Accepted
Abstract +
This study investigates gap solitons and their stability in Bose-Einstein condensates confined in Moiré optical lattices with distinct twisted angles. The results demonstrate that the twisted angle significantly modulates the Moiré periodicity and the flatness of low bands. For sufficiently large angular differences, smaller twisted angles generally lead to larger Moiré periods and flatter low bands, though this trend becomes less consistent at minimal angular differences. Moreover, smaller twisted angles yield more complex potential structures, which modify gap positions and widths, consequently affecting the properties of gap solitons. Using the Newton-conjugate gradient method, we identify various types of solitons in Moiré lattice with four different twisted angles, observing that solitons can exist over a broader range of potential depths at smaller twisted angles. The density distributions of solitons exhibit markedly different behaviors in different gaps: in the semi-infinite gap dominated by attractive interactions, deeper potentials lead to reduced soliton density, whereas in the first gap governed by repulsive interactions, deeper potentials enhance soliton density distributions. Linear stability analysis and nonlinear dynamical evolution results indicate that solitons found in the first gap(including both single-humped and multi-humped structures) demonstrate robust dynamical stability, whereas in the semi-infinite gap, single-humped solitons maintain good stability, while closely separated multi-humped in-phase solitons tend to be unstable, with enhanced stability observed for solitons located closer to the band edges. This work provides a theoretical foundation for manipulating nonlinear solitons in Moiré superlattices.
,
Abstract +
This study applies machine learning, specifically transfer learning with neural networks, to improve predictions of fission barrier heights and ground state binding energies of superheavy nuclei, which are crucial for calculating survival probabilities in fusion reactions. Transfer learning for neural networks proceeds in two stages: pre-training and fine-tuning, each driven by a distinct pre-training data set and target data set. In this work we split the pre-training data into 60 % for training and 40 % for validation, while the target data are partitioned into 20 % test, with the remaining 80 % further divided into 60 % training and 40 % validation. To construct the neural-network model we adopt the proton number Z and mass number A as the input layer, employ two hidden layers each containing 128 neurons with ReLU (Rectified Linear Unit) activation, and set the learning rate to 0.001. For the fission-barrier-height model, the pre-training dataset is either the FRLDM or the WS4 model data, and the experimental measurements serve as the target set. For the ground-state binding-energy model, we first form the residuals between WS4 predictions and the AME2020 evaluation, then separate these residuals into a light-nucleus subset and a heavy-nucleus subset according to proton number. The light-nucleus subset is used for pre-training and the heavy-nucleus subset for fine-tuning. After optimization, the root-mean-square error (RMSE) of the FRLDM barrier model falls from 1.03 MeV to 0.60 MeV, and that of the WS4 barrier model drops from 0.97 MeV to 0.61 MeV. For the binding-energy model, the RMSE decreases from 0.33 MeV to 0.17 MeV on the test set and from 0.29 MeV to 0.26 MeV on the full data set. We also present the performance of the fission-barrier model before and after refinement, together with the predicted barrier heights along the isotopic chains of the new elements Z = 119 and Z = 120, and analyzed the reasons for the differences in the results obtained by different models. We hope that these results are intended to provide a useful reference for future theoretical studies. The datasets in this paper are openly available at https://www.doi.org/10.57760/sciencedb.28388(Please use private access link https://www.scidb.cn/s/6fmeIz to access the dataset during the peer review process).
