Developing tailored heterostructures on demand is essential to meet the growing needs of semiconductor devices. However, traditional methods remain constrained by simulation-based design and iterative trial-and-error optimization. Here, we introduce SemiEpi, a self-driving platform designed for molecular beam epitaxy (MBE) that enables multi-step semiconductor heterostructure growth through in situ Reflection High Energy Electron Diffraction monitoring and on-the-fly feedback control. By integrating MBE reactors, physics-informed machine learning (ML) models, and parameter initialization, SemiEpi designs heterostructures, identifies optimal initial conditions, and proposes experiments for material growth. As a demonstration, we optimized high-density InAs quantum dot growth with a target emission wavelength of 1240 nm, achieving a density of 5 × 1010 cm-2, a 1.6-fold increase in photoluminescence intensity and a reduced full width at half maximum of 29.13 meV through feedback control of growth temperatures. We further demonstrate SemiEpi's versatility across different MBE reactors, highlighting its potential to address challenges in multi-step heterostructure growth, enable hardware-independent frameworks, and enhance process repeatability and stability.

Nuclear battery is a potential energy generator because of its long lifetime, stable output performance, high energy density and environmental resistance. It converts the energy of high energy particles into electric energy. The main constituent parts of nuclear batteries are radioisotope and semiconductor energy converter. Diamond is ultra-wide band gap semiconductor with high carrier mobilities and high chemical inertness. More importantly, it is an excellent radiation resistance material. Therefore, diamond is an ideal material for the fabrication of nuclear batteries. This paper reviews the development status of diamond-based nuclear batteries. The alpha-voltaic, beta-voltaic and gamma-voltaic nuclear batteries based on diamond Schottky junction and p–n junction are illustrated in this review. In addition, an outlook on future research of diamond-based nuclear batteries is also provided.

Traditional laser optimization relies on trial-and-error. We integrate reflection-high-energy-electron-diffraction with machine learning to correlate surface reconstructions with photoluminescence of the InAs active region. Lasers operating at 1240 nm achieved excellent performances comparable to traditional lasers.

Silicon photonics requires CMOS-compatible, on-chip, efficient laser diodes (LDs) and low-noise photodetectors (PDs). Here, we demonstrate the monolithic integration of InAs QD LDs and waveguide PDs (WPDs) on CMOS-compatible (001) silicon by employing an identical epitaxial structure. The device exhibits bias-selective operation with both components isolated by XeF2-assisted focused ion beam trenching, a strategy ideal for LDs and PDs that require precise positioning yet occupy a minimal package footprint. The LD exhibits a low threshold current density of 170 A/cm2 and a single-facet continuous-wave output power of 48 mW, while the WPD shows a low dark current density of 4.9 × 10-5 A/cm2 at -1 V and a broad O-band spectral response. The integrated LD-WPD system shows a clear correlation between LD output power and WPD photocurrent with an estimated responsivity of 0.118 A/W. Our results demonstrate the feasibility of on-chip power monitoring, paving the way for compact, reliable, and high-density photonic integration.

The synthesis of high-quality heteroepitaxial diamond films on iridium composite substrates is a critical step toward advancing diamond for electronic and optical applications. Microwave plasma chemical vapor deposition, combined with in situ optical emission spectroscopy, enables precise control over growth modes through plasma parameter tuning. In this study, we examine how methane concentration, microwave power, and gas pressure influence plasma species and, consequently, the growth modes of heteroepitaxial diamond by optical emission spectroscopy and scanning electron microscope. At low nucleation densities, increased methane concentrations promote the transition from faceted polyhedral to ballas structures, driven by elevated C2 radical concentrations in the plasma. Conversely, at higher nucleation densities, gas pressure, and substrate temperature dominate growth mode determination, leading to diverse morphologies, such as planar, polycrystalline, octahedral, and step-flow growth. These findings elucidate the interplay among plasma species, growth parameters, and growth mode, offering critical insights for optimizing growth conditions and preparing heteroepitaxial diamond films in a specific growth mode.

