Secondhand tools have capacity to aid Asian, small chipmakers

PHOTOS BY JEFF MALONEY (LEFT), PHOTOLINK (RIGHT), COURTESY OF GETTYIMAGES; COMPOSITE BY JAMES SCHLESINGER

If you're the manager of a second-tier fab looking to acquire an advanced deposition system at a decent price, you're in luck. If you're the manager of an advanced fab trying to sell off that tool, get in line. The credit—or the blame—for this state of affairs goes to the capacity glut, says the head of an on-line clearinghouse for used semiconductor equipment.

"What we're seeing is so much excess capacity and people looking to liquidate assets that [there are] brand-new tools and...a significant amount of one- or two-year-old inventory coming into the marketplace," says Mitch Lashman, president and CEO of KeyAssets.

"Cash is king, and it's a buyer's market," he adds. "If you know what you're looking for, you will have options in today's marketplace." Those options often include tools with more manufacturing oomph than what's required for the process.

A perfect example of this occurred when Microchip, the Arizona-based microcontroller manufacturer, purchased Fujitsu's fab in Oregon earlier this year, Lashman notes. The technology in Fujitsu's Gresham facility far exceeds Microchip's half-micron process requirements. The installed equipment can handle 0.18-µm processes, according to the executive.

"In today's market, Microchip and other companies that have money in the bank can procure [advanced tools] and secure their manufacturing futures," Lashman asserts. In some cases, process gear can be had for as low as 15 to 25 cents on the dollar.

When the industry was riding high a few years ago that was not the case, says Lashman, who has more than 20 years of experience with semiconductor equipment. Most secondhand gear enters the secondary tool market approximately three to five years after initial installation. "Three to five years is the typical cycle; seven is the tail end," he says.

The type of equipment available depends on its function in the fab, Lashman points out. "Metrology or lithography tools have a shorter life cycle than others. Implant, deposition, or etch tools might have a longer life cycle." Metrology and lithography equipment usually ends up for sale sooner because of the rapid rollout of next-generation ICs—one or two generations down the line and out come the For Sale signs.

A mid-September glance at Key-Assets' Web site showed an abundance of metrology and inspection gear in the front-end category. It led the list of 16 equipment types with 159 pieces, followed by lithography and photoresist gear at 103 pieces.

Brian Shekerjian, manager of manufacturing technology for ON Semiconductor, says he's not sure that any type of equipment is more readily available than another. "I have no impression one way or the other," says Shekerjian, whose responsibilities include managing front-end equipment strategy for ON. When the market was hot, some types of gear, such as GSD implanters from Axcelis and dry metal etchers from Lam Research, seemed hard to come by. At this point, though, "nothing is difficult to find."

The Arizona-based manufacturer of power management microchips has an ongoing surplus-equipment program. "We work with our factories and identify what is surplus. We manage that inventory and track it. We store the inventory at a couple of different locations, then work with our broker primarily to identify strategies for maximizing that inventory."

Shekerjian says ON redevelops or refurbishes the equipment, or uses it "to barter for other equipment we need, or we sell it and generate cash for the company." Over the last several years the program has generated "tens of millions of dollars" for the chip manufacturer, according to Shekerjian.

The manager says that the downturn hasn't necessarily brought in more money for secondhand gear sales, primarily because the dollar value of the equipment decreased. Revenues were lower in 2001 than they were the previous year and the year before that, Shekerjian says. Two factors contributed to the higher returns in previous years, he notes. "The market value of the equipment was not as depressed. We were also busier."

Shekerjian believes a healthy market is key to a healthier return than a depressed market. "By trading or cashing in on equipment that you have as the market gets healthy, you stand to gain more dollar value that way." He adds that ON's program "probably helps the company in good times in terms of bottom-line dollars [more] than in bad times."

Despite the perceived advantages of unifying what Lashman calls "a fragmented marketplace," the semiconductor industry has been slow to adopt the Internet as a vehicle for buying and selling. He attributes the slow rate of adoption to "people's fundamental resistance to change." The attitude, Lashman declares, is "what I have is working, and even if it's not working better, it's not failing."

Unsurprisingly, an additional reason for the foot-dragging is economic, the KeyAssets head asserts. "Change also requires investment. It's an investment ofttimes of money, always effort, and resources, which also obviously comes back to money, because you have to assign people to implement change."

Finding the additional resources "when you're in survival mode" leads to what Lashman calls "pushback" from management, whose response often is, 'we want to eat today, don't talk to me.' There's no revenue stream, and no definitive hard-case facts."

Nevertheless, Lashman argues that using services such as KeyAssets "is not rocket science. What we tell people is, at the end of the day, we have two customers. They are buyers and sellers. The value to the seller is to eliminate the current fragmented marketplace, which has a middleman who controls information and takes an inordinate amount of value from the sellers.

