Why do we need Millimeter Waves?


The aim of this research it to build a universal radio frontend, self healing RF circuits with Built In Self Test (BIST), and millimeter-wave circuits and systems from 30-300 GHz frequency range.

Current Trends in RF and Microwave Integrated Circuits Research (1)

RFIC and microwave-IC research can be divided into many areas such as ultra low power RF frontends, wide band circuits and cognitive radio design, the aim of this research it to build a universal radio frontend, self healing RF circuits with Built In Self Test (BIST), and millimeter-wave circuits and systems from 30-300 GHz frequency range. Within this article current trends in millimeter-wave research are addressed from devices performance metrics to highly integrated radio frontends. It also provides some design aspects and precautions for such high frequency circuits and systems.

There are two main applications for millimeter- wave systems, wireless communication as well as millimeter-wave radar and imaging. Fig. 1 shows the current trend in the wireless communication data rate and technologies. In the next few years Wireless LAN and wireless PAN data rates should be within the range of 1-10 Gbps. As the required data rate increases the RF channel bandwidth increases accordingly. Based on Shannon’s Theorem, the maximum data-rate of a communication channel, known as channel capacity, C, is related to the frequency bandwidth of the channel, BW, and the signal-to-noise ratio, SNR as in

C=BW.log2(1+SNR) (1)


The Federal Communication Commission (FCC) has allocated several frequency bands at millimeter waves for high data rate wireless communication. Fig. 2 shows selected parts of the FCC-allocated frequency spectrum. Also shown in this figure are frequency allocations for automotive radar applications. The radar azimuth resolution (perpendicular to the radar wave) and range resolution (in the direction of radar wave) are inversely proportional with the carrier frequency and bandwidth, respectively, explaining the choice of such high frequencies and bandwidths. While the 22-29 GHz frequency band is allocated for short-range applications such as park assist, stop-and-go, and blind spot detection, the 77 GHz band is used for the long-range automatic cruise control application. These radars are currently realized using compound semiconductor technologies and limited to higher end cars. Radar range resolution is inversely proportional to the bandwidth of the transmitted pulse. Therefore, the FCC has allocated a wide frequency spectrum around 24 GHz (22-29 GHz) for short range automotive radar applications. The FCC allocated frequency band allows using the ultra wideband (UWB) technology to achieve a higher resolution for short range vehicular sensing applications such as blind spot detection, side and rear impact sensing, blind spot detection, and stop-and-go. One desirable objective is pedestrian detection and protection as illustrated in Fig. 3.

New potential systems such as millimeter wave imaging and sub-THz chemical detectors are implemented in current silicon technologies with the application in astronomy, chemistry, physics and security. Those systems are designed for specific frequencies such as 90 GHz, 140 GHz, and 300 GHz; those are the attenuation windows of the millimeter wave spectrum at with the attenuation either minimum or maximum as depicted in Fig. 4. (previous sentence not understood) Another potential application for mm-wave technology is passive millimeter wave imaging. By detecting only the natural thermal radiation of objects in the mmwave band, images of objects can be formed in a very similar fashion as in an optical system. Either a group of receivers or a movable mechanical antenna is required to scan the area of interest. Unless special techniques are employed, due to the relatively large wavelength, the resolution of this approach is limited to objects on the order of a millimeter. The millimeter wave image is clearly able to penetrate through the fog and rain and provide a clear image. In security applications, passive (or active) millimeter wave images of a person can be used to find hidden weapons Unlike X-ray based imaging systems, which can only be used with limited dosage with living organisms, passive mm wave imaging does not use any additional radiation than what is naturally present.


Why Silicon for millimeter wave application?

The main premise behind using silicon at millimeter waves is the higher level of integration offered at a high yield that leads into lower cost systems. Over the relatively short span of five years, several highly integrated and complex millimeter wave systems have been reported by academia and industrial research labs such as IBM research [8]. These fully integrated chips consist of several thousand RF and digital transistors and on-chip passives in multi metal-layer silicon processes and include all the receiver, transmitter and even transceiver building blocks such as low noise amplifiers, mixers, voltage controlled oscillators, phase locked loops, power amplifiers, and in some cases on-chip antennas. Moreover, in many cases multiple receive and transmit paths are integrated in a single chip to realize fully integrated phased arrays.

