New Roles for Merchant Mask Makers
Some of the more obscure but interesting news in the patterning business in the last few months was that two of the biggest mask making companies, Hoya (Japan) and Dai Nippon Printing (DNP, Japan) are acquiring Molecular Imprints (MII, Austin, TX) process equipment for making low cost “working molds” or copies for imprint. If you have enough grey hairs you would be forgiven for thinking that making low cost copies sounds like a time warp from the ‘70s — how could this possibly be news?
It really is a return to one of the roles of mask making for contact printing, where 1000s of low cost working plates were made from an original. The significance is that the tool being acquired by Hoya is for copying molds of the type that are being used making patterned media disks, and the tool by DNP is for semiconductor device mold copying. This is the “last” infrastructure barrier to commercial imprint. Finding a commercial source of “working molds” would be the last task on the list after you had fabricated original master molds and fabricated devices on equipment that had the potential to meet commercial production targets. Furthermore, the mask makers would not invest in working copy tools without assurance that business was coming. I think that this is another signal that we will hear about production implementation of imprint very shortly.

Mold replication — The left image shows an imprint with the original master; the right image shows an imprint with a copy of the original, confirming no loss of resolution in the copying process. Work from Molecular Imprints, published by D. Resnick et al., “Step and flash imprint lithography template fabrication for emerging market applications,” Photomask Japan, April 2007.
The development of merchant sources for the production elements in patterned media follows a well established path that has been successful in semiconductors. The biggest semiconductor fabricators will do early research but will drive the solution out to their vendor base, so to get all the advantages of outsourcing exactly what you want. There is an interesting contrast with commercial roll-to-roll imprint or “embossing.” This has been around for decades in the manufacture of security holograms and optical films. Several of the major roll-to-roll imprinters have captive origination, working copy, imprint materials and process equipment fabrication. A few companies are appearing to provide merchant sourcing but they have been handicapped by the absence of standards for molds and film sizes. They are more like custom integrators rather that merchant vendors. To me this emphasizes the importance of standards, road maps, and the presence of a strong controlling force, such as Intel, in enabling an effective merchant network that can respond to new technology demands.
Any assessment of the direction of a new opportunity requires an evaluation of both the technology and business. The technology and business of patterning will be the focus of this blog. I can make a good case that patterning is the hub that supports all device manufacturing. It is the most commonly used unit operation, so it determines both device size and device cost. This blog will discuss patterning at any feature size on any substrate.
The common themes in the patterning business are mastering, resolution, how accurately layers can be aligned to each other (overlay) and defect density. For example, patterned media require very small features at 20 nm, but the media are remarkably flat, are single-layer devices that do not require overlay, and have built-in defect correction. Imprint is an optimal patterning solution for patterned media. To implement requires captive master origination because the write times are remarkably long at 1-2 months, and all the major players have established internal programs. Merchant copying is being commercialized (as discussed above), as have the materials and equipment sourcing from companies such as MII and Obducat (Sweden). As a result, Hitachi Global Storage (HTGS, San Jose, CA) and others are poised to convert to patterned media in production.
the fine print... commented:
"To implement requires captive master origination because the write times are remarkably long at 1-2 months..." whoa, I don't think my e-gun will last that long!
Tom Ortman commented:
Hi Mike: Great to see your blog. This space should provide plenty of good fodder. Good luck with it.
Tom Ortman commented:
Hi Mike: Great to see your blog. This space should provide plenty of good fodder. Good luck with it.
Lucky Somers commented:
Mike,
For an old resist guy, you seem to make a good case! Congrats on the new blog. Look forward more pearls of wisdom in the coming months.


