"For buyers, the middlemen [who] control the information flow...ultimately put a substantial premium on the value of that information," he continues. "Buyers are at their mercy in the establishment of fair market value." Buyers, he insists, know the cost of a new piece of process equipment and can therefore decide what their "breaking point" is in terms of how much to pay for a secondhand model.

Certainly, procurement specialists are aware of the distinct advantages of buying tools spanking new, Lashman says. "There's warranty and insurance that comes with a new tool. You can argue there's minimal risk because with an OEM, in theory, it's a clean sale. When you buy used, you assume some level or some degree of risk."

KeyAssets tries to minimize the risk by assuring buyers that they have the exclusive opportunity to inspect a tool before buying and renegotiating with the seller without endangering their pocketbooks or blowing the purchase because of paperwork or bureaucratic procedures.

Buyers often will turn to the used market in order to add capacity to a well-adjusted process and need additional similar tool sets. "This is not a scenario in which you can buy the latest and greatest...." Often, he says, contracts between device manufacturers and their customers insist that any process modification "requires a requalification" of the line. "So the path of least resistance for many companies adding capacity is to bring in the identical tool."

By the same token, "trailing-edge manufacturers" such as ON, Linear Technology, and some foundries realize that tools Intel, Motorola, and TI are divesting themselves of "still have a good five-year life cycle."

At least one industry expert believes that intense economic pressure is forcing a change in the nature of the used-equipment market. Bruce Huling, an Arizona-based consultant specializing in intellectual property and process equipment issues, asserts the market is "becoming saturated at the smaller wafer sizes and that...equipment is going to universities—'toy shops.'

"It isn't quite like it was in the near past, where there were factories going under but there were factories still trying to gain capacity and the used market was a viable way to add capacity. Now, people are adding capacity because of the economics. They want to add capacity at the upper end, not the middle. There's pressure on price, and it's my theory that...the market is undergoing a change in its customer base."

That base has now shifted offshore, Huling believes. China, notably, is one of the primary destinations. "I wouldn't believe, for example, that in the U.S. their sales of 5- and 6-in. equipment are increasing. I have a couple of thoughts on China. It's such a vacuum...and technology and business is in such a state that 4- and 6-in. equipment makes sense there, as would 8 in."

In China "it's not the size of the wafer; it's whether or not you have product to sell," he adds. Efficiencies are such in China that overhead makes manufacturing very cost-effective. "A lot of the talk I hear is that the majority of products made in China are going to be sold in China. The idea is to put a factory in China, and after tariffs and all that stuff you'd be competitive with putting a factory in the U.S. It's simply not the same business model."

Huling, who is a member of the editorial advisory board of MICRO, thinks the "third-world market" will be able to absorb smaller-scale equipment for some time, "whereas the U.S. and most of the first-tier players such as Germany, probably Italy, and Japan are not. It's economics. If you're in Japan or the U.S., you know what your overhead is to run a factory; the business model has already shown that your viability rests in larger wafer scaling. Why would you suddenly change the model for that structure? It's the age-old story: factories go where overhead is low and governmental restrictions, for example, are less."

TSMC recently made a big splash with the news that the foundry had applied to the Taiwanese government to build an 8-in. fab in Sonjiang Science Park in Shanghai. In September, ON Semiconductor's Shekerjian had a trip scheduled to Shanghai, where the chipmaker has plans. Even though the U.S, Japan, and Europe are under treaty provisions to restrict sales of certain process tools to China, the country's entry into the World Trade Organization may finally result in loosening the restrictions.

Huling notes that blacklisted equipment such as photolithography systems are not the only need China will have in order to grow its own viable semiconductor business. "The things that keep you in the race are both photolithography and planarization."

The buying and selling of used gear "has probably been fairly steady," he insists. "Who's doing the buying is shifting to emerging markets. That's my theory."

Asked whether he considers Key-Assets' business a leading indicator of the industry's status, Mitch Lashman replies: "Yes, we would. Typically, during an up-tick in the market, the first requisitions that you see are for used equipment, not new. There's one caveat. The exception to that is obviously R&D, where you need to invest in new technology. But for first capacity, the release of funds will typically go into the used-equipment market."

Lashman says the market right now is "very flat, just skimming along the bottom at this point." The level of inquiries has increased, an activity he calls a precursor. "We've seen much more quoting over the last month. It began in the middle of August, and we're starting to see...shopping and quoting and things of that nature."

Potential buyers are "very tight with the purse strings at this point. Most people are preparing themselves for the 'what-if' scenario. The easiest areas to justify expenditures are areas where you can justify a yield improvement, even in a down market and even in scenarios where companies have excess capacity available. There will always be the opportunity to justify expenditures for yield improvements." That opportunity will manifest itself in the sale of metrology gear "and other peripheral support items that help yield," such as particle detection and leak detection equipment.