Efficient cooling of a high performance computer chip has been an extremely important however becoming more and more tough issue. The recently invented liquid metal cooling method is expected to pave the way for high flux heat dissipation which is hard to tackle otherwise by many existing conventional cooling strategies. However, as a new thermal management method, its application also raised quite a few challenging fundamental and practical issues for solving. To illustrate the development of the new technology, this talk is dedicated to present an overview on the latest advancements made in the author’s lab in developing the new generation chip cooling device based on the liquid metal coolant with melting point around room temperature. The designing and optimization of the cooling device and component will be discussed. Several major barriers to prevent the new method from practical application such as erosion between liquid metal coolant and its substrate material will be outlined with good solutions clarified. Performance comparison between the new chip cooling method with commercially available products with highest quality such as air cooling, water cooling and heat pipe cooling devices were evaluated. Typical examples of using liquid metal cooling for the thermal management of a real PC or even super computer will be demonstrated. Further, miniaturizations on the prototype device by extending it as a MEMS cooling device or mini/micro channel liquid metal cooling device will also be explained. Along with the development of the hardware, some fundamental heat transfer issues in characterizing the liquid metal cooling device will be discussed through numerical or analytical model. Future challenging issues in pushing the new technology into large scale practices will be raised. From all the outputs obtained so far, it can be clearly seen that the new cooling strategy will find very promising and significant applications in a wide variety of engineering situations whenever thermal managements or heat transport are needed.
Anticipated transients without scram for light water reactors: implications for liquid metal fast breeder reactors
