Important market trends may be missed because no data exist that reveal them. This situation confronted the US National Institute of Standards and Technology (NIST; Gaithersburg, MD) recently when it tried to determine the magnitude of industry demand for its metrology-related services and research results. The government agency found it lacked solid data on the scale of industry`s investment in metrology equipment, including optically based equipment, also on how fast investment was growing and in what areas.
In the absence of this type of information, NIST was unable to determine if industry`s requirements were shifting or how rapidly. NIST is charged with ensuring that key metrology-related technical issues do not become barriers to the growth of the US semiconductor industry, but it could not determine whether it is investing enough to accomplish this task.
A less than perfect way to gauge the adequacy of investment is to compare the growth of industry`s demand to the scale of NIST`s programs, but relevant information was not available even for that. Comprehensive data on the semiconductor industry`s total demand for metrology-related equipment did not exist for several reasons. First, metrology-related investments tend to be embedded in other technologies, hence are difficult to measure on a stand-alone basis. Second, many metrology suppliers are relatively small companies or are divisions of far-larger companies; these are hard to identify and harder still to obtain data from about their activities.
Study takes shape
To determine if it was possible to overcome these difficulties, NIST underwrote a small investigation to measure demand for metrology equipment. Technecon Analytic Research (Washington, DC) was commissioned to perform the pilot study. After lengthy discussions with NIST and semiconductor industry observers, the firm applied two independent methods to estimate the total US semiconductor industry metrology-related investment. One method built the total market size by aggregating data from the supplier side using data developed on 130 metrology suppliers. The second method estimated total metrology investment working from the demand side covering four major end-use categories: wafer fabrication facilities (wafer fabs), semiconductor equipment suppliers, semiconductor consumables suppliers, and analytical laboratories dedicated to supporting the semiconductor industry (captive analytical labs or fab labs).
Synthesizing the results of the two methods, the researchers found that the total metrology demand in the US semiconductor industry surged from well under $1 billion in 1990 to approximately $2.5 billion by 1996-increasing at a compound annual growth rate (CAGR) of 24%. By the year 2002, total metrology-related investments in the semiconductor industry could easily more than double in size to $5 billion.
Drivers of demand
What is driving the semiconductor industry`s demand for metrology equipment? In the 1990-1996 period, it was not the better-known segments of the semiconductor industry such as wafer fabs that were driving metrology demand. Their impact on total metrology demand in the semiconductor industry was relatively modest, with growth averaging only 11%. Instead the demand was a factor of the rapid expansion of fab labs, where growth was extremely rapid-averaging 35% per year. Fab labs are analytical labs colocated with semiconductor wafer fab clean rooms. Furthermore, as the number of fab labs rapidly increased, so did the demand for optically based measurement tools.
Looking ahead, the fab labs will continue to be important drivers of demand for optically related equipment but at a slowing rate. The major wave of new fab lab expansion is over, and only as new wafer fab facilities are constructed will new fab labs be built.
A major new wave of expansion in demand for metrology generally-and optically based technologies specifically-will be the insertion of in-line sensors into the next generation of 300-mm-wafer tool sets, starting around 2005. Many observers believe that in-line optical sensor capabilities will actually displace the need for the fully equipped fab lab. Whereas today only five major semiconductor tool families rely to some degree on optical-measurement-based capabilities, this number should grow. Hence, it will be the growing need of the semiconductor equipment segment for embedded optically based sensors that will drive future demand.
Pushing the optical envelope
NIST currently has a substantial number of optically related initiatives in place, with more than 17 technical divisions working on different optical measurement issues. Overall, NIST estimates that its current budget for projects specifically related to technical issues relevant to the semiconductor industry`s technology roadmap is around $30 million.
If the semiconductor industry`s shift toward greater optically based technical solutions is representative of a wider trend, this figure may not be sufficient. The projected changes in the semiconductor industry`s demand for metrology will mean more and more optical sensors, and advances in this technology will become increasingly a limiter on the industry`s ability to shrink feature sizes. Whether NIST can push this limiting envelope ahead fast enough will be an important issue for industry to monitor. o
WILLIAM F. FINAN is a principal at Technecon Analytic Research, 2445 M Street NW, Ste. 200, Washington, DC 20037; e-mail: [email protected].