Every technologist knows about Murphy’s First Law: “Anything that can go wrong will go wrong.” Most know about the Second Law: “If it could go wrong, but hasn’t yet, it will at a more inconvenient time.” But few appreciate how these laws control the dynamics of R&D.

Before a project starts, its leaders believe they have foreseen all the difficulties. Optimism always reigns at the beginning, partly because the realm of things that could go wrong then really is limited. (How bad can a PowerPoint presentation be?)

As a project takes shape, more choices, commitments, and discoveries are made, and the potential for screwing up increases. Previously unknown (or underappreciated) effects and interactions appear, generally as problems. As these early anomalies are handled (or sidestepped) entirely new things go wrong. And as these new challenges are overcome, even more undesirable phenomena that couldn’t have arisen earlier, now do arise. For a while, new problems seem to appear faster than old ones can be resolved.

That pile-up of new problems uncovered by prior fixes constitutes Murphy’s Wall, the fundamental barrier to doing anything really new. The closer one gets to the goal, the faster more and more issues appear. Young technologists find themselves working 120 hour weeks, cynical managers begin looking for lateral transfers, pessimism increases, milestones are redefined, budgets, reapportioned. A motivated team will power though this stage, with or without additional support, but it isn’t easy.

And then, if the project goes on, the last unexpected thing that can go wrong, does go wrong, and is either handled successfully - or not.

If that last unexpected problem is overcome, the project has climbed Murphy’s Wall. It is hard to believe at first, but success appears dimly ahead and the atmosphere changes. People begin to look forward to whatever is coming.

Sometimes, though, that last problem is fatal, either because it cannot be overcome at all or because dealing with it will tie up more time and resources than seem justified. What happens then depends on the quality of the leadership. Projects can and do go on consuming resources after a fatal problem appears. (One grizzled colleague called such things “fatal features.”) Denial is always easiest. Sometimes R&D can be salvaged and redeployed to a new goal. Sometimes accepting one new idea can dislodge an impenetrable limit. And sometimes, it all falls apart.

Innovations in semiconductor lithography have two additional challenges to overcome. One is Moore’s Law: whatever the goal is, in two years it will have gotten harder because the industry will have moved on to smaller devices. That means that R&D projects have to climb Murphy’s Wall quickly, overcoming the unexpected problems in time to catch the learning curve for the next node. Otherwise, that wall will just keep getting higher as the goal recedes. Twenty years ago, there was a massive R&D program on 1× X-ray lithography that failed repeatedly to catch up to the CD in current production, in spite of consuming all available R&D resources.

The other challenge is the time and money required to scale up anything to the massive dimensions of the semiconductor industry. Just building the factories needed to supply materials for tools can take a device generation. In spite of all the effort, there never was enough high-quality CaF₂ capacity to build the proposed 157nm exposure tools, and that experience discouraged potential LuAG customers, dooming high-index immersion technology, too.

On-line at www.microlithographyworld.com and in this issue we have articles describing the progress made by EUV lithographers overcoming their expected and unexpected difficulties. They are closing in on the technical specifications defined by the industry, but the industry has moved on and the target for mass production with EUVL is now around 22nm. Will they be ready in time, or will the 1× X-ray debacle be repeated?

If you have comments on this editorial ? or anything else ? do not hesitate to e-mail me at marcl@pennwell.com or to write me at 201 San Antonio Circle, Suite 205, Mountain View, CA 94040.

M. David Levenson
Editor in Chief