Long-term outlook for MEMS: New applications and markets*
07/01/1999
Although a majority of industry observers are predicting rapid growth in microelectromechanical systems (MEMS) sales over the next five years, well-established MEMS product lines - airbag accelerometers, automotive engine manifold absolute pressure sensors, blood pressure sensors, and thermal inkjet printer heads - are expected to grow only moderately, due to high levels of market penetration. Also, significant new applications based on some mature product lines like pressure-sensing devices, for example, do not appear to be on the horizon. Rather, existing devices are likely to be modified to cover a larger number of applications. Several new products that may enter the market within the next two years, however, have the potential for rapid development and sales growth (see figure). This report assesses commercial opportunities for MEMS, forecasts future market size, and discusses future technology challenges.
 MEMS dollar sales by technology area; a) estimated for 1996, and b) projected for 2003. |
MEMS, devices produced by bulk, surface, and high-aspect-ratio micromachining methods, may be grouped into six technology categories: inertial measurement, microfluidics, optical MEMS, pressure measurement, radio frequency (RF) MEMS, and other technologies. The current market for MEMS is dominated by pressure sensors, accelerometers, microgyros, inkjet nozzles, and hard disk drive heads. Most consumer sales of these devices are in the IT (inkjet nozzles, hard drive heads) and automotive (pressure sensors, accelerometers, microgyros) sectors.
Emerging markets
Promising new applications of MEMS technologies include displays, chemical testing systems, rate sensors, and MEMS-based RF communications devices, the report says.
Displays
Sales for MEMS-based display systems are currently small, but optical MEMS have the potential to become a major application area. MEMS displays use digital light processing, which eliminates the need for digital-to-analog conversion in processing a video signal for display. There are three kinds of these digital "light switches" either on the market or nearing commercial viability. The only device presently on the market operates by using micromirrors to reflect light in either of two directions, determined by the state of the memory cell associated with the micromirror. In the on state, the light is reflected into the projection system optics. It is still a niche application, mostly for projection displays.
The other two kinds of light switches use tiny membranes suspended over an air gap. The pixel is turned on or off by manipulating these membranes. One device operates using the principle of diffraction; the other, using iridescence.
 |
Other emerging optical MEMS applications include optical switching devices, which may reach the commercial market within a year or two; and micromirror-based adaptive optics, which uses multiple micromirrors, each mounted on an actuator. This system could find use in fiber optic communications but is likely to remain a niche application. Major application areas for optical MEMS displays are projection displays, portable communications devices, other handheld devices, and instrumentation displays. The principal application area for optical switches is the communications network. Thus, the report says, sales of MEMS displays have the potential to grow by both increasing their market share and participating in the overall growth of the display market. Total sales of packaged MEMS optical devices are estimated at $440-950 million for 2003, with the majority of sales coming from displays and optical switches (see table).
Chemical testing systems
Based on microfluidics technology, chemical testing systems such as lab-on-a-chip (LOC) devices are a potentially major marketplace application of MEMS. The development of gene chips and related DNA analysis tools will enable the decoding of DNA sequences 1000x faster than possible with previously existing methods. At least one type of gene chip is in commercial production, but sales have been limited so far. Another type of LOC is based on optical analysis of blood components. Other MEMS medical devices either available or in production include a product that measures blood chemistry using micromachined, thin-film sensors for specific blood components, a glucose sensor, and a chip-based microflow cytometry device. The report goes on to say that MEMS-based gene detection devices could make up as much as 30-40% of the microfluidics market.
Rate sensors
These appear to be the most promising new type of inertial measurement sensor. Applications include stabilization devices in consumer products like camcorders, practical only if MEMS rotation sensors are in mass production at low cost. Another application of rate sensors would be car antilock braking systems or, possibly, car suspension systems. A smaller emerging market in pacemakers would monitor a wearer's activity and adjust a pacemaker accordingly. Industry observers also believe that by 2003, microaccelerometers will be used in a variety of games for new kinds of joysticks. Total sales of packaged MEMS inertial measurement components are estimated to reach $700-1400 million in 2003, with the majority of sales coming from automotive applications.
RF communication devices
According to industry observers, there are four kinds of MEMS-based devices on the horizon that would find application to RF systems: RF switches, tunable capacitors, inductors, and micromechanical filters. (The inductors would not likely have any mechanical components but would be manufactured using MEMS lithography techniques.) All of these devices are presently in the developmental stage, with RF switches close to having viable prototypes. RF switches could reach the marketplace in about two years. The other devices are probably three or four years from completion. Total sales of packaged RF MEMS components are estimated at $40-120 million in 2003, with the majority of sales coming from cellular phones and other wireless devices for the consumer market. The report was unable to project the growth rate for sales since the baseline sales are essentially zero.
Obstacles to commercialization
The report also cites a number of obstacles to MEMS commercialization, including packaging, integration, the need for computer-aided design (CAD), and low sales.
One significant obstacle in commercializing MEMS sensors involves packaging. A MEMS sensor needs to be protected from hazardous conditions such as corrosive fluids, gases, or water vapor, yet be allowed to interact with its environment enough to perform the intended physical measurement - a difficult challenge. Also, the packaging tends to be highly application specific, making standardization difficult. Seventy-five percent or more of the cost of a MEMS device is for packaging and testing; some industry observers feel that MEMS developers have not given sufficient attention to these issues. A related issue is the lack of standardized test methods and equipment. Many MEMS testing systems are custom built and expensive.
Another issue is the difficulty of integrating the sensor and the associated electronics on the same chip. Such integration offers several potential advantages such as small size and placement of the electronics and the sensor near each other. Integration is technically feasible but poses practical difficulties, such as a larger number of manufacturing steps and a resultant higher cost. The increased number of manufacturing steps also increases the possibility of manufacturing defects. Yet another challenge of micromachining is that scaling laws traditionally used in engineering do not work at micro scales; sophisticated CAD tools are required for modeling. MEMS products typically have long development times - on the order of 10 years or more. This is at least partly due to the fact that designs tend to be application specific.
In many market niches, the report states, low sales (thus far) have meant high fixed costs and high prices. For applications where MEMS is a critical enabling technology, customers are often willing to pay high prices. Market areas with high sales (e.g., accelerometers) are highly price competitive, and profits are low or nonexistent. Also, many MEMS products represent what are called discontinuous innovations, that is, they are not simply an incremental improvement on a previously existing technology. These products require some new thinking on the part of developers.
Conclusion
Unlike the microelectronics industry, where the US and Japan were the dominant participants in the early years, MEMS is emerging as a global industry. Europe and Asia have substantial MEMS activity, and US technological pre-eminence is far from assured. Also unlike the semiconductor industry, the MEMS industry has no systematic means of gathering and disseminating information on the marketplace. As a result, MEMS companies often lack the information they need to make informed decisions on commercial opportunities. The report concludes it will be difficult for MEMS to evolve from a niche market into an established, mainstream technology unless they are widely incorporated into high-sales-volume consumer products.
*This article is a summary of hte report "MEMS 1999: Emerging Applications and Markets" from System Planning Corp. (SPC). Data are based on interviews with industry participants and observers and estimates of market size from other studies. For a copy, contact Bill Detlefs, senior research analyst, SPC, 1429 North Quincy Street, Arlington, VA 22207; ph 703/351-8292, e-mail [email protected], http://memsmarket.sysplan.com.