介观尺度下液相烧结过程的数值模拟研究进展

Research Progress on the Numerical Simulation of Liquid Phase Sintering in the Mesoscopic Scale

液相烧结是粉末冶金制造的关键技术,可以获得接近全致密化的高性能材料,为高熔点合金、硬质合金和金属陶瓷等材料的制备开辟了新途径。液相烧结过程中的组织演变及致密化行为直接决定了零件的力学性能和尺寸精度。介观尺度的数值模拟着眼于数十至上千个颗粒的系统,能够明确致密化机制,并准确、直观地反映液相分布、孔隙演化、颗粒生长等组织演变过程,是联系微观结构与宏观性能的重要桥梁,已成为液相烧结过程研究的热点方向。然而,液相烧结过程涉及颗粒运动、固液相变、流动等多种因素的耦合作用,对液相烧结数值模拟研究和实际应用提出了巨大的挑战。目前,液相烧结数值模拟的一种主要思路是将液相烧结过程简化为颗粒重排、固相溶解/析出和骨架形成三个界限分明的阶段,然后分别针对每个阶段开展模拟,从而在降低模拟难度的基础上反映液相烧结的部分显微组织演化机制及影响因素。其中,围绕颗粒重排和固相溶解/析出阶段的研究最为丰富,并取得了较多成果。基于离散元法的模型和液桥粗化模型主要用于颗粒重排阶段的研究,模型考虑了颗粒间的碰撞、滑动和粘结颗粒间烧结应力等的作用,并兼顾液相烧结中存在的表面张力和液相粘性力,描述了颗粒在液相中的重排和孔洞演化及致密化等重要现象,但晶粒生长和熟化的机制通常被忽略。固相溶解/析出阶段可简化为熟化过程,主要使用蒙特卡洛方法和相场方法研究该阶段的颗粒生长、颗粒形状变化和颗粒粒径分布的变化规律等。其中大尺度三维模拟预测的粒径分布和晶粒生长情况等结果能够与液相烧结中后期的实验结果对比,并吻合良好。实际液相烧结各阶段间并无明显界限,分阶段模拟的研究思路仍难以准确描述液相烧结的全过程。近年来研究者们提出了采用耦合模型的研究思路,旨在同时描述液相烧结中的颗粒运动、颗粒生长以及液体流动等行为。耦合模型克服了分阶段简化模型带来的问题,显著提高了模拟精度,准确预测了组织演化,并获得了烧结致密化率等定量数据。为促进基于耦合模型的数值模拟在液相烧结中的实际应用,仍需克服三维模型数值求解困难、实验验证缺乏等问题。本文回顾了介观尺度下液相烧结数值模拟方法的研究进展,分析了各种模拟方法描述液相分布、孔隙演化、颗粒长大以及致密化等问题的可靠性、准确性和优缺点,最后提出液相烧结数值模拟方法的发展前景及意见。

L iquid phasesintering (LPS) is a key technology in powder metallurgy for manufacturing high performance materials.LPS provides an innovative and efficient way for manufacturing nearly fully dense components made of refractory alloys, hard alloys, cermets, etc.The microstructure evolution and densification during LPS directly determine the mechanical properties and dimensional accuracy of the part. Numerical simulation at the mesoscopic/grain scale offers direct insights into the microstructure evolution of the sintered body, and also deals with complex densification mechanisms and their interplays. Hence, numerical simulations of LPS in the mesoscopic scale have gained enormous attentions inrecent years. LPS,involving grain growth and motion, solid-liquid transition and multi-phase flow, etc., presents a great challenge to the numerical simulation and its further applicationin industry. One common way, performing the numerical simulation of LPS, is based on a general assumption that LPS could be resolved into the three stages of grain rearrangement, grain dissolution/precipitation, and skeleton formation. Therefore, each stage of LPS is independently studied for simplification to still reveal some of the microstructure evolution mechanisms and affecting factors. In all the regarding findings, those on grain rearrangement and grain dissolution/precipitation are most fruitful.For grain rearrangement, studies were carried out based on the discrete element method, the liquid bridge coalescence model, etc. During rearrangement, the displacement of each particle was simulated in the viscous liquid under various forces, including inter-particle collision force, sliding force, sintering force and capillary force, etc. In these simulations, the grain motion, pore evolution and densification during grain rearrangement were usually described ignoring the grain growth and coarsening mechanisms. Phase field method and Monte-Carlo method were often adopted to simulate the grain dissolution/precipitation stage, which was treated as the classic Ostwald ripening. Large-scale three-dimensional simulations carried out by phase field method and Monte-Carlo method provided the particle size distribution and the growth kinetics very identical to the experimental observations. However, the simulations focusing on each independent stage is not applicable to describe the whole LPS process and the mechanisms under the experimental conditions, since the ranges overlapped for the three stages of LPS. Recently, the coupling strategies were presented to simulate LPS involving the liquid flow, grain rearrangement and grain growth simultaneously. The quantitative results of microstructure evolution and the sintering kinetics were successfully achieved by several coupled models for LPS, and the results showed higher accuracy than the models simulating only one stage. However, efficient numerical solutions for the three-dimensional simulations and experimental validations are required to promote the industry application of coupled simulations in the future. This paper reviews the recent developments in numerical modeling of LPS in the mesoscopic/grain scale.The reliabilities, accuracies, advantages and disadvantages of each model/method for describing liquid redistribution pore evolution, grain growth and densification are compared and evaluated. Some advice about the future developments of the numerical simulation for LPS is thus given.