Semi suppliers court MEMS, PV, emerging markets
07/01/2007
BY TOM CHEYNEY
Despite the relative health of a semiconductor equipment sector that topped $40 billion in 2006, market maturation and consolidation have driven many companies to seek new outlets for their products and technologies. The list of chip tool companies that have either diversified into-or increased their involvement in-related micro/nano industries has grown over the past few years. Key markets drawing the lion’s share of the equipment suppliers’ attention include MEMS and solar-cell photovoltaics (PV); emerging segments such as nanotechnology applications and flexible, printed, and organic (FPO) electronics have also gained traction in certain equipment companies’ strategic plans.
Service providers’ 30,000-ft overview
As the leading outsource provider of materials characterizations services to the semiconductor industry, Evans Analytical Group (EAG) has a unique overview of emerging or growing small tech market spaces. “EAG has experienced a marked increase in demand for analytical services from customers in solar/PV, MEMS, MOEMS, carbon nanotubes, and other related sectors,” says Vince Franceschi, executive VP of sales and marketing. “Areas of interest that we see most often include surface contamination or residue analysis, chemical composition, thin film analysis, defect or failure analysis, crystal structure/phase/orientation, and chemical changes before or after processing.
“We have clearly benefited from the migration of technologists from the semiconductor industry into these sectors, [since] these people are usually familiar with the types of services we provide. The solar/PV industry in particular has utilized materials and processes quite familiar to the semi industry, allowing us to leverage proprietary, well-defined testing protocols and deliver excellent data with confidence.”
Another high-tech veteran with a multi-industry perspective is Abbie Gregg, president of AGI, an engineering/consulting services company. She sees “many commonalities across these [MEMS, PV, and FPO] platforms that have to do with proper baselining of the process, appropriate technology documentation, understanding what levels of automation are appropriate for various levels of manufacturing, defect identification, and ensuring that reasonable design rules are used for new products.”
Gregg identifies design for manufacturing as a key weakness, but sees several opportunity areas for interested suppliers. “The compatibility of design rules with manufacturing is one area that is particularly difficult. The MEMS, PV, and FPO customers are in the early stages of understanding how to capture design rules as they develop and try to manufacture new products.
“For tool suppliers that want to work with individual companies, defect inspection and substrate material properties evaluation as well as metrology for organic materials properties are other areas that are key opportunities, with both hardware and software improvements needed. Better containers, [systems that handle] the filling and handling of organic fluids, are also a good bet for equipment suppliers as volume increases.”
The photovoltaic effect
In the solar-cell production equipment market, no company has made more noise over the past year than Applied Materials. Long the dominant semiconductor/flat-panel-display tool player and a dabbler in solar for about 10 years, Applied has a solar strategy that “significantly accelerated with the acquisition of Applied Films in July 2006,” says Craig Hunter, general manager of thin-film products in the company’s solar business group. Applied has scored design wins for its tools, which process thin-film solar-cells on large glass substrates, in major PV fabs planned or under construction in Germany, India, Spain, and elsewhere. But there are risks and rewards to trying to be the new sheriff in Solar City.
 Applied Materials’ ATON is a sputter deposition system for solar cell manufacturing designed for film uniformity and high throughput. |
“As a new entrant in an existing market, you are faced with a challenge of rapidly learning the needs of your prospective new customers,” says Hunter. “It can be very easy to misinterpret market requirements if you do not listen to the customers very carefully. Entrenched suppliers to the industry had the benefit of growing up together with the customer base and learning along the way.
“On the other hand,” he continues, “we have found that our scale, experience, and business developed from the semi/LCD industries are highly appreciated by potential solar customers. We are able to offer a level of service and customer support that has not been available in solar, and our existing global infrastructure...differentiates us from most participants in the industry.”
MEMS: CMOS’s not-so-little cousin
Versions of new and used semiconductor tools have long populated MEMS fabs, but some companies have developed systems specially designed for microelectromechanical-device manufacturing. This is not a new trend: EV Group, SUSS MicroTec, and STS have offered MEMS-specific equipment for years. But the increased attention by the likes of Tegal, Lam Research, Nanometrics, and ASML signals a more focused approach to the market, especially with positive signs pointing to the arrival of “true” volume production for MEMS devices such as silicon microphones for cellular applications, game-console gyroscopes, and digital micromirrors used in the display industry.
Like its CMOS cousin, MEMS manufacturing has always had its share of tricky etch and deposition requirements, with ever-more-stringent and challenging specifications. Tegal has been involved in the MEMS equipment market for more than 15 years. However, lately the company “has been focusing on increasing market demand of specialized [plasma etch and PVD] process tools used for the commercial manufacturing of MEMS devices,” according to Murali Narasimhan, vice president of marketing. Tegal has seen a growing interest by fabricators of RF and automotive MEMS, microfluidic devices, and other components.