,
Abstract +
Octafluorocyclobutane (C4F8)-based fluorocarbon plasmas have emerged as the cornerstone of nanometre-scale etching and deposition in advanced semiconductor manufacturing, owing to their tunable fluorine-to-carbon (F/C) ratio, elevated density of reactive radicals, and superior material selectivity. In high-aspect-ratio pattern transfer, optical emission spectroscopy (OES) enables in-situ monitoring by correlating the density of morphology-determining radicals with their characteristic spectral signatures, thereby offering a viable pathway for the simultaneous optimisation of pattern fidelity and process yield. A predictive plasma model that integrates kinetic simulation with spectroscopic analysis is therefore indispensable.In this study, a C4F8/O2/Ar plasma model tailored for on-line emission-spectroscopy analysis is established. First, the comprehensive reaction mechanism is refined through a systematic investigation of C4F8 dissociation pathways and the oxidation kinetics of fluorocarbon radicals. Subsequently, radiative-collisional processes for the excited states of F, CF, CF2, CO, Ar and O are incorporated, establishing an explicit linkage between spectral features and radical densities. Under representative inductively coupled plasma (ICP) discharge conditions, the spatiotemporal evolution of the aforementioned active species is analysed and validated against experimental data. Kinetic back-tracking is employed to elucidate the formation and loss mechanisms of fluorocarbon radicals and ions, and potential sources of modelling uncertainty are discussed. This model holds promising potential for application in real-time OES monitoring during actual etching processes.
,
Abstract +
Covalent Organic Frameworks (COFs) are regarded as aclassofpromising Surface-Enhanced Raman Scattering (SERS) substrates, owing to their highly ordered porous structure, excellent molecular adsorption capacity, and structural stability have attracted widely attention. However, traditional COF materials lack plasmonic properties, making it difficult to achieve a high-intensity Raman enhancement effect, which limits their applicationin high-sensitivity detection. To address this issue, a novel ruthenium-based covalent was choosen. Organic framework composite material (Ru-COF) was designed and fabricated in this study for constructing high-performance SERS-active substrates. By directly incorporating ruthenium complexes into the COF skeleton, astable Ru–N/Ocovalent coordination structure was formed, which effectively improved the loading capacity and dispersibility of ruthenium, while significantly enhancing the electromagnetic field coupling strength and electron transfer capability ofthesubstrate.Compared with pure COFs, the Ru-COF substrate exhibited excellentSERS response performance in the detection of MethyleneBlue (MB) molecules. Specifically,it achieved a low limit ofdetection (LOD) down to10 12 mol·L 1,alinearcorrelation coefficient (R2) ofno less than 0.99, and a high enhancement factor (EF) of up to 1.83×101. Additionally, the substrate showed good signal reproducibility(relative standard deviation, RSD < 5%) and retained over 90% o its initial signal intensity even after exposure to air for four months, demonstrating outstanding stability and durability. Further application studies indicated that the Ru-COF substrate could still realize stable detection of trace MB molecules in complex water samples, with the LOD remaining at the1012 mol·L1 level, along with excellent anti-ioninterference ability and signal consistency. This suggests that the substrate notonlyexhibits exceptional sensitivity and reproducibility under standard conditions but also holds potential for high-sensitivity quantitative detection in real environmental samples.The designstrategyofthismaterialprovidesanewresearchdirectionformetal-organic synergistically enhanced SERS systems and lays a crucial foundation for their practical applications in fields such as environmental pollutant detection,foodsafety analysis, and clinical diagnosis.
,
Abstract +
High-pressure science has emerged as one of the core frontiers in exploring novel states of matter and phenomena under extreme conditions. In high-pressure environments, the in situ detection of physical quantities such as magnetic fields and pressure is crucial for understanding material behavior under extreme conditions. However, conventional high-pressure magnetic sensing techniques often face challenges such as low spatial resolution, poor sensitivity, and difficulties in achieving in situ magnetic detection.
In recent years, quantum sensors based on solid-state color centers—such as nitrogen-vacancy centers in diamond, silicon-vacancy/double-vacancy centers in silicon carbide, and color centers in hexagonal boron nitride—have enabled high-pressure quantum metrology with micrometer-scale spatial resolution, high sensitivity, and superior in situ detection capabilities, offering innovative solutions for high-pressure research.
