Paul van Gerven is an editor at Bits&Chips.

23 January 2017

‘It’s like you actually believe him,’ wrote ASML’s PR representative when she emailed me back the article I’d sent her so she could correct any factual errors. It was early 2008; I’d interviewed Phil Ware, a senior fellow for lithography at Canon and a vocal opponent of EUV. Hardly a neutral party, of course, but one of the few critics within my reach who was willing to speak on the record. I presented his arguments to ASML and I wrote the article ‘No guarantee that EUV will succeed’. I thought the piece was balanced to the point of boredom, but the PR rep clearly thought otherwise.

So who should I believe? The question had me in its grip for quite a long time, and I suspect the fate of EUV lithography occupied a larger-than-average share of your attention in recent years, too, even if you don’t work in semiconductors. We’re talking about the heart of the high-tech industry, after all.

I wanted to believe that EUV would succeed. ASML is a successful company that has blown away its competition through a combination of technological innovation and prudent investment policy. But the fact the EUV development process was largely a black box didn’t make it easy to keep the faith. Now and then we’d get an update on source power or some other performance criterion, but the how and the why usually remained unclear. What’s more, the deadlines on the roadmap kept getting pushed further and further into the future.

Meanwhile, the foreign press was generating an unending stream of sceptical reports. As a journalist, you know they should be taken with a grain of salt. The media are often used to bolster agendas, and hyperbole is the order of the day. And the seductive pull of sensation is always lurking, of course. But when TSMC’s (since retired) litho director Burn Lin paints a stygian picture of the whole EUV undertaking, it makes an impression. And I’ve spoken with many people who have a far better understanding of physics and machinery manufacturing than I, who also thought success would be hard to come by.

And then the question starts to eat at you. What if it doesn’t work out after all? I’ve been worried more than once that I might go down in history as ASML’s messenger boy, who, like many a cycling correspondent, stood too close to his heroes to see the needles still hanging from their glutei maximi.

That torment is now over: I’m a believer. Even gentleman scientist Chris Mack, though without a doubt far better steeped in lithography than I, can’t change my mind now.

My faith grew gradually, thanks largely to the frank and open way in which CTO Martin van den Brink spoke about EUV – and speaks of it again now. He’s not one to dabble in mealy-mouthed sound bites – to the occasional despair of his colleagues in PR. The fact that the ranks below him now also dare to admit the mistakes they’ve made, feels like a switch has been thrown. You only dare to talk about what’s gone wrong once you’re convinced that success is inevitable.

But mistakes is actually too harsh a word. The development of production-ready EUV scanners is the Apollo programme of the twenty-first century, and it deserves deep respect. If ASML can be accused of anything, it’s that at first it badly underestimated the endeavour and then clung ferociously to its over-optimistic plans for too long. The source, the vacuum, the optics: these were completely new technologies that ASML and its suppliers first had to master. In the haze of victory after the Twinscan success, that fact went unacknowledged. The repeated delays that followed blew more wind in the sceptics’ sails than was called for.

But no matter what, there would have been resistance. A great many Americans didn’t believe in the Apollo project, either, until Neil Armstrong set foot on the moon. Now it’s a symbol of courage and ingenuity. The same will be true of the EUV project.