Traditional material discovery methods， including experimental trial-and-error and high-throughput screening， are constrained by database scalability， hindering efficient exploration of the vast chemical space. Generative artificial intelligence （AI） is revolutionizing materials science by enabling a new paradigm for the inverse design of functional materials. This paper centers on the landmark work published in Nature—the MatterGen generative model， detailing its diffusion model-based approach for achieving stable and controllable inorganic crystal material generation. MatterGen not only generates diverse and stable crystal structures across the periodic table but also facilitates conditional generation with fine-tuning for target chemical compositions， spatial symmetries， and multiple performance constraints （e.g.， mechanical， electrical， and magnetic properties）. By examining the technical principles， performance advantages， and experimental validation of MatterGen， this paper illustrates how generative models are transforming material design from “screening” to “creation”， while also discussing the challenges and future development trends of this technology.

The detection of electromagnetic waves—including infrared rays， visible light， ultraviolet rays， X-rays， gamma rays—is one of the most essential approaches for humankind to understand the world. In the ultraviolet-visible light band， detection technologies are relatively mature due to the well-developed materials such as crystalline silicon. However， numerous challenges still persist in the detection of extremely short-wave X-rays and gamma rays as well as long-wave infrared rays. In the field of infrared detection， HgCdTe （MCT） focal plane array detectors are capable of capturing thermal imaging that is invisible to the naked eye. Nevertheless， they are unable to provide spectral information due to the lack of infrared spectroscopic devices at the front end. Therefore， developing optical crystals for infrared polarizing prisms and infrared tunable filter devices forms the foundation for advancing infrared multispectral and hyperspectral detection. In addition， for the detection of highly penetrating X-rays and gamma rays， semiconductor-based detection and scintillator-based detection are the most extensively studied approaches. However， the commercial semiconductors CdTe and CdZnTe （CZT） suffer from non-uniform defect distribution and high production costs， while traditional halide and oxide scintillators are hygroscopic， costly， or have insufficient energy resolution. Hg₂X₂ （X=Cl， Br， I） crystals have emerged as a promising multifunctional alternative， combining exceptional infrared polarizing optical properties， strong acousto-optic performance， and high-performance radiation detection capabilities. Hg₂X₂ crystals crystallize in the I4/mmm space group， featuring one-dimensional X-Hg-Hg-X chains stabilized by intra-chain covalent bonding and inter-chain van der Waals interactions. Halogen substitution from Cl to Br and I enables systematic tuning of key material parameters： lattice constants increase， band gaps decrease （from 2.9 to 2.1 eV）， infrared transparency extends up to 40 μm （in Hg₂I₂）， and both birefringence and the acousto-optic figure of merit improve significantly （2 600×1.5×10^-18 and 3 200×1.5×10^-18 s³/g for Hg₂Br₂ and Hg₂I₂， respectively）. Notably， researchers at Shandong University have successfully grown high-quality and large-sized Hg₂Cl₂ and Hg₂Br₂ crystals with diameters up to 54 mm via physical vapor transport （PVT） method by optimizing temperature gradient， growth rate， and cooling profile to suppress defects such as impurities and cloud-like inclusions. Hg₂Cl₂ crystals have been employed for fabricating various polarizing devices such as Glan-Taylor prisms and Wollaston prism， which are operable within the spectral range of 3 μm to 20 μm. They exhibit an extinction ratio of 30 000∶1 and a laser-induced damage threshold （LIDT） of 8.06 J/cm² at 3.5 μm. Moreover， the anti-reflection-coated polarization beam combiner （PBC） achieves combination efficiencies of 93.5% at 4.6 μm and 92.4% at 9.2 μm， with excellent beam quality （M²=1.23/1.17）. For acousto-optic applications， Hg₂Br₂-based acousto-optic tunable filter （AOTF） represents the only viable solution for MWIR/LWIR operation， demonstrating internal diffraction angles greater than 7° and a peak diffraction efficiency of 97.35%， achieved through engineered acoustic absorption layers. In radiation detection， Hg₂X₂-based detectors achieve an energy resolution of 1.35% for 662 keV γ-rays from ¹³⁷Cs. Furthermore， Hg₂Br₂ enables dual-mode detection of γ-rays and thermal neutrons due to the high neutron capture cross-section of ¹⁹⁹Hg， and has been applied in wearable α-particle detectors for real-time monitoring of water contamination. This review highlights the pivotal role of Hg₂X₂ crystals in the fields of infrared and nuclear radiation detection， and offers perspectives on the challenges associated with their crystal growth and device fabrication.

