Current Issue An Independent Journal Dedicated to the Advancement of Chip - Scale Electronics July 2001 Email the editor

By Ron Iscoff, Editor

Control of the photolithography tool market for IC packaging applications is rapidly developing into a two-technology race, split between wafer steppers and proximity aligners (Figures 1-3).

Figure 1. The EV Group's IQ Aligner accommodates wafers from 100 mm to 300 mm.

Figure 2. The Suss MA300 proximity aligner offers a 300 mm range of focus.

Figure 3. Ultratech's 1:1 Saturn Spectrum 300 stepper was introduced last December.

To date, the aligner market for packaging applications is controlled by two European firms with extensive U.S. operations, while the stepper market for packaging is dominated by a Silicon Valley company.

If you're wondering about the traditional lithography tool giants, Canon and Nikon, it appears that the flip-chip/ wafer-level packaging market holds little interest for them.

As one bump foundry executive mused aloud to Chip Scale Review, "Would you rather sell 10 machines at $1 million a piece (about the cost of a low-end stepper for packaging) or two machines at $5 million each (the price of a high-end Canon or Nikon stepper for a wafer fab)?" The answer seems obvious to the Japanese stepper giants, too.

The Competition

If the market competition between proximity aligner and wafer stepper were a horserace, the odds would go something like 50/50 as to which technology will achieve the largest market share, ultimately-at least where 200 mm wafers and below are concerned.

Although 300 mm wafers are far from a production item, the latest generation steppers and aligners are built to handle them.

Selection of aligner or stepper is going to boil down to a few key items (in no particular order):

• Wafer size and handling capability

• Process yield considerations (when isn't this a key item?)

• Film thickness

• The critical dimensions required by the flip-chip or wafer-level process

• The money available for capital equipment

• Ease of operation, ease of maintenance

• Cost of ownership

Key challenges in producing equipment for wafer-level packaging, according to an aligner maker, involve the patterning of a thick photoresist and the need to plate thick metal layers, which may not follow a step-and-repeat pattern. (Figure 4 shows a plate patterned for bumping with a Suss aligner.)

However, depending on your wafer-level process, you might not need the precision alignment that a stepper or a high-end proximity aligner will deliver.

New Foundry

Victor Batinovich, founder of ASAT, IPAC and Swire packaging foundries, is setting up a new WLP foundry, Advanced Inter-connect Solutions (AIS) in Silicon Valley.

Employing its proprietary technology, AIS will provide RDL, UBM, fine-pitch bumping, multi-solder screen printing and related services, as well as what the company terms the WL-TCSP (True Chip-Size Package) shown in Figure 4.

Batinovich says, "For the type of work we do, it's not a question of expensive stepper vs. expensive proximity aligner. The important factors are cost of ownership, flexibility, wafer sizes, photo-image material thickness and productivity."

"At 10-15 micron lines who cares? We can achieve the desired results with last-generation aligners. At some point in the company's growth," he admits, "We'll use both steppers and aligners."

Ron Blankenhorn, president of Pac Tech USA, is also about to establish a bumping facility in Silicon Valley, using Pac Tech's electroless nickel process. "Currently," Blankenhorn reports, "we're looking at both steppers and contact aligners."

Figure 4. Cross-section of the AIS True CSP

Different Requirements

"The requirements in photolithography are completely different for wafer bumping compared to front-end technology," observes Prof. Michael Töpper of Fraunhofer-IZM in Berlin, Germany. Fraunhofer and Dr. Tšouml;pper are pioneers in wafer-level packaging.

Prof. Töpper notes that in packaging, "We are talking about structure sizes well above 10 µm for solder bumping (Figure 5). Only in gold bumping is there a need for smaller gaps, but these are still above 1 µm, such as in LCD drivers."

Figure 5. This SEM, courtesy of Fraunhofer IZM, shows an exposure on dry film resist with a Suss mask aligner.

Proponents of steppers insist, however, that aligners can't support the requirements for gold bumping.

The increasing demands of gold bumping, including resolving finer features and printing high-aspect ratio images in thick resist with improved sidewall profiles, will enable steppers to dominate this lithography area.

Among the key users today is the Flip Chip Division (formerly Flip Chip Technologies) of Kulicke & Soffa Industries, Phoenix, Ariz., which currently owns steppers and aligners.

The division processes 100-200 mm wafers, and is getting ready to offer 300 mm, according to David Stepniak, vice president of engineering. (Figure 6 shows a SEM of what the division claims is the world's first commercially bumped 300 mm wafer.)

According to the K&S division, the technology for 300 mm is scaleable and is available for license.

Multiple Sizes, Multiple Masks

Stepniak says the requirements for an acceptable photolithography tool include the ability to handle multiple sizes, with multiple masks and multiple passivations. Since 300 mm wafers are waiting in the wings, Stepniak says he also looks for "scaleability."

While most bumping processes tend to be somewhat similar, the chief commonality is the need to use thickfilms which respond well to "broadband-type" tools, adds Stepniak.

Although wafer steppers are more expensive than proximity aligners, Stepniak says their steppers offer other advantages. These include, he adds, the lower cost of small masks over plates.

However, according to Fraunhofer's Tšouml;pper, the reduced cost of stepper reticles may only be a valid argument in prototyping.

Why Three Different Tools?

But, to return to user K&S, why is Stepniak using different tools? "When we were first buying tools, stepper suppliers were not prepared to support the bump market with a broadband tool. Now that's been corrected and it opens up the market."

Aligners, on the other hand, can match the throughput and yield of a 1:1 stepper, according to the aligner fraternity boosters. (This point is untrue, according to stepper boosters.)

Aligners are also less expensive and easier to operate. At least until some time into the future, when narrow alignment tolerances are an important factor with 300 mm, the future of the aligner seems to be secure.

Unitive Electronics Inc., Triangle Research Park, N.C., is one of the leading U.S.-based wafer-bumping services providers. The company is now carefully considering the costs and equipment involved in moving to 300 mm wafers.

Figure 6. SEM illustrates what is claimed to be the first commercially bumped 300 mm wafer. (Ultratech Stepper).

"We have proposals out to customers who have asked us about processing 300 mm wafers," reports Dan Mis, a Unitive vice president.

However, before jumping on the 300 mm wafer bandwagon, Mis says, Unitive has to answer a key question: "How many foundries are going to use outside bumping for 300 mm?

"Spending the money you have to spend for a 300 mm fab this year, and having that fab remain static for a year until somebody starts to ramp, is very expensive," Mis observes.

$10 Million-$50 Million

Unitive, he believes, could put together a 300 mm fab now, "However, we believe the way to proceed is with a strong customer commitment or partnership. At this point, we're not ready to bankroll the whole project on our own."

While Mis won't disclose the potential cost of a 300 mm fab at Unitive, he says numbers between $10 million and $50 million, are being tossed around the industry.

Conclusion

A market that barely existed five years ago is being eyed hungrily by three competitors-two on the mask alignment side and one who is currently the only viable maker of 1:1 wafer steppers, a stepper technology that seems well-fitted to packaging tasks.

Will steppers end up owning the WLP lithography market? Or will continued improvements in contact aligners (as well as their typically lower cost) push steppers into second place as the WLP technology of choice? The answer, despite what the proponents of one camp or another would have you believe, is in the future.

The 300 mm wafer, is now one of the many market wildcards-a wildcard each technology is eyeing with some degree of reverence. Will the stepper dominate this market. Or will it be the contact aligner that pulls ahead. As always, the market will determine the ultimate success of these contrasting technologies.