The heteroepitaxial growth of large-area thick single-crystal diamond is strongly limited by stress accumulation and crack formation during prolonged deposition. Laser-patterned templates have been demonstrated as an effective approach for obtaining crack-free freestanding diamond layers. In this work, the influence of laser grooves on the growth behavior of diamond films grown on Ir/YSZ/Si (001) substrates by microwave plasma chemical vapor deposition (MPCVD) is systematically investigated. Laser grooves introduced into a thin diamond seed layer generate a spatially non-uniform growth template. During the early growth stage, polycrystalline diamond forms inside the grooves, whereas the surrounding regions maintain stable (001)-oriented epitaxial growth. As deposition proceeds, the faster vertical growth of single-crystal diamond progressively overgrows the groove regions, leading to gradual groove filling and closure, ultimately forming a laterally continuous epitaxial diamond film. The buried grooves with polycrystalline diamond act as mechanically compliant zones that facilitate stress redistribution during thick-film growth, thereby suppressing crack formation and enabling the fabrication of millimeter-scale freestanding diamond layers with high crystalline quality. This work clarifies how laser-patterned templates regulate growth evolution and stress accommodation during diamond heteroepitaxy, providing insights for the controlled growth of large-area thick single-crystal diamond films on highly mismatched substrates.

As an ultra-wide bandgap semiconductor, diamond possesses many excellent properties, making it a highly promising material for applications in high-frequency, high-power, and deep-ultraviolet semiconductor devices. Achieving efficient p-type and n-type doping is a fundamental requirement for the fabrication of diamond-based semiconductor devices. Currently, the mainstream doping elements are boron (p-type) and phosphorus (n-type). However, both elements introduce relatively deep impurity levels at room temperature, which hinders further applications. With the advancement of first-principles calculations and high-performance computing using supercomputers and computing clusters, it has become feasible to theoretically predict the doping properties of diamond. This paper primarily focuses on the progress in theoretical research on p-type and n-type doping in diamond based on first-principles methods, covering both single-element doping and co-doping with multiple elements. In addition, it summarizes the challenges in this field and provides an outlook on future research.

Traditional methods for optimizing light source emissions rely on a time-consuming trial-and-error approach. While in situ optimization of light source gain media emission during growth is ideal, it has yet to be realized. In this work, in situ reflection high-energy electron diffraction (RHEED) is integrated with machine learning (ML) to correlate the surface reconstruction with the photoluminescence (PL) of InAs/GaAs quantum dots (QDs), which serve as the active region of lasers. A lightweight ResNet-GLAM model is employed for the real-time processing of RHEED data as input, enabling effective identification of optical performance. This approach guides the dynamic optimization of growth parameters, allowing real-time feedback control to adjust the QDs emission for lasers. InAs QDs on GaAs substrates are successfully optimized, with a 3.2-fold increase in PL intensity and a reduction in full width at half maximum (FWHM) from 36.69 to 28.17 meV. Automated, in situ self-optimized lasers with 5-layer InAs QDs achieved electrically pumped continuous-wave operation at 1240 nm with a low threshold current of 150 A cm- 2 at room temperature, an excellent performance comparable to samples grown through traditional manual multi-parameter optimization methods. These results mark a significant step toward intelligent, low-cost, and reproductive light emitters production.

