A Pre-trained Deep Potential Model for Sulfide Solid Electrolytes with Broad Coverage and High Accuracy

Ruoyu Wang,Mingyu Guo,Yuxiang Gao,Xiaoxu Wang,Yuzhi Zhang,Bin Deng,Xin Chen,Mengchao Shi,Linfeng Zhang,Zhicheng Zhong

2024-07-24

Abstract:Solid electrolytes with fast ion transport are one of the key challenges for solid state lithium metal batteries. To improve ion conductivity, chemical doping has been the most effective strategy, and atomistic simulation with machine-learning potential helps find optimized doping by predicting ion conductivity for arbitrary composition. Yet most existing machine-learning models are trained on narrow chemistry, and new model has to be trained for each system, wasting transferable knowledge and incurring significant cost. Here, we propose a pre-trained deep potential model purpose-built for sulfide electrolytes with attention mechanism, known as DPA-SSE. The training set encompasses 15 elements and consists of both equilibrium and extensive out-of-equilibrium configurations. DPA-SSE achieves a high energy resolution of less than 2 meV/atom for dynamical trajectories up to 1150 K, and reproduces experimental ion conductivity of sulfide electrolytes with remarkable accuracy. DPA-SSE exhibits good transferability, covering a range of complex electrolytes with mixes of cation and anion atoms. Highly efficient dynamical simulation with DPA-SSE can be realized by model distillation which generates a faster model for given systems. DPA-SSE also serves as a platform for continuous learning, and the model fine-tune requires only a portion of downstream data. These results demonstrate the possibility of a new pathway for AI-driven development of solid electrolytes with exceptional performance.

Chemical Physics

What problem does this paper attempt to address?

The paper aims to address the issue of low ionic conductivity in solid-state electrolyte materials, particularly focusing on sulfide solid-state electrolytes. Specifically, the goals of the paper include:

1. **Developing a new machine learning potential model**: A pre-trained deep potential model (DPA-SSE) is proposed to simulate sulfide solid-state electrolytes, with particular attention to its applicability and high accuracy across a broad chemical space.

2. **Improving ionic conductivity**: By employing chemical doping strategies to optimize the ionic conductivity of sulfide electrolytes, which is one of the key factors in enhancing the performance of solid-state lithium batteries.

3. **Overcoming limitations of existing models**: Current machine learning models are typically limited to narrow chemical ranges, requiring retraining for each new system, which not only wastes transferable knowledge but also incurs high costs. The DPA-SSE model aims to overcome this limitation by achieving good transferability.

4. **Accurately predicting non-equilibrium configurations**: Ion transport calculations require accurate predictions of non-equilibrium configurations along dynamic trajectories, an area where existing general models perform poorly. DPA-SSE improves the prediction accuracy for such configurations by including a large amount of non-equilibrium configuration training data.

5. **Model fine-tuning and distillation**: The pre-trained model can be fine-tuned for specific tasks with only a small amount of additional data to achieve good results. Additionally, a distillation scheme is proposed to generate faster and lighter models from the pre-trained model, suitable for large-scale dynamic simulations.

In summary, the main goal of the paper is to accelerate the design process of sulfide solid-state electrolyte materials by developing a new pre-trained deep potential model, particularly optimizing in the area of chemical doping, with the aim of improving the energy density, safety, and sustainability of solid-state lithium batteries.