This review systematically summarizes the effects of extreme high-pressure conditions on the optical and spin properties of these solid-state defects. Furthermore, taking high-pressure magnetic phase transition studies in magnetic materials and Meissner effect measurements in superconductors as examples, we highlight recent advances in in situ magnetic sensing using solid-state color centers under high pressure. This overview aims to provide technical guidance for the future development of high-pressure quantum precision measurement techniques based on solid-state defects.
In recent years, quantum sensors based on solid-state color centers—such as nitrogen-vacancy centers in diamond, silicon-vacancy/double-vacancy centers in silicon carbide, and color centers in hexagonal boron nitride—have enabled high-pressure quantum metrology with micrometer-scale spatial resolution, high sensitivity, and superior in situ detection capabilities, offering innovative solutions for high-pressure research.
This review systematically summarizes the effects of extreme high-pressure conditions on the optical and spin properties of these solid-state defects. Furthermore, taking high-pressure magnetic phase transition studies in magnetic materials and Meissner effect measurements in superconductors as examples, we highlight recent advances in in situ magnetic sensing using solid-state color centers under high pressure. This overview aims to provide technical guidance for the future development of high-pressure quantum precision measurement techniques based on solid-state defects.
,
Abstract +
Photodetectors play an essential role in optical communications, environmental monitoring, and medical imaging, and their performance strongly depends on the properties of the optoelectronic materials. Therefore, the exploration of high-performance optoelectronic materials has long been a research focus in the field of materials science. Viologen-based organic materials, owing to their unique redox and chromic characteristics, have been extensively utilized in electrochromic devices, biosensors, and flow batteries. In this work, a viologen complex containing the transition metal element Co, {[Co(BPYBDC) (H2O)5]·(BDC)·H2O} (denoted as 1-Co) was designed and successfully synthesized. A series of in-situ high-pressure characterization techniques were employed to systematically investigate its structural and optoelectronic behaviors. The results reveal that 1-Co crystallizes in the Pc space group and remains structurally stable up to 11.6 GPa without any phase transition. UV-visible absorption spectroscopy shows a red-shift of the absorption edge upon compression, accompanied by a color change from colorless and transparent to yellow, indicating a pressure-induced narrowing of the optical bandgap. Consistent with the bandgap narrowing, impedance measurements demonstrate a significant reduction in the total resistance under compression, which remains about two orders of magnitude lower than the initial value after decompression. Furthermore, the photocurrent response is markedly suppressed under compression and barely recovers upon pressure release. This behavior can be attributed to the enhanced recombination of electrons with viologen groups under compression, leading to the formation of stable viologen radical states. These localized radicals cannot effectively participate in the separation and transport of photogenerated carriers, thereby contributing little to the photocurrent. These findings suggest that high pressure effectively modulates the optical and electrical behaviors of 1-Co by tuning intermolecular interactions and the electronic band structure, providing valuable insights into the pressure-dependent behavior of viologen-based materials.
,
Abstract +
The uncertainty principle, a cornerstone of quantum mechanics, has evolved from a fundamental limitation into a manageable resource in quantum information science. Precise control over quantum uncertainty is crucial for ensuring the security of quantum cryptography and the advantage of quantum computation. This work investigates the control of the quantum-memory-assisted entropic uncertainty relation in a noisy two-particle qutrit system, using quantum feedback control as a suppression strategy. In our model, Bob prepares a system AB composed of two V-type three-level atoms and sends atom A to Alice. Atom A interacts with a bimodal dissipative cavity. To suppress decoherence, a photodetector is used to monitor the dissipative cavity. Once a photon is detected, a local quantum feedback control is applied to atom A. Meanwhile, Bob’s atom B is assumed to be isolated from the noisy environment. To quantify the uncertainty, we select two incompatible observables, Sx and Sz, corresponding to the spin-1 components. We analyze the evolution of the entropic uncertainty and its lower bound, with the system initialized in two distinct states: an excited state and a maximally entangled state. Our findings demonstrate that applying appropriate quantum feedback control to the system can significantly suppress decoherence, leading to a marked reduction in both the entropic uncertainty and its lower bound. Through numerical simulations, we identify the optimal feedback strength for minimizing the entropic uncertainty and its lower bound to be p=2 for both initial states. Furthermore, examination of the system’s steady-state behavior after prolonged evolution reveals a key insight: under optimal feedback, the initial maximally entangled state evolves into a state with maximal classical correlation. Although no quantum correlation exists in this steady state, the strong classical correlation provides Bob with partial information about atom A, thereby enhancing his prediction accuracy for the measurement outcomes and leading to the observed reduction in the entropic uncertainty. Additionally, we explore the dynamics of the system’s purity. The results show a clear negative correlation, indicating that the reduction in entropic uncertainty is directly attributable to the purification of the system effected by the feedback control. In conclusion, this study establishes quantum feedback control as an effective theoretical protocol for suppressing the entropic uncertainty in realistic noisy environments. It provides a viable pathway for manipulating quantum uncertainty to enhance the robustness and performance of quantum information processing tasks.