Diamond is a typical representative of ultra-wide bandgap semiconductors. Theoretically， it has the advantages of a large bandgap， extremely high thermal conductivity， and high carrier mobility， making it an ideal material for high-frequency， high-power， and high-temperature electronic devices. Achieving efficient and stable semiconductor doping is an inevitable requirement for the application of diamond semiconductor electronic devices. Currently， through surface modification methods such as hydrogen termination/silicon termination and boron doping， diamond has achieved relatively excellent p-type doping， and p-type devices have also continuously made new breakthroughs. However， suitable dopants or material modification methods for diamond n-type semiconductor doping have not yet been found， and it still faces problems such as low doping efficiency， high activation energy， and difficult material growth. This paper systematically reviews the research progress at home and abroad on achieving n-type semiconductor doping in diamond through single-element doping and multi-element co-doping methods， analyzes the advantages and disadvantages of various doping schemes， and looks forward to the development prospects of diamond n-type doping， hoping to provide a reference for solving the problem of diamond n-type semiconductor doping.

Perovskite single crystals are regarded as highly promising materials for optoelectronic devices， owing to their low defect density， outstanding photovoltaic characteristics， high resistance to moisture， and effective suppression of ion migration. However， the efficiency of perovskite single crystals-based devices in applications such as photovoltaics and light-emitting diodes has remained inferior to that of polycrystalline thin-film counterparts， primarily due to limitations in material processing and crystal growth quality. To address these challenges， a low-temperature crystal growth strategy was developed using 2-methoxyethanol （2ME） as the solvent. This approach was designed to minimize thermal fluctuations during the crystallization process， thereby suppressing the formation of defect states. In addition， owing to the low boiling point of 2ME， solvent retention on the perovskite surface was significantly reduced. When compared with MAPbI₃ single crystals prepared using conventional gamma-butyrolactone （GBL） solvent， the 2ME-grown crystals exhibited a 2.7-fold enhancement in photoluminescence （PL） intensity and a 1.5-fold extension of carrier lifetime. A reduction in defect state density by 16% was achieved， along with a 19% increase in the activation energy for ion migration. Furthermore， a self-powered photodetector fabricated from 2ME-derived MAPbI₃ single crystals demonstrated a responsivity of 0.55 A·W^-1 and a specific detectivity of 0.80×10¹³ Jones at 0 V bias. These values exceed those of devices based on GBL-grown crystals by factors of 1.72 and 1.59， respectively. The response speed was also markedly improved， with rise and fall times of 2.1 ms and 3.2 ms—representing a 54% and 49% enhancement over the GBL-based device （~4.6 ms rise， ~6.3 ms fall）. This study establishes a reliable low-temperature pathway for the production of high-quality， low-defect MAPbI₃ single crystals via 2ME. The findings indicate that this method effectively enhances the optoelectronic performance and stability of perovskite single crystals， facilitating their application in advanced devices such as high-sensitivity imaging systems， optical communication modules， and quantum detection platforms.

Cadmium zinc telluride（CZT）is a crucial material in the fields of infrared detection and nuclear radiation detection， garnering significant attention as a foundational material for manufacturing next-generation ultra-large array detectors. To address the technical challenges in growing 150 mm diameter CZT crystals via the Vertical gradient freeze （VGF） method， finite element simulation was employed to regulate the heater output power during the crystal growth process. This approach enabled precise control over the temperature field during CZT crystal growth， achieving a convex solid-liquid interface shape throughout the entire growth process. Consequently， high-quality 150 mm diameter CZT single crystals were successfully grown. The grown crystals enabled the preparation of 100 mm×100 mm infrared CZT substrates. Test results demonstrate uniform composition distribution in the CZT crystals. The full width at half maximum （FWHM） of the rocking curve for the （111） plane of the CZT substrates is below 15″， and the average etch pit density （EPD） is less than 1×10⁴ cm^-2.

