Commercial SiC Power Electronics

点击翻译按钮获取中文摘要

The current power semiconductor market ($10bn) is dominated by silicon. The performanceof which has improved steadily over the past 20 years, but has now reached the stage whereits fundamental material properties limit the capability and efficiency of power semiconductorsystems. The superior material properties of Silicon Carbide (SiC) have been recognised asthe way forward yet to date SiC transistors have very poor performance and production costsare prohibitive. This proof of concept proposal is to prove Anvil Semiconductor’s solutionsto the two main issues preventing SiC taking its potential market share ($2.6bn): the cost ofthe substrate and the ability to produce good quality gate oxides (and thus produce goodquality transistors).This work enables the development of low cost Silicon Carbide (SiC) power electroniccomponents suitable for mass market applications, the first of which are electric vehicles andpower inverters for domestic photovoltaics. Together these are predicted to account for 60%of the potential SiC market (2019).The current generation of SiC components based on the hexagonal form (4H-SiC) areinherently expensive; a 100mm wafer of 4H-SiC costs ~£1,000 compared to ~ £20 for a Siwafer. Further it is likely that 100mm will be the main 4H-SiC wafer diameter for at least thenext six years- restricting size based economies of scale. Thus 4H-SiC components areunlikely to be viable in cost sensitive applications. A low cost alternative route using a thinSiC layer deposited on a Si wafer has been demonstrated on 150mm wafers and provides aroute to even larger diameters (3C-SiC/Si). This has technical challenges but AnvilSemiconductors has some key IP which it needs to demonstrate has no technological barriersto prevent it making the components needed by system builders.An ideal application of SiC is the drive inverter for electric vehicles because of its improvedpower efficiency, increased robustness in high temperature environments and importantly thedramatic reduction of size/mass of the system. Yole (Appendix A) show that the use of SiCcomponents instead of Si would result in an overall system cost increase of $31; not anacceptable solution. However with the lower cost version of SiC proposed here, there wouldbe a system saving of ~$100. This would enable car designers to take advantage of the space,energy and robustness savings of SiC.Another application for low cost SiC devices is the power conversion in domestic scale(~5kW) photovoltaic panels. Although current Si-based inverters are very efficient (97%),they are bulky and expensive, and have slow switching speeds leading to the need forphysically large passive components and heatsinks. SiC components will reduce the size andcosts of the capacitors, inductors and system cooling needed. However, a componenttechnology with similar costs to Si is essential in order to achieve an overall reduction insystem cost, ie 3C-SiC/Si.