Rapid thermal processing market
07/01/1998
Rapid thermal processing market
John Salzer, Salzer Technology Enterprises Inc., Santa Monica, California
The RTP market has withstood the uncertainties of the marketplace better than hot-wall furnaces or, indeed, better than the overall semiconductor equipment market. In 1997, a weak growth year for the total field of semiconductor production equipment, RTP still grew over 20%, just as we expected it would a year ago. Table 1 shows the recent growth patterns of the RTP market divided into its three segments - basic RTP, RTEpi, and RTCVD. Basic RTP employs heat to affect such processes as annealing, silicidation, glass reflow, and oxidation. RTEpi is a single-wafer epitaxial deposition method where heat is applied by RTP techniques. RTCVD is a CVD process employing rapid thermal heating.
 |
This article reports some of the findings in Salzer Technology Enterprises` annual worldwide survey and report of the RTP field.
Still an emerging market
With single-wafer and small-batch chambers as well as cluster tools becoming the preferred equipment configurations, large batch furnaces have found a more limited role. They are better suited to large-scale production, such as DRAMs; are less practical with larger wafers, particularly 300 mm; are not cost-competitive with processes that have shorter times at temperature; and are facing increasing difficulties with uniformities as geometries recede below 0.25 ?m.
So what is the outlook for 1998? One of the reasons that 1997 performed as well as it did was that it started to gain strength after the low point of the second half of 1996. As it turned out, 1997 showed quarter-by-quarter improvements, but turned in a more moderate performance than 1996.
Although 1997 appeared to be a hopeful year, the Asian economic crisis put us on notice about the uncertainties of projecting economic data. Even now, the world remains confused about the global effects of the Pacific turmoil. Our projection for 1998 is for modest growth in the RTP market - more modest than even in 1997. But one thing considered certain is that RTP will outpace the overall wafer processing equipment market.
There are several other reasons that the RTP market outperforms the overall equipment market. First, as an emerging technology, RTP finds its way into new markets as its various applications are developed. Second, the thermal equipment market is very large, and RTP`s penetration is still very low. Third, while mature equipment with a sizable installed base depends mostly on quantity purchases for higher production rates - which is rare during a market slow-down - newer equipment is bought for both development and production.
Table 2 shows some dollar figures supporting the indicated growth rates for all RTP. In 1997, the total RTP market passed the half-billion dollar mark. Beyond 1998, the equipment markets will be in a recovery phase, and the total RTP business will grow beyond $1 billion by 2000.
Broadening technology base
The core technology of RTP equipment was, and still is, a combination of tungsten/halogen lamps applying radiant heat to the wafer; the use of one or more pyrometers to remotely sense the emission from the wafer; the use of a thermocouple-instrumented wafer to calibrate the temperature control system; and sophisticated software to correct for the emissivity characteristics of the wafer.
Within this framework, there are still significant variations. The lamps may be long and linear, lined up parallel to the wafer`s surface. An alternative solution uses short bulbs perpendicular to the wafer arranged in circles or near-circles. These have the valid claim that a circle matches the wafer`s shape and should provide better uniformity radially. But rotating the wafer and other compensations with linear lamps may achieve the same effect. Furthermore, in a circular arrangement, uniform illumination has to be designed between circles and between the discrete lamps within each circle.
Where temperature is measured by one or more pyrometers, the control software can be quite elaborate, and companies have distinguished themselves by their level of success in providing emissivity corrections and other control features. The lamps can be controlled in zones, with single-lamp zones providing the most versatile control possibilities.
A great simplification is achieved by measuring wafer temperature with a thermocouple (TC) in direct contact with the wafer, which is an alternative offered by AG Associates and Dainippon. Contamination by the TC and TC wires has been circumvented by placing the TC into silicon-carbide or quartz tubing. Excellent uniformity has been achieved with these systems at only modest sacrifices in ramp rate.
The next area of technical variation is based on different heating methods. Eaton Thermal Process Systems uses a small, vertical furnace in which a temperature gradient is maintained. The wafer is elevated to the desired temperature level and is heated rapidly due to its small thermal mass.
Another heating method is used by Mattson Technology and Kokusai. In this approach, a plate or chuck is heated resistively, and the wafer is held at a certain distance. The temperature is measured by a TC embedded in the heat plate, and the wafer temperature is also measured by a pyrometer and calculated by an algorithm based on previous calibrations. This arrangement achieves rather good uniformity, but the practical temperature limit at the wafer is about 900?C or less, excluding some desirable applications.
Yet another source of illumination, which seems to go in and out of fashion, is an arc lamp. Arc lamps are expensive and use a lot of power so that it is practical to use only one lamp in the chamber. Uniformity over larger wafers may be a problem. But Vortek`s arc lamp is making new waves both figuratively and literally. Cooling water is in helical motion inside the glass envelope, and this cooling allows the use of very high-power light sources. These in turn provide extreme flexibility in system design such as a black-body chamber.
Leading companies
We want to distinguish leadership in basic RTP and in total RTP, which includes RTEpi and RTCVD. For basic RTP, only four companies shipped over $25 million in 1997 (Table 3).
 |
AG Associates pioneered RTP technology. Its leadership in basic RTP has only been recently challenged by Applied Materials. STEAG AST Elektronik has grown very rapidly and is establishing a strong presence in both Europe and the US. Dainippon has been the most persistent Japanese participant in this field, and its RTP shipment level towers over any other Japanese company`s.
Applied Materials achieved leadership by its presence in all three areas; it is the leader in basic RTP and in RTCVD with its polysilicon deposition system, and it was a strong second to ASM Epitaxy in 1997 in RTEpi. Table 4 shows the significance of the RTEpi field, in which the trend from batch to single-wafer systems has been very strong over the past years.
 |
Salzer Technology Enterprises Inc. provides strategic advice to high-technology industries, with special orientation toward the semiconductor equipment and process field. 909 Berkeley St., Santa Monica, CA 90403; ph 310/828-9628, fax 310/828-9386, email [email protected].
Dr. John M. Salzer passed away April 26 in Santa Monica, CA. Salzer was a consultant and president of Salzer Technology Enterprises Inc., Santa Monica, CA, and was author of Deposition, Etching and RTP Equipment and Processes, a SEMI newsletter.
Salzer received his BS and MS degrees at Case Western Reserve and his doctorate in EE at MIT. He taught various courses at both of these universities and at UCLA. Salzer was a Fellow of IEEE and a recipient of the IEEE Centennial Medal. He is listed in American Men of Science, Who`s Who in the West and Who`s Who in the World. He co-authored two books as well as numerous scientific and management articles.
His friends and colleagues at SEMI and in the industry will miss him.