Mass Storage Pins Progress Hopes on Aggressive Litho
Storage requirements in HDDs and SSDs are increasing at a rate that will require fast progress in lithography, etch and metrology, speakers at the recent DiskCon conference said.
Alexander E. Braun, Senior Editor -- Semiconductor International, 9/26/2008 12:45:00 PM
 |
This year’s Diskcon, held in Santa Clara, Calif., offered a special session, “Future Storage Lithography: A Crossroads in Technology?” Session chair Paul Hofemann, vice president of HDD and emerging markets business development at Molecular Imprints (Austin, Texas), said there is no doubt that patterned media will happen in the HDD industry. “Now it’s no longer a question of whether, but when. It won’t be as soon three months, but not as long as three years, either.”
HDD areal density must keep up. “Most present technology extensions are running out of steam, and going to >1 Tb/in.2 won’t be easy,” Hofemann said. “We’re at about 700-800 Gb/in.2, and these extensions are hitting limits. Discrete track media (DTM) will get us past 1.5 Tb/in.2, and then it’ll be necessary to go to bit patterning. Heat-assisted recording theoretically could work at a 10-30 Tb/in.2 level, taking us to the end of the decade.”
 |
| The HDD Areal Density Roadmap shows current technologies are running down. At present, storage density is at ~700-800 Gb/in.2, and achieving >1Tb/in.2 will be difficult. (Source: Molecular Imprints) |
The current storage transition is different from previous ones, because it requires lithographic resolution that goes beyond the most aggressive semiconductor roadmap. Patterned media also introduces processes — such as litho and etch — not currently found in disk media production fabs. “This will force unprecedented collaboration and coordination among suppliers and customers,” Hofemann said. Imprint technology is expected to deliver <20 nm resolution, limited only by template resolution. Productivity requirements will require imprint lithography modules to be clustered in order to scale to >1000 disks per hour.
 |
| Dimensions for 1 Tb/in.2 require 12.5 nm CDs, and must be printed at low cost in high-volume manufacturing in only a few years. (Source: Molecular Imprints) |
Thin-film head challenges
Peter ten Berge, product marketing manager for ASML (Veldhoven, Netherlands), spoke about the lithography challenges in thin-film head (TFH) manufacturing. Improved overlay is one challenge. Additional measures must be taken for TFH manufacturing on AlTiC, to control thermal fluctuations and drift effects. “Litho tools are available and ready today to support the areal density roadmap, both the KrF and ArF imaging and overlay requirements,” he said. Meeting the litho roadmap for TFH manufacturers also depends on developments in resist processing, reticle quality and wafer flatness, he said.
Robert Allen, manager for lithography and wafer purification at the IBM Almaden Research Center in San Jose, spoke on “Functional Materials for Advanced Patterning.” He showed that chemical amplification is the great enabler for modern lithography because it makes possible high-speed, high-resolving power resists. “We’re trying to extend immersion litho; not too long ago, there was talk of things like ultrahigh NA, immersion litho using oil instead of water, etc. — all this is dead,” he said. “Most recognize double patterning as the way of extending litho to finer features without buying new tools that don’t exist.” Allen stated that for a variety of reasons his group is working on materials for extreme ultraviolet (EUV) as well as e-beam lithography. “One of these involves imprint litho for template processes; nanoimprint litho interests us,” he added.
Allen expounded on beyond-193 nm expectations, noting that the leading contenders are EUV, e-beam and imprint. “E-beam in terms of direct-write is very interesting, but there’s the matter of throughput,” he said. “There’s considerable effort directed at maskless lithography, to build tooling with multiple, parallel e-beams for good throughput.” He said that EUV has similar challenges to e-beam direct-write: complex tooling, vacuum-based litho, extremely high-speed resists. “Imprint still has tooling challenges: not too many offer tools, throughput is a bit sketchy, defectivity must be controlled, and there are many template challenges — we’re in the learning stage.”
Allen also reviewed polymers. “What’s new is the use of these thin films as organizational structures directed by a preliminary lithography step,” he said. “The pre-patterns come in various types. We’re working on chemical patterns, where you have a difference in surface chemistries. Then you apply the polymer and, by selecting the right system, you can get different morphologies, where you have a chemical bond connecting polymer A to B, forming a nanoscale morphology of 10-20 nm. This can enable complex morphologies and stacks of materials. In this process, orientation is all-important and is driven by the surface chemistry.”
 |
| Directed polymer self-assembly is directed by lithographically defined pre-patterns. Current challenges lie in developing materials for litho-friendly directed self-assembly. (Source: IBM) |
“Patterned Media Technology — An Equipment Perspective,” was a presentation by Ren Xu, key account technologist at Intevac (Santa Clara, Calif.). “Key equipment trend drivers are process development for advanced continuous media, and the investigation of patterning approaches, such as etching strategies and planarization,” Xu said, adding that the winning process approach for storage must be manufacturable, cost-effective and scalable. “Equipment platforms must be modular and offer throughput and low cost. Although the processes aren’t down yet, the solutions should be all-encompassing, universal enablers for industry-wide process development and technology evolution.”
Plasma etch challenges
Deirdre Olynick, staff scientist at Lawrence Berkeley National Laboratory, in a presentation titled, “Nanoscale Plasma Etching — Challenges for Patterned Media,” concluded that etching research is critical for <20 nm patterns. Resists and self-assembled layers are thin and not ideal for pattern transfer. Olynick said there is potential for <10 nm etching to run into fundamental problems, particularly because profile control is difficult and process windows are shrinking. “An unprecedented understanding of the etching process will be required, and additional knobs to turn must be put into place,” she said. “We may require alternate sources, such as atomic-layer etching. Only by breaking up processes further can we expect to better control them.”
Sri Venkataram, director of field marketing and customer support at KLA-Tencor (Milpitas, Calif.), took a snapshot of the International Technology Roadmap for Semiconductors (ITRS) over the next few years, in particular the smaller features that drive it from the flash memory perspective. “The annual shrink rate will come to ~30% over the next five years,” he said. “There are minor differences between semiconductor metrology and data processing, but we’re breaching those and taking advantage of the similarities. The three big challenges are overlay alignment, CD and defect control.”
According to Venkataram, overlay no longer lags behind ITRS requirements, and media imprint lithography is required for higher areal density, and to remain competitive with flash. There are some technology extensions on the inspection side for scatterometry, ellipsometry, reflectometry and topographical analysis, to non-destructively inspect wafer surfaces for defects and film thickness uniformity, and potentially, even SEM applications for media imprint process control.