“Our PVD technology is widely used by the acoustic filter industry leaders for depositing high-quality, piezoelectric aluminum-nitride films, with low nonuniformity and highly oriented crystal structures,” he continues. “This has resulted in bulk acoustic wave (BAW) and film bulk acoustic resonator (FBAR) devices exhibiting high-quality factors and low insertion losses. Proprietary temperature treatments, thickness profile control, stress control, and preclean sputter etch are critical in enabling high-performance FBARs in production.”
Another familiar name in the MEMS equipment space is the semiconductor etch-tool leader, Lam. The company has “been selling into the MEMS market for over 10 years” and has an “installed base of over 350 chambers,” says Jackie Seto, managing director of Lam’s MEMS business unit. She cites a range of complex challenges that differ from the CMOS process norm and depend on what kind of MEMS are being made, starting with profile control in accelerometers, gyroscopes, and resonators.
“The profile needs to be very symmetrical, with no change in the angle on either side of the critical feature,” explains Seto. “The profile across the wafer needs to be consistent, with no profile tilting. If the deep silicon etcher defining the critical features does not have a high level of profile control, the yield across the wafer will be low.” For other applications, she mentions etch-rate and sidewall surface uniformities as critical factors.
 Tegal’s AMS MMT AlN PVD cluster tool is designed for aluminum nitride and electrode deposition in the creation of bulk acoustic wave MEMS. |
MEMS metrology faces “more varied requirements than in CMOS applications,” due to factors such as thicker films, higher aspect ratios, larger features, and “the high rate of innovation in the field,” notes Nanometrics’ technical marketing manager, Paul Knutrud. He believes that metrology will become more indispensable as volume MEMS manufacturing lines proliferate.
“A number of MEMS fabs run very low production rates and don’t require full automation,” explains Knutrud. “These fabs either do without metrology [and are] running blind, buy used equipment, or purchase equipment that is less expensive and may not meet their needs in full. More and more fabs, however, are hiring engineers from the CMOS [industry] who understand the need for critical metrology.”
The lithography of MEMS devices has its own suite of specific patterning and overlay concerns, some of which may surprise a fab engineer accustomed to CMOS manufacturing. “Critical dimension uniformities (CDUs) for a mechanical structure, such as a gear or cantilever, are very tight, but the actual size of the structure may be quite large compared to integrated circuits,” says Keith Best, a director in ASML’s Special Applications group. “The mindset of a CMOS engineer would be that a large feature requires a tool with lower-resolution specifications. But in fact it is the opposite, because the CDU requirement is that of the advanced high-resolution tooling.”
Reverse technology transfer
The technology and expertise transfer can cut both ways, with equipment suppliers learning valuable lessons from their MEMS and PV work that can be fed back to their mainstream semiconductor products. ASML’s Best relates one example:
“During the early development of our MEMS 3DAlign option [which was built for double-sided wafer processing], we discovered that the hardware could also be used as an alignment solution for a variety of difficult front-side processes, such as CMP and thick epitaxy.
“In the case of thick epi, the alignment markers were typically destroyed when epitaxy thicknesses exceeded 20 microns. But with alignment markers on the backside of the wafers, the 3DAlign option gave CMOS process engineers an alternative alignment strategy to improve their process overlay.”
 ASML’s Twinscan exemplifies the arrival of “true” volume production for MEMS. |
Lam’s Seto believes that “all of our learning in deep silicon etch will be directly applicable for through-silicon-via (TSV) formation for 3D interconnects” in next-generation chip manufacturing. RF MEMS switches use hafnium-oxide and other high-k dielectric materials for bypass capacitors that require plasma dry etching, a skill set that Tegal can employ in its CMOS customers’ high-k applications. Nanometrics’ Knutrud says “there are a number of algorithms and processes that we developed for MEMS that have applications in CMOS and have allowed us to develop a better understanding of our equipment.”
Although Applied’s solar-cell manufacturing efforts are at “an early stage,” Hunter believes that “learning how to reliably handle thousands of 150-micron-thick wafers per hour should significantly enhance our existing automation know-how. Providing an integrated production line...[should] allow us to better understand the impact of tool performance on production line performance and help us develop better software products and easier-to-service hardware platforms.”
For further information on semiconductor equipment, materials, and services suppliers involved in the MEMS and photovoltaic markets, see www.semi.org/mems and www.semi.org/pv, respectively.