Gastric ulcers often cause postprandial epigastric pain, especially in acute cases. Abnormal motility, with about 50 % of patients having delayed gastric emptying, contributes to ulcer development. Costunolide (COS) and dehydrocostuslactone (DEH), derived from "Mu xiang" herbs, show potential in treating ulcers and regulating gastrointestinal motility. However, their poor solubility and bioavailability limit in vivo use. This study uses electrospinning to develop hydrophilic nanofibers loaded with COS and DEH in a polyvinylpyrrolidone (PVP) matrix for treating acute gastric ulcers. The production process for costunolide / dehydrocostuslactone nanofibers (COS/DEH/NFs) was optimized, characterized, and tested in pharmacodynamic studies. The results showed that COS and DEH remained in a non-crystalline state within COS/DEH/NFs, enhancing their in vitro release. With 21 % drug incorporation, COS/DEH/NFs released over 70 % of COS and more than 50 % of DEH within 20 min in body fluids. In treatment, COS/DEH/NFs suppressed pro-inflammatory cytokines, resisted oxidative stress, promoted gastric mucosal repair, and enhanced gastrointestinal motility. In a mouse model of acute gastric ulcer, high-dose COS/DEH/NFs achieved a 77.09 % ulcer inhibition rate, and low-dose COS/DEH/NFs resulted in gastric residual and intestinal propulsion rates of 73.55 % and 69.89 %, respectively. The drug loading of COS/DEH/NFs is 14.76 ± 0.26 %, with an encapsulation efficiency of 68.77 ± 1.14 %. COS/DEH/NFs is a new choice for treating acute gastric ulcers with gastrointestinal bloating due to its convenience and swallow-free use, providing rapid symptom relief.

The semiconductor industry is increasingly reliant on advances in epitaxial technologies to meet the demands of high-performance applications such as advanced photonics, quantum computing, and power electronics. However, the nonlinear dynamics of crystal epitaxial growth, combined with stochastic interfacial fluctuations, present significant challenges. These challenges create fundamental difficulties between the control of empirical parameters and the stringent material quality requirements, which hinder systematic improvements in material performance. This paper reviews recent progress in Intelligent Epitaxy, a transformative framework that employs an autonomous architecture consisting of sensing, decision-making, and execution. This framework integrates machine learning with precise characterization and control through three core modules: the Multimodal Sensing Module, the Knowledge-Informed Decision Module, and the Adaptive Control Module. Together, these modules enable comprehensive monitoring of growth dynamics, causal analysis of the relationships between parameters, growth states, and outcomes, as well as the autonomous regulation of growth processes. Additionally, we discuss and address current challenges and issues in this field, providing insights and perspectives for future research. Our review aims to guide the development of a new technological trajectory that goes beyond traditional approaches, positioning intelligent epitaxy as the foundation for next-generation autonomous semiconductor manufacturing.

Self-assembled InAs/GaAs quantum dots (QDs) have properties highly valuable for developing various optoelectronic devices such as QD lasers and single photon sources. The applications strongly rely on the density and quality of these dots, which has motivated studies of the growth process control to realize high-quality epi-wafers and devices. Establishing the process parameters in molecular beam epitaxy (MBE) for a specific density of QDs is a multidimensional optimization challenge, usually addressed through time-consuming and iterative trial-and-error. Here, we report a real-time feedback control method to realize the growth of QDs with arbitrary density, which is fully automated and intelligent. We developed a machine learning (ML) model named 3D ResNet 50 trained using reflection high-energy electron diffraction (RHEED) videos as input instead of static images and providing real-time feedback on surface morphologies for process control. As a result, we demonstrated that ML from previous growth could predict the post-growth density of QDs, by successfully tuning the QD densities in near-real time from 1.5E10 cm-2 down to 3.8E8 cm-2 or up to 1.4E11 cm-2. Compared to traditional methods, our approach, with in situ tuning capabilities and excellent reliability, can dramatically expedite the material optimization process and improve the reproducibility of MBE, constituting significant progress for thin film growth techniques. The concepts and methodologies proved feasible in this work are promising to be applied to a variety of material growth processes, which will revolutionize semiconductor manufacturing for optoelectronic and microelectronic industries.

>As an ultra-wide bandgap semiconductor, diamond garners significant interest due to its exceptional physical properties [1–3] . These superior characteristics make diamonds highly promising for applications in power electronics [4] , deep-ultraviolet detectors [5] , high-energy particle detectors [6] , and quantum devices based on color centers [7] .

