Low-cost packaging is priority for millimeter-wave MMICs

Figures released by ABI Research - a New York-based technology research think tank and consultancy - reveal that power amplifiers for wireless networks represented 68% of the total RF market at the end of 2003, valued at almost $2 billion, but that proportion is set to fall to 49% by 2008. According to Lance Wilson, senior analyst for ABI, the only option for RF-device manufacturers is to diversify: "No matter whether the [cellular] business goes up or down, the price pressure on [the device manufacturers] is so great that their problem is one of revenue. Their average selling price is falling, and they are converting to lower-cost packaging that will drive prices down even further."

These challenging market conditions are prompting many MMIC suppliers to apply their existing know-how to MW frequencies at 20-100 GHz. A number of applications exist, but only a few commercial products are currently available. Indeed, TriQuint of the US - one of the world s leading MMIC suppliers - estimates that the market for commercial MW MMICs based on GaAs technologies is set to grow from $163 million in 2003 to $400 million in 2007.

MMICs operating in the MW spectrum offer several major benefits over their low-frequency counterparts. For a start, the higher operating frequencies enable more information to be encoded in the signal, which make MW solutions ideal for high-bandwidth, high-capacity communications systems. High-frequency MMICs also work for applications requiring narrow antenna beams or high spatial resolution in a compact size, since the shorter wavelengths of MWs enable the use of smaller receiver and transmitter elements.

These advantages have already been put to good use in the space and defense sectors, but continuing advances in circuit design are enabling smaller and cheaper components to be developed for commercial end-users. For example, vendors of fixed-wireless equipment, such as Terabeam Wireless, Ceragon Networks and Bridgewave, are exploiting MW MMICs in 60 GHz communications links that support data rates in excess of 1 Gbit/s - enough for enterprise customers to connect their internal Ethernet networks to the fiber infrastructure. Systems operating beyond 70 GHz are now being developed to provide data rates of up to 10 Gbit/s with a range of up to one mile.

Satellite network operators are also eager to harness the benefits of higher frequencies. Two-way communications based on very-small-aperture terminal satellite technology at 12-18 GHz already represents a significant commercial market for microwave MMICs, while next-generation systems at 26-40 GHz are set to be deployed in 2005 and 2006. These deployments, which will offer better services at lower cost by enabling smaller antennas, dynamic bandwidth allocation and narrower beams, are likely to fuel strong demand for MW MMICs through 2008.

For true mass-market success, however, MW suppliers are looking to the emerging market for automotive anticollision radar at 76-77 GHz. Radar-based systems providing adaptive cruise control are now being fitted to luxury cars - such as the Mercedes-Benz S-Class and the Volkswagen Phaeton - and falling component costs look set to open up the mid-price car market. Dan Green, TriQuint s director of broadband technology, estimates that 4.5 million MW MMICs will be deployed for automotive radar applications by 2007, while total MW MMIC usage in this market could approach 100 million units by 2010.

There is no doubt that MW MMICs offer serious potential in the commercial market-place, but the challenge for suppliers is to deliver packaged and tested parts that meet the performance, cost and size criteria demanded by commercial end-users.

These demands are already being addressed at the chip level by the major manufacturers of MW MMICs. "The latest generation of MMIC manufacturing processes can deliver the gain and bandwidth performance required for applications at frequencies of up to 100 GHz, while further optimization of MW circuit designs are likely to reduce both the size and cost of MMIC solutions," said Jeff Powell and Dave Bannister of the MMICs and Millimeter-Wave Modules group at QinetiQ.

According to Powell and Bannister, the leading MMIC foundries are ready to introduce price structures that would allow significant rate reductions for high-volume orders, even with existing designs. But the big obstacle remains the cost of packaged MW transceivers, which are often about five to six times as expensive as bare MMIC die. "Only by reducing packaging and assembly costs can the MW community hope to achieve high-volume sales of MMIC devices and transceivers," said Powell and Bannister.

To tackle these issues, package designers are now investigating whether low-cost, lightweight materials could replace the milled-metal packages that have traditionally been employed for MW components. Most attention is focusing on laminates and low-temperature co-fired ceramic (LTCC) materials, which have already been exploited in MMIC technology at lower frequencies.

