Search

SignificanceBig Models, or Foundation Models, have offered a new paradigm in smart agriculture. These models, built on the Transformer architecture, incorporate numerous parameters and have undergone extensive training, often showing excellent performance and adaptability, making them effective in addressing agricultural issues where data is limited. Integrating big models in agriculture promises to pave the way for a more comprehensive form of agricultural intelligence, capable of processing diverse inputs, making informed decisions, and potentially overseeing entire farming systems autonomously.ProgressThe fundamental concepts and core technologies of big models are initially elaborated from five aspects: the generation and core principles of the Transformer architecture, scaling laws of extending big models, large-scale self-supervised learning, the general capabilities and adaptions of big models, and the emerging capabilities of big models. Subsequently, the possible application scenarios of the big model in the agricultural field are analyzed in detail, the development status of big models is described based on three types of the models: Large language models (LLMs), large vision models (LVMs), and large multi-modal models (LMMs). The progress of applying big models in agriculture is discussed, and the achievements are presented.Conclusions and ProspectsThe challenges and key tasks of applying big models technology in agriculture are analyzed. Firstly, the current datasets used for agricultural big models are somewhat limited, and the process of constructing these datasets can be both expensive and potentially problematic in terms of copyright issues. There is a call for creating more extensive, more openly accessible datasets to facilitate future advancements. Secondly, the complexity of big models, due to their extensive parameter counts, poses significant challenges in terms of training and deployment. However, there is optimism that future methodological improvements will streamline these processes by optimizing memory and computational efficiency, thereby enhancing the performance of big models in agriculture. Thirdly, these advanced models demonstrate strong proficiency in analyzing image and text data, suggesting potential future applications in integrating real-time data from IoT devices and the Internet to make informed decisions, manage multi-modal data, and potentially operate machinery within autonomous agricultural systems. Finally, the dissemination and implementation of these big models in the public agricultural sphere are deemed crucial. The public availability of these models is expected to refine their capabilities through user feedback and alleviate the workload on humans by providing sophisticated and accurate agricultural advice, which could revolutionize agricultural practices.

The solidification microstructure of the Mg-5Al alloy is controlled by the Gd content (0.00%, 0.25%, 0.50%, 0.75%, 1.00%, mass fraction, the same below) and high pressure (3 GPa), and the correlation between microstructure and compressive properties at room temperature was studied. The results show that the microstructure is single solid solution only when the the content of Gd is 0.75%, for the Mg-5Al-xGd alloys solidified under atmospheric pressure graphit mold casting. In the Mg-5Al-0.75Gd alloy, no eutectic phase form at the grain boundaries, granular Al2Gd phase is dispersed on the matrix, and the average grain size is about 85 μm. Moreover, the compressive strength and the maximum compressive strain of the Mg-5Al-0.75Gd alloy solidified under atmospheric pressure are 379 MPa and 33.46% respectively, which are higher than those of alloys with intergranular second phase. For the Mg-5Al-xGd alloys solidified under 3 GPa pressure, the microstructure is single solid solution when the Gd content is ≤ 0.25%, and there are eutectic phase Al2Gd at the grain boundaries (among dendrites) when the Gd content is ≥ 0.50%. In addition, the grain sizes of the Mg-5Al-0.25Gd and Mg-5Al-0.75 Gd alloys solidified under 3 GPa pressure are 74 μm and 38 μm respectively. However, the compressive strength and the maximum compressive strain of the Mg-5Al-0.25Gd alloy are 402 MPa and 33.61% respectively, which are superior than that of the Mg-5Al-0.75Gd alloy with the compressive strength of 341 MPa and the maximum compressive strain of 25.12%. Therefore, the existence form of Al and Gd at grain boundaries is an important factor affecting the mechanical properties of the Mg-5Al-xGd alloys.