,
Abstract +
The neutron total cross section is fundamental nuclear data crucial for the design of nuclear energy systems and research in nuclear physics. For graphite, an important reactor moderator, significant discrepancies exist among the major evaluated nuclear data libraries concerning its high-energy neutron total cross section, particularly in the resonance structures and the regions above 20 MeV. These uncertainties constrain the precise design of advanced nuclear systems. To resolve these controversies and provide benchmark experimental data, this study performed a high-accuracy measurement of the neutron total cross section of natural carbon from 0.3 eV to 50 MeV using the transmission method combined with the time-of-flight (TOF) technique. The experiment was conducted at the back-streaming white neutrons (Back-n) at the China Spallation Neutron Source (CSNS), utilizing the NTOX spectrometer equipped with a multi-cell fission chamber. The neutron emission time (t0) was precisely calibrated using the prompt γ-flash from the spallation reaction. The 77-meter flight path was accurately calibrated using the known standard fission resonance peaks of 235U at 8.774 eV, 12.386 eV, and 19.288 eV. For the energy region above 100 keV, a Bayesian iterative algorithm was applied to unfold the double-bunch problem, effectively resolving the overlap of TOF spectra from neutrons produced in different beam bunches. The experimental results show excellent agreement with the ENDF/B-Ⅷ.1 evaluation and existing experimental data in the EXFOR database within the 0.3 eV-100 keV region. Owing to the high statistical accuracy, approximately 97.6% of the data points have statistical uncertainties of less than 1%, with the vast majority of total uncertainties better than 2%, significantly reducing the uncertainty level in this energy region. In the 100 keV– 50 MeV energy range, the data align with the overall trends observed in mainstream evaluation databases. No significant resonance effect was detected at 4.93 MeV, providing high-quality reference data for clarifying the resonance structure at this energy point. Systematically evaluated data above 20 MeV are currently only available in JENDL-5. The measurement results of this work provide indispensable high-quality experimental data to fill the high-energy data gap and to drive updates of the relevant evaluated libraries. This study not only provides critical benchmark data for the international nuclear data re-evaluation, especially for the CENDL-3.2 library which currently lacks complete data for natural carbon, but also systematically validates the methodological reliability of obtaining wide-energy-range, high-precision neutron total cross section data at the CSNS Back-n beamline.