If silicon technology has adequate performance to implement the front-end portions of the transceiver, the ability to integrate digital logic in CMOS increasing densities offers the opportunity to drastically lower overall system cost.

Lower cost could be the prime motivator for the use of BiCMOS or CMOS over III-V technologies. Again considering fT as a measure of performance, the SiGe BiCMOS HBT has comparable performance to the NFET at roughly twice the minimum feature size. For standalone RF functions, where area is dominated by passive devices and I/O pads, BiCMOS may be the lower-cost option despite the approximately 20% additional process complexity required to form the HBT. III-V transistor performance at substantially relaxed lithography dimensions is comparable with leading edge CMOS. So, again for purely RF devices, III-V implementations may offer a lower cost especially when utilizing existing designs and time-to-market is considered.However, when even modest amounts of digital logic are to be integrated, CMOS has a clear advantage as circuit density and chip size scale with the square of the minimum lithographic dimension.


Limitations at Millimeter wave frequencies for silicon technologies

The current silicon technologies suffer from high noise and lower output power at millimeter wave frequencies compared to III-V counterparts. It seriously limits the link budget of Gbps transmission. The SNR affects both the communication data rates and distance.

For a given distance, the received signal experience higher attenuation as the frequency increases. It is due to smaller antenna size higher absorption in air and other materials. In a multipath environment, multiple replicates of the transmitted signal that are reflected from various objects reach the receiver at different times with different amplitudes and phases, causing unwanted signal fading. The amount of attenuation due to unwanted multi-path effects depends on the size of scattering objects relative to the carrier frequency as well as their type and location. Both high intrinsic noise of the current silicon technologies and low received signal power due to attenuation and multi-path, results in lower SNR. And lower SNR is translated to either lower distance or data rate. Before discussing how to overcome these limitations let’s present the performance of the state of the art silicon technologies at millimeter wave frequencies.


Millimeter wave Silicon Devices

As stated earlier the performance of silicon technologies is inferior if compared to the III-V semiconductors technologies. They suffer from relatively low carrier motilities and hence low devices figure of merit (FOM). High resistive or semi insulating silicon substrate is very hard to implement resulting in lower isolation and higher substrate losses in passives devices and interconnects at millimeter wave frequencies.

However, the recent advances in silicon technologies driven by high performance digital circuits enhanced the performance of the active devices in millimeter wave frequencies. The performance of the active device is quantified by fT , fmax or NFmin . The performance is dramatically increased with geometry scaling and technology enhancements in both CMOS and SiGe HBT [7]. The roadmap of the cutoff frequency (fT ) comparing a number of IIIV semiconductor devices with the silicon CMOS NFET and SiGe HBT as taken from the 2006 ITRS are plotted in Fig. 5. It is evident that silicon technology currently exhibits small signal gains that are competitive with those of III-V transistors and are predicted to scale at least as quickly in the near-term future. Fig. 6 shows the different types of RF devices.


Active Devices

Bipolar Devices

Silicon Heterojunction Bipolar transistors offer some advantages compared to CMOS devices such as lower 1/f noise, higher output resistance and higher voltage capability for a given speed. The range of technologies on the market today offers HBTs with fT > 200GHz and sometimes fmax > 300GHz [9] as shown in Fig. 7.


CMOS devices

CMOS transistors follow the well known Moore’s Law of scaling, thus leading to always increasing functional integration. The 65nm node still uses poly silicon gate, but the carrier mobility is sometimes increased by using several technological solutions as described previously. As depicted in Fig. 8 , fT as high as 150GHz and 200GHz are reached in the 65nm node for Low Power (LP) and General Purpose (GP) devices, respectively.












[5]L.Yujiri,M. Shoucri, P.Moffa, “Passvie mm-Wave Imaging,” IEEE Microwave Magazine, vol. 4, issue 3, pp. 39-50, Sept. 2003.


[6]International Technology Roadmap for Semiconductors,





[8]P. Chevalier, et al., “Advanced SiGe BiCMOS and CMOS platforms for Optical and Millimeter-Wave Integrated Circuits,” IEEE CSICS 2006


Muhammad Aly El-Kholy

Microwave, milli-meter researcher , IHP

M.Sc Electronics and communications


Related posts:


Post a Comment

Your email is never published nor shared. Required fields are marked *


Facebook Comments