电子器件的热流密度越来越大，已远超风冷、热电制冷等技术的性能极限，因此散热效率高、易于集成的微纳米通道热沉成为国际研究热点。由于提高温度均匀性及减小流动损耗的难题，亟需新结构微通道热沉来满足散热需求。本文提出一种将歧管和二次流道相结合的新型混合结构微通道热沉，使用模拟优化和实验手段对其流动和传热特性开展了较系统的研究，另外，毛细填充是实现纳米通道热沉热量排散的基础，本文对纳米通道中的毛细填充开展了模拟和实验研究。提出微通道热沉的歧管-二次流混合设计结构，通过歧管降低压损并强化换热，通过二次流道增强扰动并减薄边界层。模拟发现混合结构热沉有“优化设计区”，在该区域，热沉的压损?P和总热阻Rt可被同时降低，当雷诺数为295时，混合结构热沉可使?P降低1.91%的同时，Rt降低19.15%。搭建了“试验设计+Pareto图分析+近似建模+遗传算法+TOPSIS”的模拟优化系统，可对完整混合结构热沉单元自动取样并优化，大幅提高了优化计算效率。提出了基于Pareto优化的不同类型热沉综合性能的对比评价方法，对于给定工况的散热问题，该方法以Pareto前沿为基础，通过比较优化设计区的大小即可判断热沉性能的优劣，该方法进一步证实混合结构热沉的散热性能相比单一强化结构更优。采用光刻、刻蚀等工艺加工了优化混合结构的深为60 μm和100 μm的嵌入式硅基微通道热沉，开展了单相流动传热实验，证实混合结构热沉有“优化设计区”，对于60 μm深热沉，当流量为580 mL/min，热流密度为20 W/cm2时，混合结构热沉可使?P降低11%的同时，Rt降低24%。流量增大会使混合结构与歧管热沉的热阻比(Rt/Rt0)减小而压损比(?P/∆P0)增大，不过即使对于最大实验流量580 mL/min（Re≈900），压损比仍然小于1，这表明混合结构热沉具有优异的流动特性。模拟发现纳米通道热沉可排散的热流密度比微通道热沉高一个数量级，且毛细填充越快，散热效果越好。对硅纳米通道热沉中的毛细填充机理开展了实验研究，发现实验斜率Aexp小于Lucas-Washburn(LW)模型预测值ALW，分析表明动态接触角是造成偏离的主要原因。基于实验结果修正了LW模型，与本文及文献中的实验数据符合较好，表明当通道越浅或毛细填充速度越快时，实验斜率相比理论斜率越小。建立了适用于纳米通道极低毛细数(10-6~10-4)的动态接触角模型，表明当通道越浅或毛细填充速度越快时，动态接触角相比静态接触角越大。

The increasing heat dissipation of electronic devices has been beyond the performance limits of traditional cooling technologies, such as air cooling and thermoelectric refrigeration. Therefore, micro- and nanochannel heat sinks with high heat removal efficiency and compact structure have already become an international research hotspot. New microchannel structures are urgently needed to meet the heat dissipation demand for the temperature uniformity and pressure loss. In this thesis, a novel hybrid microchannel heat sink combining manifold arrangement with secondary oblique channel (MMC-SOC) is proposed. A systematical study on the flow and heat transfer characteristics of the hybrid heat sink is conducted using the numerical simulation, geometry parameters optimization and experimental methods. In addition, the capillary filling process is the basis of heat dissipation for nanochannel heat sinks. The capillary filling process in nanochannels is studied by numerical and experimental methods.A hybrid microchannel heat sink with manifold arrangement and secondary oblique channels is proposed. The manifold structure is used to reduce pressure loss and enhance heat transfer, and the secondary channel can promote fluid mixing and decrease the thermal boundary layer thickness. The simulation results show that the hybrid microchannel heat sink has a Design Optimization Area (DOA), where the pressure drop ?P and the total thermal resistance Rt can be both reduced compared to manifold microchannel heat sink. The hybrid heat sink can reduce the pressure drop by 1.91%, and simultaneously decrease the total thermal resistance by 19.15% at Re=295. A simulation and optimization system, which combines design of experiment, Pareto chart analysis, approximation modeling, genetic algorithm method and TOPSIS, has been built. Sampling of complete hybrid heat sink unit can be performed automatically, which improves the optimization efficiency greatly. An overall performance evaluation method of different types of heat sinks based on Pareto optimization is proposed. For a cooling problem under given working conditions, the proposed evaluation method can judge the performance of various types of heat sinks by comparing the DOA based on the Pareto front. The proposed evaluation method further proves that the hybrid heat sink has superior performance compared to traditional passive enhancement strategies.Lithography, etching and other processes are used to fabricate the optimized hybrid microchannel heat sinks embedded in sililcon with heights of 60 μm and 100 μm. The DOA is also verified by single-phase experimental test. The experimental results show that for channel height of 60 μm, an 11% reduction in ?P and a 24% reduction in Rt are obtained for a volume flow rate of 580 mL/min and a heat flux of 20 W/cm2. Increasing volume flow rate will bring smaller total thermal resistance ratio (Rt/Rt0) and larger pressure drop ratio (?P/∆P0). The values of (?P/∆P0) are all smaller than 1 under the working conditions, even though the volume flow rate reaches up to 580 mL/min (Re~900), indicating that the proposed hybrid microchannel heat sink has a superior hydraulic performance.The simulation results show that the heat flux that can be dissipated by nanochannel heat sink is one order of magnitude higher than that of microchannel heat sink. The nanochannel heat sink will have an enhanced heat removal capability if the capillary filling is faster. The mechanism of the capillary filling in silicon nanochannel heat sink is studied experimentally. The extracted experimental slope Aexp is smaller than the prediction ALW of the theoretical model Lucas-Washburn (LW), which is mainly caused by the dynamic contact angle. A modified LW model is obtained based on the experimental results and it can describe the experimental results of our study and other literature well. The new model shows that Aexp will get smaller than ALW if the channel height decreases or the capillary filling speed increases. A new dynamic contact angle model is also developed for very low capillary number (10-6~10-4) in nanochannels. The model shows that the dynamic contact angle will get larger than the equilibrium one if the channel height decreases or the capillary filling speed increases.