,
Abstract +
Accurate prediction of the state of health (SOH) of lithium-ion batteries in electric vehicles is crucial for ensuring the safety of drivers and passengers, optimizing battery management systems (BMS), and extending battery life. Reliable SOH prediction underpins essential BMS functions, including charge–discharge control, remaining useful life (RUL) prediction, and fault diagnosis. However, existing data-driven methods still face two long-standing challenges. First, most models rely excessively on a large number of handcrafted health features derived from voltage, current, and capacity data, resulting in feature redundancy and low computational efficiency. Second, SOH-related time series exhibit strong nonlinearity and non-stationarity, causing traditional neural networks to suffer from prediction drift, instability, and performance degradation under varying conditions. To address the first challenge, we propose a lightweight health feature selection mechanism that combines incremental capacity analysis (ICA) with correlation analysis to automatically identify compact and physically meaningful degradation features. Only four key health features that are highly correlated with capacity fading are selected, which effectively reduces model complexity and computational cost while maintaining high SOH prediction accuracy. To overcome the second challenge, we develop a hybrid neural network model (KanFormer) integrating the Kolmogorov–Arnold (KAN) representation theory with a Transformer-based temporal modeling framework for accurate and robust SOH prediction. Specifically, the proposed KanFormer framework consists of three hierarchical modules: (1) the local feature extraction module, which leverages the smooth interpolation capability of KAN to capture fine-grained degradation characteristics from voltage–capacity and incremental capacity (IC) curves, modeling local nonlinear behaviors in the degradation process; (2) the global feature extraction module, which employs a multi-head Transformer encoder to learn long-range dependencies and cross-scale temporal relationships, enabling the joint modeling of short-term dynamics and long-term aging evolution; and (3) the prediction output module, which uses nonlinear KAN layers to adaptively fuse local and global representations, producing numerically stable and highly accurate SOH prediction results. By combining the mathematical expressiveness of KAN with the temporal reasoning capability of the Transformer, KanFormer effectively mitigates prediction drift and oscillations induced by data nonlinearity and non-stationarity. Compared with conventional deep-learning models, the proposed method improves training efficiency by 15.32%. Experimental validation on three publicly available battery-aging datasets—Michigan Formation, HNEI, and NASA—demonstrates its superior performance, achieving MSE = 0.0045, MAE = 0.041, R2 = 0.978 on the Michigan dataset, MSE = 0.00055, MAE = 0.0175, R2 = 0.996 on the HNEI dataset, and MSE = 0.0056, MAE = 0.017, R2 = 0.984 on the NASA dataset. These results substantially outperform mainstream baselines, confirming the high accuracy and robustness of KanFormer. In summary, KanFormer unifies lightweight feature selection, nonlinear functional representation, and cross-scale temporal modeling, providing a scalable and interpretable solution for high-accuracy and high-efficiency SOH prediction.
Abstract +
Based on the generalized reduced R-matrix theory, this work performs a comprehensive analysis of all available experimental data for the 6He system using the RAC (R-matrix Analysis Code). A complete set of evaluated nuclear data has been obtained for major reaction channels induced by triton beams in the energy range of 10-2 ~ 20 MeV. The evaluated integral cross sections include T(t,2n)4He, T(t,n)5He, and T(t,d)4H reactions; the differential cross sections include T(t,2n)4He, T(t,n)5He, T(t,d)4H, and T(t,t)T. The evaluation results show good agreement with experimental data and the evaluated data of ENDF/B-VIII.1. In particular, for the T(t,2n)4He reaction, the evaluated cross sections are consistent with existing experiments over the full energy range, and a resonance dominated by the 2+ level is observed near 2.9 MeV. At 1.9 MeV, where experimental measurements of both integral cross sections and angular distributions are available, the evaluation reproduces both observables well. The combined constraint of integral and differential data significantly improves the stability of R-matrix parameters and the reliability of the evaluation. Based on the global analysis of the 6He system, this work also provides supplementary cross section data for the T(t,n)5He and T(t,d)4H reactions. The results contribute to the nuclear data foundation for fusion-related reactions and lay the groundwork for future joint evaluation with the mirror 6Be system.
