Interview

Abstract

Wooyeol Choi from the University of Dallas at Texas talks to Electronics Letters about his paper ‘450 × 580 µm2 pixel incorporating TX and coherent RX in CMOS for mm-wave active imaging using a single reflector’, page 982. Wooyeol Choi My main research field is integrated circuits (ICs) for high frequency operation. These are already a crucial part of everyday devices, such as cellular phones, wireless LANs, and automotive radars. With my research, I am trying to improve the performance of such devices, push the limits of the technologies, and explore new and exciting applications. Now, I am interested in potential applications of extremely high frequency (EHF) electromagnetic waves commonly known as millimetre (mm)-wave and terahertz. Many useful everyday applications, such as imaging through obstacles, high-data-rate wireless and wired communications, and gas sensing/electronic smelling with absolute specificity, should become possible using these parts of the spectrum. Millimetre-wave radars provide an effective way to ‘see’ objects through obstacles or harsh weather conditions. One of the fundamental ways to improve the lateral resolution of active imaging radar at a given form factor is to use a higher frequency. For instance, 300 GHz radars will have 3.8 times better resolution than conventional 79 GHz ones with the same form factor. However, existing radar systems at around 300 GHz are prohibitively bulky and costly for high volume consumer applications. With this work, my colleagues and myself are trying to demonstrate the feasibility of 300 GHz imaging radar using CMOS technologies that can remove these limitations so that this technology can hopefully be widely used in the near future. The paper reports a compact transceiver that can perform signal generation and detection concurrently at 260 GHz. A voltage controlled oscillator, which doubles as a transmitter and a local oscillator, a downconversion mixer, and an antenna are integrated into an area of less than half wavelength in width and height (450 × 580 µm2). This technology enables multi-pixel digital cameras which can see at mm-wave instead of visible lights without using external sources. More importantly, since the transceiver is implemented using CMOS technologies, it should be possible to integrate hundreds or thousands of the transceivers in a single chip or package, and manufacture a large number of them affordably. The most challenging part is achieving the useful performance of a radio that includes both transmitter and receiver in such a small area. This is even more challenging because we are using a CMOS technology whose transistors have almost no gain at 260 GHz. We have exploited the nonlinearity of the devices in the process technology to overcome the limitation. We are now working to make this pixel operate with a reflector to form an image. We are also working on a linear array of pixels. This will be followed by that on 2-dimensional arrays. This paper reported a compact pixel that has more than 10 times better sensitivity. In addition it provides phase information. Medium-to-long range active imaging through obstacles will be an immediate application. As the community figures out ways to improve the performance of pixel, this imaging technology can enable self-driving cars and other autonomous systems to work in harsh conditions such as fog, dust storm, smog, smoke and others. Since the first report of terahertz circuits using CMOS technologies in 2008, more than 10 to 100 times improvements have been achieved for almost all the performance metrics. It is clear now that it is possible to build affordable terahertz systems using CMOS technologies, and it is being transferred to various industrial partners who are investigating commercialization for everyday applications. I really hope within next ten years there will be products based on this technology for everyday applications.