Integrated GaN-Diamond Microwave Electronics: From Materials, Transistors to MMICs

全球对能够提供远远超过当前技术的功率密度的高功率微波电子器件的需求正在增加。特别地，基于氮化镓（GaN）的高电子迁移率晶体管（HEMT）是高效军用和民用微波系统的关键使能技术，并且越来越多地用于低碳经济中的功率调节应用。这种材料和器件系统远远超过了现有Si LDMOS、GaAs PHEMT或HBT技术所允许的性能。GaN基HEMT已经达到高达40 W/mm的RF功率水平，并且频率超过300 GHz，即，出色的性能使许多系统应用程序能够进行中断性更改。然而，晶体管的可靠性是由电场和沟道温度驱动的，因此自加热实际上意味着可靠的器件只能在高达10 W/mm的RF功率密度下工作，而文献中公布的数据为40 W/mm。英国和整个欧洲也存在相当大的担忧，即最先进的GaN微波技术受到美国ITAR（国际武器贸易条例）限制的限制。GaN HEMT技术的所有元素的最先进的能力，使用传统的SiC衬底，外延和器件处理目前驻留在美国，与英国行业的限制访问。集成GaN金刚石微波电子的愿景：从材料、晶体管到MMIC（GaN-DaME）是开发变革性的GaN-on-Diamond HEMT和MMIC，该技术超越了GaN-on-SiC，这将彻底改变目前限制GaN电子器件的热管理。如何将这种不同的材料集成到可靠的设备技术中存在挑战。结果将是与GaN on SiC相比RF功率增加> 5倍的器件，或者可替代地并且同样有价值地，MMIC或PA尺寸的显著“阶跃变化”收缩，并且因此通过去除有损组合网络以及降低功率放大器（PA）成本来提高效率。这代表了能力的颠覆性变化，将允许实现新的系统架构，例如RF导引头和医疗应用，并实现提供5G及更高带宽所需的带宽。冷却需求的降低/可靠性的提高将在系统一级节省大量成本。为了使我们的愿景成为现实，我们将开发新的金刚石生长方法，最大限度地提高接近有源GaN器件区域的金刚石热导率。在目前的金刚石上GaN器件中，在GaN表面上需要薄的介电层，以使得能够将金刚石播种和成功沉积到GaN上。不幸的是，这些器件中的大部分热障存在于GaN-电介质-金刚石界面处，其具有比期望的差得多的热导率。通过消除对用于金刚石生长的介电晶种层的需要，或者通过优化晶种附近的金刚石的晶粒结构，这种热阻的任何降低都将具有巨大的益处。新的金刚石生长将与使用相变材料的创新微流体相结合，这是一种比传统微流体更强大的方法，以进一步帮助热量提取。使用金刚石的一个未讨论的结果，其低介电常数，这对微波设计提出了挑战和机遇，将被利用。在最基本的层面上，这项技术的可靠性尚不清楚。例如，在材料水平上，金刚石和GaN具有非常不同的热膨胀系数（CTE）。将需要开发机械刚性界面，包括叉指型GaN-金刚石界面。

Global demand for high power microwave electronic devices that can deliver power densities well exceeding current technology is increasing. In particular Gallium Nitride (GaN) based high electron mobility transistors (HEMTs) are a key enabling technology for high-efficiency military and civilian microwave systems, and increasingly for power conditioning applications in the low carbon economy. This material and device system well exceeds the performance permitted by the existing Si LDMOS, GaAs PHEMT or HBT technologies. GaN-based HEMTs have reached RF power levels up to 40 W/mm, and at frequencies exceeding 300 GHz, i.e., a spectacular performance enabling disruptive changes for many system applications. However, transistor reliability is driven by electric field and channel temperature, so self-heating means in practice that reliable devices can only be operated up to RF power densities of 10 W/mm in contrast to the 40 W/mm hero data published in the literature. Considerable concern also exists in the UK and across Europe that access to state-of-the-art GaN microwave technology is limited by US ITAR (International Traffic in Arms Regulation) restrictions. The most advanced capabilities for all elements of GaN HEMT technology, using traditional SiC substrates, epitaxy and device processing currently reside in the US, with restricted access by UK industry.The vision of Integrated GaN-Diamond Microwave Electronics: From Materials, Transistors to MMICs (GaN-DaME) is to develop transformative GaN-on-Diamond HEMTs and MMICs, the technology step beyond GaN-on-SiC, which will revolutionize the thermal management which presently limits GaN electronics. Challenges occur in terms of how to integrate such dissimilar materials into a reliable device technology. The outcome will be devices with a >5x increase in RF power compared to GaN-on-SiC, or alternatively and equally valuably, a dramatic 'step-change' shrinkage in MMIC or PA size, and hence an increase in efficiency through the removal of lossy combining networks as well as a reduction in power amplifier (PA) cost. This represents a disruptive change in capability that will allow the realisation of new system architectures e.g. for RF seekers and medical applications, and enable the bandwidths needed to deliver 5G and beyond. Reduced requirements for cooling / increased reliability will result in major cost savings at the system level. To enable our vision to become reality, we will develop new diamond growth approaches that maximize diamond thermal conductivity close to the active GaN device area. In present GaN-on-Diamond devices a thin dielectric layer is required on the GaN surface to enable seeding and successful deposition of diamond onto the GaN. Unfortunately, most of the thermal barrier in these devices then exists at this GaN-dielectric-diamond interface, which has much poorer thermal conductivity than desired. Any reduction in this thermal resistance, either by removing the need for a dielectric seeding layer for diamond growth, or by optimizing the grain structure of the diamond near the seeding, would be of huge benefit. Novel diamond growth will be combined with innovative micro-fluidics using phase-change materials, a dramatically more powerful approach than conventional micro-fluidics, to further aid heat extraction. An undiscussed consequence of using diamond, its low dielectric constant, which poses challenges and opportunities for microwave design will be exploited. At the most basic level, the reliability of this technology is not known. For instance, at the materials level the diamond and GaN have very different coefficients of thermal expansion (CTE). Mechanically rigid interfaces will need to be developed including interdigitated GaN-diamond interfaces.