Crystal products are prone to scratch defects during production， manufacturing， and processing， and accurate identification of such defects remains a technical challenge in this field. To address the urgent demand for multi-type scratch detection under complex imaging conditions， this paper proposes an improved approach based on YOLO11 by introducing a deformable attention transformer （DAT） and large separable kernel attention （LSKA）. The method improves adaptive modeling and recognition of surface defects by leveraging deformable convolutions and multi-scale receptive field fusion. Compared with traditional models such as YOLOv3， YOLOv5， YOLOv8， and the baseline YOLO11， the improved YOLO11-DAT_LSKA achieves performance improvements of 2.5， 2.3， 1.9 and 1.5 percentage points respectively in terms of mAP@0.5 for defect detection. These results demonstrate the effectiveness of the proposed method in enhancing feature modeling capabilities and improving the perception of complex scratches， thereby improving the detection accuracy for surface defects on crystal materials. In particular， to address low-contrast and irregular crystal surface defects， the DAT and LSKA modules enhance global contextual modeling and multi-scale receptive field adaptation， enabling effective capture of defect shape variations and multi-scale defect structures.

Under the background of the rapid development of third-generation semiconductors and the growing demand for large-size， low-warp 8-inch （1 inch=2.54 cm） 4H-SiC substrates， the traditional multi-wire sawing suffers from large material loss and limited capability in precisely controlling residual stress and wafer shape. However， the existing laser slicing studies are mostly focused on small-size or semi-insulating 4H-SiC， and a systematic understanding of the correlation among process parameters， residual stress， and wafer shape for conductive 8-inch 4° off-axis ingots under engineering conditions is still lacking. In this work， the 8-inch conductive N-type 4° off-axis 4H-SiC ingots are used as the research object. Internal modification is introduced by 1 064 nm picosecond laser irradiation combined with ultrasonic slicing， while the scanning speed （50~400 mm/s） is systematically varied. By means of SEM/TEM/SAED， Raman depth profiling， high-resolution XRD combined with the Stoney equation， lattice stress analysis and full-aperture wafer-shape measurements， a multi-scale evaluation chain is established from the microscopic modified layer to the macroscopic warp of the whole wafer. This paper clarifies the nucleation and propagation mechanism of SiC cracks along the ［11\bar{\mathrm{2}}0］ crystallographic direction induced by picosecond laser irradiation， and reveals influence of laser scanning speed on the slicing efficiency， stress distribution and wafer shape quality of 8-inch SiC substrates from both micro- and macro-perspectives. The paper provides theoretical support and process-optimization pathways for high-quality laser slicing fabrication of 8-inch 4H-SiC wafers， and offers a reference research idea for achieving controllable wafer shape tuning of SiC substrates based on laser-modified layers.

4H-SiC metal-oxide-semiconductor field-effect transistor （MOSFET） has become one of the most promising semiconductor devices due to their excellent properties such as high breakdown voltage， fast switching， and low loss. However， the yield and reliability of MOSFET can be degraded by morphological defects formed during the epitaxial process. In this study， the influence of three types of morphological defects on the electrical characteristics of 6.5 kV MOSFET was investigated. The results show that triangle defects can lead to early breakdown of the device， with a reverse breakdown voltage not exceeding 3 V. Furthermore， the presence of these defects in the active region also induces gate control failure. In contrast， triangle-like defects and linear defects have no significant impact on the breakdown voltage of the device， but the on-resistance of device with the triangle-like defect increases by 4.56%. Although triangle-like defects exhibit smaller band-gap shrinkage compared to linear defects， they introduce higher localized stress in their morphology and have a stacking fault area five times larger than that of linear defects， resulting in increased on-resistance of the device. In addition， both triangle-like defects and linear defects exhibit leakage current when a negative voltage was applied by the conductive atomic force microscope. Although neither type of defect caused significant changes in the static characteristics of the device， this behavior remains detrimental to long-term reliability.