The datasets presented in this paper, including the ScienceDB, are openly available at https://www.doi.org/10.57760/sciencedb.j00213.00202 (Please use the private access link https://www.scidb.cn/s/7jMryq to access the dataset during the peer review process)
The datasets presented in this paper, including the ScienceDB, are openly available at https://www.doi.org/10.57760/sciencedb.j00213.00202 (Please use the private access link https://www.scidb.cn/s/7jMryq to access the dataset during the peer review process)
,
Abstract +
Covalent organic frameworks (COFs) have been a potential candidate for applications in photocatalysis due to its periodically porous structures and tunable structure. The COF skeletons consisted of different building blocks may result in different performance. Investigating the effects of different building blocks on energy levels and excitons for COF can provide some insight for designing excellent COF catalysts. Based on the first-principles many-body Green’s function theory, the electronic structures and optical properties of the three donor-acceptor COFs by employing the monomer 2,4,6-trimethyl-1,3,5-triazine (TMT) as the key acceptor subunit and the trigonal aldehyde monomers including the tris(4-formylphenyl) amine (TPA), 1,3,5-tris(4-formylphenyl) benzene (TFPB) and 2,4,6-tris(4-formylphenyl)-1,3,5-triazine (TFPT) as the donor subunit are calculated in this work. Regulation of the donor unit and interlayer interactions on the electronic structures and excitonic properties are analyzed. The results show that the valence band maximum (VBM) and conduction band minimum (CBM) energies of the system are varied by the alteration of donor subunit. From TPA to the TFPB or TFPT, the bandgaps of the system increase, the light absorption blue shift, and the exciton binding energies gradually increase. There is little effect on the band gap and excitation energy by replacing the TFPB with the TFPT. Among the three COFs, the positions of both CBM and VBM of the TFPT-TMT COF only match well with the chemical reaction potential of H2/H+ and O2/H2O, which is capable of photocatalytic overall water splitting. But the photocatalytic performance for the TFPT-TMT COF might be inhibited by the higher exciton binding energy. The exciton for the TPA-TMT COF is easier to separate according to the exciton distributions and the exciton binding energy. The effect of different building units on the electronic structure, excitation energy, and excitonic properties of COFs in monolayer COFs are in line with that in multilayer and bulk COFs. The variation of the energy levels and excitation energies of all the three COFs as the number of layers are consistent. With the increasing number of layers, the VBM and CBM shift up and down with respect to the vacuum level, respectively. The band gap gradually decreases. The energy tend to decrease slower with the more layer. The exciton energy for multilayer COFs is close to the bulk state. These results are significant to design and modify COFs.
,
Abstract +
In this paper, a novel scheme is proposed and experimentally demonstrated. It is based on a directly modulated laser (DML) and all-optical mode-locking for generating tunable microwave frequency combs (MFCs). Theoretical analysis reveals that harmonic or rational harmonic mode-locking can be achieved by adjusting the parameters of the fiber ring cavity, which enables the generation of MFCs with adjustable comb spacing. Based on this, experimental verification shows that the DML can be driven to exhibit various typical dynamical states under sinusoidal modulation with different frequencies and amplitudes. These states serve as seeding signals that subsequently undergo all-optical mode-locking within the ring laser cavity, resulting in the generation of MFCs. The bandwidths of the MFCs are 13, 15, 19.5, 19.8, and 22 GHz, respectively, all of which satisfy the ± 5 dB flatness criterion. A continuously tunable comb-spacing range of 200 MHz to 3 GHz is attained through the effective combination of the DML and all-optical mode-locking. The single-sideband (SSB) phase noise of the first comb line remains below −100 dBc/Hz at a 10 kHz offset. Theoretical analysis and experimental results demonstrate that the modulated signals of the proposed scheme support flexible parameter tuning over a wide range. Furthermore, the generated MFCs have remarkable advantages in flatness, bandwidth, and tunability.
,
Abstract +
Accelerating the application of lead-free inorganic halide perovskites in solar cells necessitates the development of novel perovskite materials with suitable bandgap widths, high stability, and environmental friendliness. This represents a crucial pathway for driving photovoltaic technology innovation and reducing reliance on conventional fossil fuels. However, traditional material development paradigms heavily depend on trial-and error experimental screening or pure density functional theory (DFT) calculations, which incur significant time and material costs.
