by Debra Vogler, Senior Technical Editor, Solid State Technology

Some speakers at last week’s ISSM conference (Oct. 15-17, Santa Clara, CA) speculated that the industry’s eventual transition to 450mm wafers would offer a chance to reformulate the concept of the fab. Not all IC manufacturers would be in the elite group that would benefit from the transition. However, most chipmakers would welcome the opportunity to add capacity in existing facilities while decreasing cycle times.

As a number of presenters at ISSM made clear, there is a movement in the industry that favors the ability to utilize small-lot (<25 wafers) manufacturing as a way to provide fab agility and drive down cycle times. Aquest Systems believes its vehicle-free transport (VFT) technology, FabEX Transporter, released in May of this year, can change the way existing or new fabs implement automation, and enable both small lots and a possible wafer size transition.

The company’s AMHS technology can be deployed in new fabs, but was also designed to work with existing automation systems, so fabs constrained with current layouts and tools can add capacity without making major changes. Available with straight segments of varying lengths and “turn” segments (0-270° of rotation), stocker and overhead transport (OHT) I/O, the system mounts directly under the OHT. According to the company, the controls leverage current ASIC, VoIP, and Ethernet technologies.

The breakthrough with the VFT concept is the decoupling of transportation with the loading operation — a concept that ISSM presenter Mike Brain, Aquest VP of technology and marketing, sees as one of the ways in which “no-wait manufacturing” can be used to drive down cycle time.

Brain told WaferNEWS that he envisions a time when automation will be tied more closely with process and metrology equipment. A key issue is that fabs, especially foundries, want the flexibility to start <25 wafer lots, but those smaller lots require higher movement capacity, and therefore, more FOUPs handled by equipment front-ends. "Double the number of FOUPs at the same wafer out rate translates to twice as much FOUP traffic," he said. This would cause even more traffic jams with OHTs, he noted. "Today, most people believe that an OHT system could handle 13 wafer lots, but it means putting up double the number of vehicles." It is this traffic congestion that Aquest wants to address with VFT.

Additional AMHS components in the FabEX product line are expected to be released in 1Q08. Brain couldn’t provide details at this time, but acknowledged that these components are also designed to reduce the need for wafers to wait for material handling systems, and for equipment to wait for wafers. — D.V.

by Ed Korczynski, Senior Technical Editor, Solid State Technology

Jeannine Sargent, recently appointed CEO of Oerlikon Solar, sat down with WaferNEWS for an exclusive interview to discuss the future of solar energy and her company’s plans to bring photovoltaic (PV) technology to the point of “grid parity” (the point at which the cost to generate electricity is no more than the cost to buy it from the distribution grid).

Oerlikon has committed itself to the future of PV success beyond the current amorphous/micromorph-tandem silicon thin-film on glass technology, and so they consider the eventual use of alternate absorbers as well as alternate substrates.

The company currently sells equipment and complete turnkey manufacturing lines to global customers. Established companies in the solar manufacturing market may only need certain hardware to add to their line, while a new company to the industry may opt for a complete turnkey line from Oerlikon. “Some people getting interested in renewable energy may have tried thin-film some years ago and it didn’t work for some reason,” explained Sargent. “Now they see that a turnkey line offers them the ability to enter the market.” She contends that the PV manufacturing industry is poised to take the last step that hasn’t happened in the semiconductor manufacturing space: buying a complete turnkey line, along with contracts for all the personnel to run it.

In general, one gets a better economy of scale on a per-module basis if you increase the capacity of the line,” Sargent explained. Today’s 20-40MW capacity is relatively small; 60-80 MW is the current scale ordered, and 120 MW lines are now being spec’d.” Even greater capacity would likely come from multiple 120MW lines, perhaps distributed around the world (each producing ~4100 modules/day at 1.3 m²).

Oerlikon considered using both 1.3 and ~4 m² glass substrate sizes (having worked with both for TFT), and after consulting with potential customers decided that the smaller size was preferable for device performance and to minimize installation costs for solar farms/plants. Commercial plants use modules less than a maximum size so as to avoid the costs and hassles of cranes for installation, and to minimize the “sail effect” of stressing or breaking panels in the wind. Consequently, even the 1.3m² substrates need to be cut in two prior to module assembly, so a larger substrate size just results in more cutting. “If you look at solar parks and solar farms, they all cut up large plates,” Sargent noted.

Manufacturing efficiency for Oerlikon’s line would not necessarily be helped by a larger substrate size, since the absorber deposition chamber processes a batch of 30 substrates at a time. Ensuring that film properties remain constant across a larger substrate requires a reduction in the deposition rate and thus the throughput based on area would not be helped. Larger panels might be advantageous for custom installations on the sides of office buildings, but Sargent pointed out that this category is <10% of current forecasts.

Oerlikon plans to drive final module costs down to grid parity primarily through the rigor of industrial engineering benchmarks, based on diverse manufacturing industries such as automotive, aerospace, and paper, and the IC fab industry still provides reference technologies as well. “We have developed some in-line metrology technologies that we’re starting to look at for feedback process control,” said Sargent. “The automotive and aerospace industries really understand reliability, and we want to align the energy market with these industries.”

Pure PV technology development continues for thin-films beyond silicon, today primarily for CdTe and CIGS, though much work is still needed on materials science and engineering to get to a viable manufacturing flow, with peak efficiencies for CIGS seen in the 16%-20% range. Meanwhile, Oerlikon will be building up a silicon-based product line that it expects to be a $1B/year business in <5 years, and if other thin-films or even roll-to-roll substrates eventually seem likely the company will work on them too, Sargent noted.

Sargent thinks that CIGS absorbers could get costs near the $0.25/W-peak level, which she suspects may be the parity level in 2012. “A lot of people will be happy because it’s ‘green,’ but it’ll be driven by economics,” she explained. “We talked to China’s government officials, and they can’t even build up the infrastructure for coal fast enough. They can’t build pipelines from Russia for oil fast enough. They’re fast-forwarding into renewable energy because they don’t have a choice.” Also, the “emerging markets” of the world without substantial existing energy infrastructures should find it relatively much easier to adopt all manner of renewable energy sources including PV. With a terawatt challenge facing the world, it looks like there should be nearly boundless demand for all types of PV modules for the foreseeable future. — E.K.