Targeting the urgent demand for high-quality P-type highly doped epitaxial layers in high-voltage silicon carbide （SiC） power devices， this work systematically investigated the homoepitaxial growth technology of 8-inch （1 inch=2.54 cm） 4H-SiC based on trimethylaluminum （Al（CH₃）₃， TMA） precursor. By optimizing key parameters in the process of high-temperature chemical vapor deposition （CVD）， the controllable doping of aluminum （Al） in the epitaxial layer a doping concentration exceeding 1.00×10¹⁹ cm^-3 was successfully achieved on 8-inch and 4° off-axis 4H-SiC substrates， and the longitudinal doping uniformity in the epitaxial layer was good. The influence of doping concentration on defect morphology was analyzed by surface defect detection technology. The results show that when the Al doping concentration surpasses 1.35×10¹⁹ cm^-3， the lattice mismatch stress induces significant degradation of the surface morphology， and the degree of deterioration will increase with the increase of the Al doping concentration. Through further optimization of the growth conditions， the fatal defect density is successfully suppressed to 0.156 cm^-2 at a high doping concentration higher than 1.00×10¹⁹ cm^-3， so that the usable area ratio of 3 mm×3 mm chip area reaches 99.0%. This study provides an effective technical pathway for fabricating high-quality， large size P-type 4H-SiC epitaxial layers， laying a solid material for their industrial application in high-voltage power devices.

The n-type 4H-SiC crystal samples prepared by the physical vapor transport （PVT） method in this study were subjected to etching treatment using molten KOH in a nickel crucible at 520 ℃. Dislocations in the wafer， including threading edge dislocation （TED）， threading screw dislocation （TSD）， and threading mixed dislocation （TMD）， were distinguished and identified using optical microscopy （OM）， atomic force microscopy （AFM）， and scanning electron microscopy （SEM）. Micro-Raman spectroscopy was employed to characterize the spectral features of these different dislocation defects. The Raman spectra from the defect regions were compared with those from defect-free regions， with particular focus on the longitudinal optical phonon plasmon coupling （LOPC） mode near 984.609 cm^-1， which is influenced by increased carrier concentration. The Raman spectra were fitted using Matlab， and the carrier concentration （n） in different regions was calculated based on theoretical formulas. Comparing the n values at the defect cores with those in defect-free areas revealed a significant carrier trapping effect at the defect cores. The full width at half maximum （FWHM） of the Raman spectra at each defect core was further calculated. By jointly analyzing the FWHM and the carrier concentration n， the relative strength of the carrier trapping effect for different dislocation types was assessed， yielding the following order： TSD， TMD， TED. Some optimization suggestions are proposed for the crystal growth process and device fabrication process.

In this study， GaN crystals were grown by Na-flux liquid-phase epitaxy method on patterned substrates （PS） and non-patterned substrates （NPS） at various growth temperatures. COMSOL numerical simulations were employed to calculate and analyze the nitrogen concentration distribution and supersaturation evolution in the Ga-Na melt under different temperature conditions. The combined experimental and simulation results systematically reveal the differences in epitaxial growth behaviors of GaN crystals on PS and NPS. The results indicate that epitaxial growth on PS was successfully achieved only at 840 and 850 ℃. The obtained GaN crystals exhibit a regular hexagonal pyramid morphology， with partially unfilled gaps observed in the upper region of the cross-sections. As the temperature increases from 840 ℃ to 850 ℃， the nitrogen concentration in the melt increases while the supersaturation decreases， leading to smoother pyramid facets and an increase in crystal thickness from approximately 570 μm to 810 μm. When the temperature raises to 860 ℃， the supersaturation decreases further， resulting in complete dissolution of the seed points and no epitaxial growth occurring. In contrast， stable epitaxial growth on NPS was achieved over the temperature range from 840 ℃ to 860 ℃. With increasing temperature， the surface morphology evolves from ridge-like structures at lower temperatures to flat， cell-like structures at higher temperatures， while the cross-sections exhibit dense and laterally continuous growth. The crystal thickness first increases and then decreases with temperature increasing， reaching a maximum of approximately 1 450 μm at 850 ℃. In addition， the thickness of the GaN crystals obtained on PS is significantly thinner than that on NPS. Overall， compared with NPS， PS has a narrower stable epitaxial growth window， which requires more precise control of growth parameters to achieve high-quality epitaxial crystal growth.