To address these challenges, this study innovatively proposes and implements an efficient screening strategy based on the synergy between deep learning and DFT calculations. By constructing a database containing 1181 inorganic halide double perovskite materials, we systematically trained and compared the performance of five mainstream machine learning models for the bandgap prediction task: Random Forest Regression (RFR), Gradient Boosting Regression (GBR), Support Vector Regression (SVR), eXtreme Gradient Boosting Regression (XGBR), and a Deep Neural Network (DNN) model. Results demonstrate that the DNN model, leveraging its powerful nonlinear mapping capability and advantage in automatic high-dimensional feature extraction, achieved exceptional prediction accuracy on the test set, with the Mean Absolute Error (MAE) significantly reduced to 0.264 eV and the coefficient of determination (R2) reaching 0.925. Its performance was markedly superior to other compared models, highlighting the immense potential of deep learning in predicting complex material properties.
Using this optimized DNN model, this study successfully screened four promising inorganic double perovskite candidates from 55 potential materials: Cs2GaAgCl6, Cs2AgIrF6, Cs2InAgCl6, and Cs2AlAgBr6. Among them, Cs2AgIrF6 and Cs2AlAgBr6 performed particularly well, with predicted bandgaps of 1.36 eV and 1.20 eV, respectively. This range ideally matches the requirement for efficient light absorption in solar cells. Further device performance simulations revealed that the solar cell based on Cs2AgIrF6 achieved a simulated power conversion efficiency (PCE) of 23.71%, with an open-circuit voltage (VOC) of 0.94 V, a short-circuit current density (JSC) of 31.19 mA/cm2, and a fill factor (FF) of 80.81%. Cs2AlAgBr6 also exhibited a simulated efficiency of 22.37%, corresponding to VOC=0.78 V, JSC=36.73 mA/cm2, and FF=77.66%. Notably, both materials demonstrated high open-circuit voltages and fill factors, clearly indicating excellent carrier separation efficiency and significantly reduced nonradiative recombination losses within these materials.
In summary, this study successfully validates the significant efficacy of the deep learning-DFT synergistic strategy in accelerating the discovery of novel lead-free perovskite materials. This method not only substantially enhances the efficiency of DFT data analysis and the depth of pattern mining, overcoming some bottlenecks associated with traditional highthroughput calculations, but more importantly, it provides a practical and highly innovative approach for the rational design of high-performance, stable, and environmentally friendly lead-free perovskite solar cells, holding positive implications for advancing green, low-carbon energy technologies.
To address these challenges, this study innovatively proposes and implements an efficient screening strategy based on the synergy between deep learning and DFT calculations. By constructing a database containing 1181 inorganic halide double perovskite materials, we systematically trained and compared the performance of five mainstream machine learning models for the bandgap prediction task: Random Forest Regression (RFR), Gradient Boosting Regression (GBR), Support Vector Regression (SVR), eXtreme Gradient Boosting Regression (XGBR), and a Deep Neural Network (DNN) model. Results demonstrate that the DNN model, leveraging its powerful nonlinear mapping capability and advantage in automatic high-dimensional feature extraction, achieved exceptional prediction accuracy on the test set, with the Mean Absolute Error (MAE) significantly reduced to 0.264 eV and the coefficient of determination (R2) reaching 0.925. Its performance was markedly superior to other compared models, highlighting the immense potential of deep learning in predicting complex material properties.