Ex situ characterization techniques in molecular beam epitaxy (MBE) have inherent limitations, such as being prone to sample contamination and unstable surfaces during sample transfer from the MBE chamber. In recent years, the need for improved accuracy and reliability in measurement has driven the increasing adoption of in situ characterization techniques. These techniques, such as reflection high-energy electron diffraction, scanning tunneling microscopy, and X-ray photoelectron spectroscopy, allow direct observation of film growth processes in real time without exposing the sample to air, hence offering insights into the growth mechanisms of epitaxial films with controlled properties. By combining multiple in situ characterization techniques with MBE, researchers can better understand film growth processes, realizing novel materials with customized properties and extensive applications. This review aims to overview the benefits and achievements of in situ characterization techniques in MBE and their applications for material science research. In addition, through further analysis of these techniques regarding their challenges and potential solutions, particularly highlighting the assistance of machine learning to correlate in situ characterization with other material information, we hope to provide a guideline for future efforts in the development of novel monitoring and control schemes for MBE growth processes with improved material properties.

Traditional methods for optimizing light source emissions rely on a time-consuming trial-and-error approach. While in-situ optimization of light source gain media emission during growth is ideal, it has yet to be realized. In this work, we integrate in-situ reflection high-energy electron diffraction (RHEED) with machine learning (ML) to correlate the surface reconstruction with the photoluminescence (PL) of InAs/GaAs quantum dots (QDs), which serve as the active region of lasers. A lightweight ResNet-GLAM model is employed for the real-time processing of RHEED data as input, enabling effective identification of optical performance. This approach guides the dynamic optimization of growth parameters, allowing real-time feedback control to adjust the QDs emission for lasers. We successfully optimized InAs QDs on GaAs substrates, with a 3.2-fold increase in PL intensity and a reduction in full width at half maximum (FWHM) from 36.69 meV to 28.17 meV under initially suboptimal growth conditions. Our automated, in-situ self-optimized lasers with 5-layer InAs QDs achieved electrically pumped continuous-wave operation at 1240 nm with a low threshold current of 150 A/cm2 at room temperature, an excellent performance comparable to samples grown through traditional manual multi-parameter optimization methods. These results mark a significant step toward intelligent, low-cost, and reproductive light emitters production.

In previous studies, the influence of gas phase and surface reactions on the growth of GaN was mainly calculated through simulations. In this study, a novel gas pre-decomposition device (GPDD) was designed to experimentally investigate the effects of gas phase and surface reactions on GaN growth by changing the length and height of the isolation plates (IPs). By varying the structure of the GPDD, the effects on the growth rate and thickness uniformity of the GaN films were studied. The growth rate of the GaN sample slowed with the extension of the IPs because the longer partition plates led to insufficient gas mixing and premature consumption of the precursor trimethylgallium (TMG). The use of GPDD simultaneously achieves high crystal quality and smooth surface morphology of the GaN film. Owing to the use of GPDD, the decomposition of TMG in the pyrolysis pathway was promoted, which suppressed Ga vacancies and C impurities, resulting in weak yellow luminescence bands in the photoluminescence. This study provides a comprehensive understanding of the chemical reaction mechanism of GaN and plays an important role in promoting the development of metal-organic chemical vapor deposition equipment.

The semiconductor industry has prioritized automating repetitive tasks through closed-loop, self-driving experimentation, accelerating the optimization of complex multi-step processes. The emergence of machine learning (ML) has ushered in self-driving processes with minimal human intervention. This work introduces SemiEpi, a self-driving platform designed to execute molecular beam epitaxy (MBE) growth of semiconductor heterostructures through multi-step processes, in-situ monitoring, and on-the-fly feedback control. By integrating standard reactor, parameter initialization, and multiple ML models, SemiEpi identifies optimal initial conditions and proposes experiments for multi-step heterostructure growth, eliminating the need for extensive expertise in MBE processes. SemiEpi initializes material growth parameters tailored to specific material characteristics, and fine-tuned control over the growth process is then achieved through ML optimization. We optimize the growth for InAs quantum dots (QDs) heterostructures to showcase the power of SemiEpi, achieving a QD density of 5E10/cm2, 1.6-fold increased photoluminescence (PL) intensity and reduced full width at half maximum (FWHM) of 29.13 meV. This work highlights the potential of closed-loop, ML-guided systems to address challenges in multi-step growth. Our method is critical to achieve repeatable materials growth using commercially scalable tools. Furthermore, our strategy facilitates developing a hardware-independent process and enhancing process repeatability and stability, even without exhaustive knowledge of growth parameters.