This study systematically investigated the effects of surface morphology， yield and quality of GaN crystals grown for different durations by adopting the Na flux liquid phase epitaxy method. Combined with the change of material in the crucible and calculation of N ion concentration during the growth process， the growth process of GaN crystal by flux-excess-assisted liquid phase epitaxy was elucidated. The results indicate that with the extension of growth time， the surface morphology of the crystals gradually evolves from the initial small-sized ridge-like to the pyramids shape， and eventually develops into large-sized ridge-like morphologies. The epitaxial thickness of the crystals increases with the extension of growth time， and the growth thickness is approximately 1 500 μm when the growth time is 100 h. Meanwhile， the crystal yield is significantly improved， exhibiting an approximately linear correlation with the growth time. When the growth time is 100 h， the single crystals yield， polycrystals yield and total yield of GaN are about 65.5%， 18.5% and 84.0%， respectively. In the initial growth stage， the full width at half maximum （FWHM） of the X-ray rocking curve （XRC） for the （0002） plane is less than 270″， and it gradually increases with the extension of growth time. The calculation results of the residual materials in the melt during growth process show that the mass of residual metallic Ga decreases linearly with growth time， whereas the mass of residual metallic Na increases slightly in the initial growth stage and then stabilizes in the later stage. Numerical calculation results reveal that the N ion concentration in the melt shows an increasing trend with the extension of growth time. This work provides important experimental and theoretical basis for regulating the morphology， improving the yield and optimizing the growth process of GaN crystals.

This study systematically investigates the effects of the miscut angle of c-plane GaN substrates on the morphology， indium incorporation behavior， and optical properties of InGaN epitaxial layers grown by metal-organic chemical vapor deposition （MOCVD） method. The results demonstrate that as the substrate miscut angle increases， the morphology of InGaN transitions from two-dimensional island-like structures to step-like structures， primarily driven by a reduction in surface supersaturation. The critical miscut angle for the transition from two-dimensional islands to step structures increases with higher vapor-phase supersaturation， which can be achieved by lowering the growth temperature or increasing the growth rate. Furthermore， both the InN mole fraction and photoluminescence intensity reach their maximum values near the critical miscut angle. This study elucidates the mechanism by which the miscut angle and growth conditions jointly regulate the morphology and properties of InGaN， providing both theoretical underpinnings and experimental guidance for the fabrication of high-quality InGaN-based optoelectronic devices.

β-phase gallium oxide （β-Ga₂O₃） crystals have become a key material for high-power devices due to their ultra-wide bandgap characteristics. The vertical Bridgman （VB） method is currently the most promising approach for commercial-scale growth of gallium oxide single crystals. However， the semi-transparent nature of gallium oxide crystals and melts can cause significant internal radiation， which affects the temperature and flow fields during crystal growth process， and thus the crystal quality. Therefore， in this paper， a heat transfer numerical model for the growth process of gallium oxide crystals by VB method was established using the finite element software Comsol Multiphysics. The influence of internal radiation on the temperature field， melt flow field， melt-crystal interface， and crystal thermal stress was systematically investigated. The numerical simulation results show that the internal radiation in the crystal significantly enhances the thermal transport of the crystal. The radiation heat from the melt-crystal interface can directly penetrate the semi-transparent crystal to the crucible wall， reducing the temperature gradient and thermal stress inside the crystal. This radiation directly cools the melt-crystal interface， causing a downward trend in the temperature at the interface. To maintain the melting point temperature， the melt-crystal interface must move towards the upper high-temperature melt， increasing the convexity of the interface. The internal radiation in the melt also affects the heat transfer in the melt region. The radiation from the hot zone can penetrate the melt to the melt-crystal interface， radiatively heating the interface. Therefore， the melt-crystal interface moves towards the crystal side， and the convexity of the interface shape decreases， presenting a W-shaped distribution. However， since the isothermal lines and thermal stress of the crystal mainly accumulate at the bottom of the crystal， the effect on the temperature gradient and thermal stress inside the crystal is small. In addition， the sensitivity of internal radiation to the absorption coefficients of the crystal and melt was systematically analyzed. It was found that as the absorption coefficient of the crystal decreases， the internal radiation in the crystal increases， the temperature gradients in the melt and crystal decrease， the thermal stress of the crystal decreases， and the convexity of the melt-crystal interface increases， leading to an uneven radial distribution of solutes. As the absorption coefficient of the melt decreases， the internal radiation in the melt increases， the temperature gradient and thermal stress at the bottom of the crystal slightly decrease， the convexity of the melt-crystal interface at the center decreases， the W-shaped distribution becomes more obvious， and the edges are more prone to polycrystalline nucleation， thereby affecting the crystal quality.