A NiCoV layered double hydroxide (LDH) nanosheet array structure grown on Ni foam was synthesized through V-doping coupled H2O2 pre-oxidation using a hydrothermal method. Systematic characterization proved that NiCoVx-LDH/NF-y (x: 0.5, 1.0, and 1.5; y (the amount of H2O2 used): 50, 100, and 200) composites featured vertically interlaced LDH nanosheets with more highly valent metal species Ni3+ and Co3+. Especially, NiCoV1.0-LDH/NF-100 demonstrated the best electrocatalytic performance for the urea oxidation reaction (UOR) and hydrogen evolution reaction (HER). Furthermore, the NiCoV1.0-LDH nanosheet arrays were in situ grown on reduced graphene oxide (rGO) uniformly modified Ni foam (∼135 nm × 15 nm), only requiring a low potential of 1.33 V for the UOR and a low overpotential of 70 mV for the HER to offer 10 mA cm−2. When it is used as both the cathode and anode in urea-assisted water electrolysis, the as-prepared array only requires 1.45 V to offer 10 mA cm−2, outperforming many non-precious metal catalysts. In situ characterization and theoretical calculations showed that V-doping coupled H2O2 pre-oxidation can regulate the electronic environment of the catalyst via (1) producing more highly valent Ni3+ active sites and Co3+ auxiliary sites; (2) reducing the adsorption energy of urea at the Ni site in the UOR and simultaneously reducing the water adsorption energy and water dissociation energy in the HER, especially the active site of H* adsorption was shifted from Co sites to V sites, which optimized the Gibbs free energy of H* adsorption; and (3) reducing the band gap, thus accelerating the reaction kinetics of the UOR and HER. Additionally, the synergistic coupling between NiCoV-LDH and rGO further improved the electrocatalytic performance. [ABSTRACT FROM AUTHOR]

The electrochemical oxygen reduction reaction (ORR) to produce hydrogen peroxide (H2O2) is appealing due to its sustainability. However, its efficiency is compromised by the competing 4e− ORR pathway. In this work, we report a hierarchical carbon nanosheet array electrode with a single-atom Ni catalyst synthesized using organic molecule-intercalated layered double hydroxides as precursors. The electrode exhibits excellent 2e− ORR performance under alkaline conditions and achieves H2O2 yield rates of 0.73 mol gcat−1 h−1 in the H-cell and 5.48 mol gcat−1 h−1 in the flow cell, outperforming most reported catalysts. The experimental results show that the Ni atoms selectively adsorb O2, while carbon nanosheets generate reactive hydrogen species, synergistically enhancing H2O2 production. Furthermore, a coupling reaction system integrating the 2e− ORR with ethylene glycol oxidation significantly enhances H2O2 yield rate to 7.30 mol gcat−1 h−1 while producing valuable glycolic acid. Moreover, we convert alkaline electrolyte containing H2O2 directly into the downstream product sodium perborate to reduce the separation cost further. Techno-economic analysis validates the economic viability of this system. Electrochemical oxygen reduction is a promising method for H2O2 production but suffers from poor activity and low yield. In this work, the authors present Ni single atom catalyst for efficient H2O2 production which is further coupled with organic oxidation reaction to enhance the economic feasibility of the system. [ABSTRACT FROM AUTHOR]

The dry reforming of methane provides an attractive route to convert greenhouse gases (CH4 and CO2) into valuable syngas, so as to resolve the carbon cycle and environmental issues. However, the development of high-performance catalysts remains a huge challenge. Herein, we report a 0.6% Ir/CeO2−x catalyst with a metal-support interface structure which exhibits high CH4 (~72%) and CO2 (~82%) conversion and a CH4 reaction rate of ~973 μmolCH4 gcat−1 s−1 which is stable over 100 h at 700 °C. The performance of the catalyst is close to the state-of-the-art in this area of research. A combination of in situ spectroscopic characterization and theoretical calculations highlight the importance of the interfacial structure as an intrinsic active center to facilitate the CH4 dissociation (the rate-determining step) and the CH2* oxidation to CH2O* without coke formation, which accounts for the long-term stability. The catalyst in this work has a potential application prospect in the field of high-value utilization of carbon resources. Dry reforming of methane is an attractive route to convert greenhouse gases into valuable syngas. Here, the authors demonstrate design strategy to prepare efficient catalysts towards activation of both C–H and C = O bonds and provide atomic-level insights into interfacial synergistic catalysis. [ABSTRACT FROM AUTHOR]