Using this optimized DNN model, this study successfully screened four promising inorganic double perovskite candidates from 55 potential materials: Cs2GaAgCl6, Cs2AgIrF6, Cs2InAgCl6, and Cs2AlAgBr6. Among them, Cs2AgIrF6 and Cs2AlAgBr6 performed particularly well, with predicted bandgaps of 1.36 eV and 1.20 eV, respectively. This range ideally matches the requirement for efficient light absorption in solar cells. Further device performance simulations revealed that the solar cell based on Cs2AgIrF6 achieved a simulated power conversion efficiency (PCE) of 23.71%, with an open-circuit voltage (VOC) of 0.94 V, a short-circuit current density (JSC) of 31.19 mA/cm2, and a fill factor (FF) of 80.81%. Cs2AlAgBr6 also exhibited a simulated efficiency of 22.37%, corresponding to VOC=0.78 V, JSC=36.73 mA/cm2, and FF=77.66%. Notably, both materials demonstrated high open-circuit voltages and fill factors, clearly indicating excellent carrier separation efficiency and significantly reduced nonradiative recombination losses within these materials.
In summary, this study successfully validates the significant efficacy of the deep learning-DFT synergistic strategy in accelerating the discovery of novel lead-free perovskite materials. This method not only substantially enhances the efficiency of DFT data analysis and the depth of pattern mining, overcoming some bottlenecks associated with traditional highthroughput calculations, but more importantly, it provides a practical and highly innovative approach for the rational design of high-performance, stable, and environmentally friendly lead-free perovskite solar cells, holding positive implications for advancing green, low-carbon energy technologies.
,
Abstract +
Crystallization of ions in aqueous micro-droplet or nano-droplet on solid surfaces is ubiquitous, with applications ranging from inkjet printing to pesticide spraying. The substrates involved are typically nonpolar. Yet, the atomistic mechanism of crystallization within sessile droplets on such nonpolar substrates remains elusive. Here, we employ molecular dynamics simulations to investigate the crystallization of sodium chloride inside an aqueous nano-droplet resting on a nonpolar face-centered-cubic (111) surface. Crystallization occurs inside the droplet rather than at the liquid–gas or solid–liquid interface, when the concentration of the sodium chloride in the droplet exceeds 3.76 mol/kg. The phenomenon originates from the spatial distributions of water molecules and ions: a dense interfacial water layer forms at the solid–liquid interface, whereas ions accumulate in the droplet interior, increasing the local concentration. The ion–water hydration due to the electrostatic interaction dominates over ion–solid interaction. The spatial confinement provided by the solid, rather than the physical properties of the solid, enriches ions inside the nano-droplet and thereby triggers the crystallization. We further generalize this mechanism to the isolated aqueous sodium chloride nano-droplet, where the gas phase breaks the continuous spatial distribution of ions as that in the droplet. Analogous crystallization is observed for the sessile droplet of potassium chloride solution on nonpolar solid surfaces, indicating the generality of crystallization in nano-droplets. These findings offer atomic-scale guidance for controlling crystallization in nano-droplets relevant to microelectronics, inkjet printing and related technologies.
,
Abstract +
The study of unstable nuclei beyond the nucleon drip line is an important method to study the nuclear interaction and structure in the extremely neutron- or proton-rich system, and various nuclides beyond the proton drip line mainly decay by spontaneous one-proton emission. Using the deformed Woods-Saxon potential, spin-orbit potential and the expanded Coulomb potential to construct the daughter-proton potential, based on the quantum tunneling model and the microscopic Gamow state theory, the half-lives data of various proton emitters are systematically calculated. By using nuclear data from different source and comparing to experiments, the dependence of proton emission on decay energy and spectroscopic factors is evaluated. Additionally, based on previous observations, the half-life of the possibly lighter proton emitter in the fpg-shell below has been theoretically predicted. Our results are compiled into a comprehensive dataset of half-lives for both experimentally confirmed emitters (50 < Z < 84) and theoretically predicted emitters (30 < Z < 50), providing a useful reference for future experimental investigations related to the proton drip line. The datasets presented in this paper, including our results of calculation, are openly available at https://www.doi.org/10.57760/sciencedb.27551 (Please use the private access link https://www.scidb.cn/s/zQzA3e to access the dataset during the peer review process).
- 1
- 2
- 3
- 4
- 5
- ...
- 15
- 16