Semiconductor laser diodes (LDs), with merits of little volume, lightweight, low power consumption, ease of modulation, and high data rates, are great candidates for space laser communications. However, operating in a radiation environment can result in various damages to LDs. Hence, the growing focus on satellite laser communications necessitates LDs with improved radiation hardness. This review covers the efforts made in investigating the radiation effects on LDs induced by various types of radiation, including neutrons, γ-rays, protons, electrons, and other radiation particles. The conditions of radiation experiments and the behaviors and mechanisms of the degradation of LD material and device performance after being radiated are surveyed and discussed. It has been revealed that quantum dot (QD) LDs typically exhibit superior radiation tolerance compared to quantum well LDs or LDs with bulk active layers due to the enhanced confinement of carriers and reduced active areas in QD LDs, indicating significant potential for space applications. Furthermore, current challenges and issues in this field are discussed and addressed, providing a perspective and outlook for further research. Our review aims to guide the development of suitable light sources for future space laser communications, fostering high-performance satellite communications networks.

Thin film deposition is an essential step in the semiconductor process. During preparation or loading, the substrate is exposed to the air unavoidably, which has motivated studies of the process control to remove the surface oxide before thin film deposition. Optimizing the deoxidation process in molecular beam epitaxy (MBE) for a random substrate is a multidimensional challenge and sometimes controversial. Due to variations in semiconductor materials and growth processes, the determination of substrate deoxidation temperature is highly dependent on the grower's expertise; the same substrate may yield inconsistent results when evaluated by different growers. Here, we employ a machine learning (ML) hybrid convolution and vision transformer (CNN-ViT) model. This model utilizes reflection high-energy electron diffraction (RHEED) video as input to determine the deoxidation status of the substrate as output, enabling automated substrate deoxidation under a controlled architecture. This also extends to the successful application of deoxidation processes on other substrates. Furthermore, we showcase the potential of models trained on data from a single MBE equipment to achieve high-accuracy deployment on other equipment. In contrast to traditional methods, our approach holds exceptional practical value. It standardizes deoxidation temperatures across various equipment and substrate materials, advancing the standardization research process in semiconductor preparation, a significant milestone in thin film growth technology. The concepts and methods demonstrated in this work are anticipated to revolutionize semiconductor manufacturing in optoelectronics and microelectronics industries by applying them to diverse material growth processes.

Nuclear batteries have attracted much attention because of their long lifetime. Due to the large band gap and high radiation tolerance, diamond is an ideal material for the fabrication of the energy converter for nuclear batteries. In this work, a diamond Schottky barrier diode energy converter was prepared, and the performance of the device was studied by the electron-beam-induced current (EBIC) method. The energy converter can work under an electron beam current of mu A order. The open-circuit voltage of the device is around 0.8 V, and a short-circuit current of several milliamps can be achieved. The maximum energy conversion efficiency of 3.19% is obtained under a beam current of 0.2 mu A at 26 keV.

In order to better understand the influence of different complexes on the diamond co-doping system, N and Ga are chosen as co-dopants in diamond. The properties of several substitutional NmGan (m+n <= 3) complexes with vacancy (Va) in the bulk diamond have been investigated by ab initio density functional technique, including their optimized lattice structures, formation energies, binding energies and thermodynamic transition levels. The calculational results show that NmGan complexes in the donor-acceptor-donor (DAD) configuration can provide ionization energies similar to phosphorus-doped diamond. All other complexes provide deep impurity levels. For the DAD configuration, the adsorption process on the diamond surface has been studied to demonstrate the feasibility of growing diamonds containing N-Ga-N in experiments. The desired complex configuration is not uniquely present in the co-doped system. Investigating these properties of different complexes beyond NGaN provides insight into the N and Ga codoped diamond system, yielding a more comprehensive understanding of its potential and limitations. Our research ideas can also be extended to other co-doped systems and help to evaluate other potential co-dopants for diamond.