In this paper， the first-principles calculation method was adopted to systematically investigate the effects of Ga vacancies and 3d transition metal elements Ti， V， Cr， Mn， Fe， Co， Ni and Cu doping on the geometric structure， electronic structure， stability and magnetic properties of β-Ga₂O₃.The calculation results show that both Ga vacancies and doped atoms cause local distortions of different degrees in the geometric structure of β-Ga₂O₃， but do not destroy the overall symmetry of β-Ga₂O₃ structure. The formation energy calculation results indicate that β-Ga₂O₃ systems containing Ga vacancies or doped atoms are all stable and are more likely to be formed in an oxygen-rich environment. More importantly， the ground states of the β-Ga₂O₃ systems containing Ga vacancies and Ti， V， Cr， Mn， Fe， Ni and Cu doping are magnetic， with magnetic moments of 2.51， 0.67， 0.12， 3.00， 2.05， 1.00， 1.00 and 1.92 μ_B， respectively. Based on the analysis， it can be concluded that the magnetic moment distribution in β-Ga₂O₃ systems containing Ga vacancies or doped with Ti， V， Cr， Mn， Fe， Ni， and Cu is associated with the hybridization between the vacancy/dopant atom and its neighboring oxygen atoms. In transition metal-doped β-Ga₂O₃ systems， the magnetic moments primarily originate from the contribution of 3d transition metal dopants.

In this study， first-principles calculations were employed to construct armchair （AC） and zigzag （ZZ） type lateral heterojunction models， denoted as AC-（SnSe） _m /（SnS） _n and ZZ-（SnSe） _m /（SnS） _n （m/n=1/11， 6/6， 11/1）. The structural stability， electronic structure， optical properities， and strain regulation effects of these heterostructures were systematically investigated. The results show that the band gap of AC-（SnSe） _m /（SnS） _n decreases with increasing m， which is benefical to promote the generation of photoexcited electron-hole pairs and enhances photocatalytic activity. Among all the constructed configurations， only ZZ-（SnSe）₆/（SnS）₆ exhibits a typical type-Ⅱ band alignment structure， which effectively promotes the spatial separation of photogenerated carriers， increases the probability of electron excitation and transfer， and improves the overall optoelectronic performance. Optical absorption spectrum analysis indicates that both ZZ-（SnSe）₆/（SnS）₆ and AC-（SnSe）₆/（SnS）₆ possess stronger polarization responses and better carrier transport potentials， with ZZ-（SnSe）₆/（SnS）₆ displaying a wider optical absorption range and higher absorption intensity. Moreover， strain engineering is demonstrated to further regulate the properties. For AC-（SnSe）₆/（SnS）₆， a +4% tensile strain induces an indirect-to-direct band gap transition， while a -12% compressive strain significantly enhances optical absorption， suggesting superior photocatalytic efficiency. This study reveals the synergistic regulation mechanism of component ratio and strain engineering in SnSe/SnS lateral heterojunction， which provides a theoretical basis for the design of efficient two-dimensional photocatalytic/photovoltaic devices.

Numerical simulation studies on gas flow patterns in the deposition chamber during chemical vapor deposition （CVD） of zinc sulfide （ZnS） were conducted using FLUENT software. Focusing on the flow characteristics of zinc（Zn） vapor and hydrogen sulfide （H₂S） gas， a three-dimensional physical model of the deposition chamber was established， and the distribution characteristics of the gas flow field under the synergistic effects of various process parameters were systematically analyzed. The study emphasized the distribution of gas flow patterns， density flow fields， velocity flow fields， and the evolution of temperature fields under parameters such as chamber pressure and reactant inlet velocity. It aims to reveal how gas flow patterns influence the uniformity of ZnS deposition rates and the quality of material growth. Results indicate that as the deposition pressure increases from 3 000 Pa to 6 000 Pa and the nozzle velocity synchronously increases， the gas flow patterns in the deposition chamber exhibit a synergistic evolution trend of “disordered-ordered-stable-instable”. This work provides a theoretical basis and process optimization directions for the controllable growth of high-quality ZnS materials， and offers significant guiding value for the preparation of infrared optical materials.