin and its very low toxic hazard.

Effects of melatonin and age on gene expression in mouse CNS using microarray analysis.

The expression levels of a number of genes associated with inflammation and immune function change with advancing age. Melatonin modulates gene expression levels of several of these genes. Therefore the declining levels of melatonin associated with age may play a role in the physiological effects of aging. We used oligonucleotide microarrays to measure age-related changes in mRNA expression in the murine CNS, and to study the effect of prolonged administration of dietary melatonin upon these changes. CB6F1 male mice were fed 40 ppm melatonin for 2.1 months prior to sacrifice at age 26.5 months, and compared with both age-matched controls and young, 4.5-month-old untreated controls. Total RNA was extracted from whole brain (excluding cerebellum and brain stem) and individual samples were hybridized to Affymetrix Mouse 430-2.0 arrays. The expression of a substantial number of genes was modulated by melatonin treatment and changes in selected genes were validated by quantitative reverse transcription polymerase chain reaction (qRT-PCR). A subset of these genes did not change with age. Conversely, some genes modulated by age were also modulated by melatonin treatment. In general, melatonin treatment drove the expression levels of these genes closer to the expression levels detected in the younger animals. Notably, the abundance of lipocalin 2 (Lcn2) mRNA increased with age and was decreased in old animals treated with melatonin. Lcn2 is a member of the acute phase response family of proteins and its mRNA levels in the brain increase in response to inflammation. Many of the genes with expression reduced by melatonin are involved in inflammation and the immune system. This suggests that melatonin treatment may influence the inflammatory responses of old animals, driving them to resemble more closely those occurring in young animals.