One crucial advantage of LTCC and laminate materials is their ability to be formed into multilayer packages that can replace a number of individual modules with a common substrate incorporating both passive components and active devices. This approach reduces the overall size of the MMIC subsystem, minimizes parasitic losses caused by electrical interconnections, lowers assembly costs and enables high-level integration of functions demanding different device technologies.

For example, Kenjiro Nishikawa and colleagues at NTT in Japan have exploited laminate multilayers to develop a fully integrated transmitter operating at 57-60 GHz and requiring a chip area of just 2.89 mm2. The group has also built a 5.04 mm2 integrated receiver, achieving a noise figure of less than 6 dB.

To build these devices, Nishikawa and colleagues form a 3D interconnection layer - consisting of four layers of 0.25 μm-thick polyimide film and 1 μm-thick metal (0.2 μm for the top-level metal) - on top of European MMIC foundry United Monolithic Semiconductor s (UMS s) commercial 0.15 μm PHEMT process technology (figure 2). Once the process is complete, the PHEMT has a cut-off frequency of 110 GHz.

This type of integration alleviates many of the headaches associated with interconnecting MW components, since traditional wire bonds lead to high parasitic reactances at these frequencies and often require labor-intensive assembly techniques. In contrast, multifunction modules enable low-loss interconnects to be made between the devices and the package as part of the fabrication process.

However, connecting the package to the next-level subsystem remains a major problem. One of the most promising solutions is flip-chip bonding, which allows low-loss interconnects to be achieved with automated pick-and-place machines that also ensure high throughput and yield.

For example, researchers at NEC in Japan have combined flip-chip bonding with an LTCC multilayer package to build 60 GHz MMIC modules for wireless applications. Eudyna Devices has also indicated that flip-chip modules might be used for automotive radar systems, while fellow Japanese companies Matsushita and NTT are investigating the potential of the technology for high-performance packages.

Other promising interconnect technologies include ball-grid arrays, which exploit an array of small metallic balls to provide a reliable and low-cost interconnect between a ceramic package and a circuit board. Meanwhile, UMS is using coupled electromagnetic transitions between the chip and the substrate to provide W-band interconnects in a module designed for automotive radar at 77 GHz (see MMICs provide the key for lower-cost automotive radar).

These developments are certainly promising, but more work is needed to develop standard packaging solutions that are compatible with high-volume manufacturing techniques. One problem is that most module makers are aligned with their end-user market sectors, which reduces the opportunity for transferring new ideas between subsystem manufacturers serving different markets. It also creates a general lack of appreciation among the end-user community of the importance of packaging in low-cost MMIC solutions. These issues are now being addressed by the International Wireless Packaging Consortium (IWPC), which aims to bring together MMIC suppliers and system manufacturers to close the "packaging knowledge gap".

According to the IWPC, the introduction of cost-effective packaging technologies has been hampered by poor communication between the foundries, the component and module makers and the system manufacturers. Although the organization covers all wireless technologies, the packaging challenges at MW frequencies have made this a crucial area of activity. For example, a recent workshop brought together device manufacturers, subsystem suppliers and end-users in the automotive industry to share ideas on the development of MW chips and modules for low-cost radar systems.

The IWPC is certainly attracting plenty of interest, with almost all the industry heavyweights appearing in its membership list. By raising the profile of packaging issues along the whole supply chain, it hopes that suppliers will better meet the needs of their customers and, in so doing, reduce unit costs, deliver the required performance and speed up the time to market.

• These and other issues related to MW MMICs are explored in detail in a new report, Commercial Applications for Millimeter-Wave MMICs. The report is published by Technology Tracking, a partnership between QinetiQ, Europe s largest science and technology organization, and Institute of Physics Publishing, the publisher of Compound Semiconductor magazine. See www.technology-tracking.com for more details.

K Nishikawa et al. IEEE GaAs IC Digest 2003 p97.
M Ito et al. IEEE MTT-S Int. Microwave Symp. Digest 2000 p57.