Delta-doped diamond structures for high performance electronic devices

化学气相沉积（CVD）技术生产的合成金刚石具有极高的电子和热性能，这一发现使人们对其作为半导体材料的潜在用途感到相当兴奋。实验研究表明，载流子迁移率>3000cm2V-1s-1，热导率> 2000wm - 1k -1。据预测，该材料的击穿场强超过10 MVcm-1。这些数字表明，提供一系列可以克服的技术挑战，金刚石将特别适合作为半导体材料，在需要高频率，高功率，高温或高电压的情况下运行。该提案解决了“德尔塔掺杂”的新用途，以实现这种装置。在传统的器件技术中，对给定半导体内迁移率值大小的主要限制是导致载流子散射的电离杂质的存在。然而，正是这些电离杂质是n或p掺杂材料中自由载流子的来源。正是杂质与自由载流子的物理分离，使得散射减少，迁移率值增加，这是最近使用III-V半导体技术改进高频器件性能的核心。实现这一目标的一种方法是在同质结构中形成非常薄的高掺杂区域。如果掺杂层只有几个原子层厚，载流子将在靠近该层的区域内移动，而不是在该层的外部。载流子和产生它们的供体/受体原子之间的分离导致迁移率增强。d掺杂在其他体系中提供的优势将适用于金刚石，另外的特点是可以克服硼的大活化能问题，因为d层中需要非常高的浓度。然而，可以用于III-V型半导体生长的分子束外延（MBE）技术不能用于金刚石；使用等离子体增强CVD工艺的需要使实现原子尺度调制掺杂金刚石结构所需的方法显着复杂化。虽然Si和GaAs器件在固态微波器件市场占据主导地位，但它们无法与真空管的功率性能相匹配。金刚石作为半导体的一个驱动因素是人们对在小众应用中取代真空管的兴趣。固态替代方案的开发将具有许多优点，包括体积小、重量轻、工作电压低（与真空管设备相比）以及更强的稳健性。目前的真空管设计，如磁控管、速调管和行波管（TWT）通常体积庞大，往往易损坏，而且价格昂贵（微波炉的磁控管除外，其产量巨大，成本仅为10-20美元/千瓦）。如果金刚石的固有特性可以通过新型掺杂器件的设计和制造得到充分利用，它不仅可以与现有的宽带隙器件（基于SiC和GaN）竞争，还可以与整个射频（RF）产生市场上高达100 GHz的行波管竞争。在高电压下的功率控制是钻石装置的另一个潜在用途，可能来自拟议的研究方案。理论上，一个金刚石开关可以用来在接近50千伏的电压下开关电源。这是目前任何其他电子材料都无法实现的。

The combination of extreme electronic and thermal properties found in synthetic diamond produced by chemical vapor deposition (CVD) is raising considerable excitement over its potential use as a semiconductor material. Experimental studies have demonstrated charge-carrier mobilities of >3000cm2V-1s-1 and thermal conductivities >2000 Wm-1K-1. The material has been predicted to have a breakdown field strength in excess of 10 MVcm-1. These figures suggest that, providing a range of technical challenges can be overcome, diamond would be particularly well suited to operation as a semiconductor material wherever high frequencies, high powers, high temperatures or high voltages are required. This proposal addresses the novel use of 'delta-doping' to realise such devices.In conventional device technology a major limitation to the magnitude of mobility values within a given semiconductor is the presence of ionised impurities which cause carrier scattering. However, it is these ionised impurities that are the origin of the free carriers within n- or p-doped material. It is the physical separation of the impurities from the free carriers, such that less scattering occurs and mobility values increase, that lies at the heart of recent improvements in high frequency device performance using III-V semiconductor technology. One approach to achieve this the formation of very thin, highly doped regions within a homostructure. Provided the doped, or d, layer is only a few atom layers thick, carriers will move in a region close to, but outside, this layer. The resultant separation between carriers and the donor/acceptor atoms that created them leads to enhanced mobility. The advantages offered by d doping in other systems will be valid for diamond, with the additional feature that the problem with the large activation energy of boron can be overcome, as very high concentrations are desirable in the d-layer. However, the molecular beam epitaxy (MBE) techniques that can be used for III-V semiconductor growth cannot be used with diamond; the need to use plasma-enhanced CVD processes significantly complicates the approach needed to realise atomic-scale modulation-doped diamond structures.While Si and GaAs devices dominate the solid-state microwave device market, they cannot match the power performance of the vacuum tube. One driver for diamond as a semiconductor stems from an interest in replacing vacuum tubes in niche applications. The development of a solid-state alternative would have many benefits including small size, low weight, low operational voltage (compared with vacuum tube devices), and greater robustness. Current vacuum tube designs, such as magnetrons, klystrons, and traveling-wave tubes (TWT) are usually bulky, often fragile, and expensive (with the exception of magnetrons for microwave ovens, which are manufactured in huge volumes and cost only $10-20/kW). If the intrinsic properties of diamond could be fully exploited through novel delta-doped device design and fabrication, it could compete not only with existing wide-bandgap devices (based on SiC and GaN) but also with TWTs in the entire radio frequency (RF) generation market up to 100 GHz. The control of power at high voltages is another potential use of the diamond devices that may arise from the proposed programme of study. Theoretically, a single diamond switch could be used to switch power at voltages approaching 50 kV. This is not currently achievable with any other electronic material.