Van den Brink & Van Hout

A chip is made by shining light through a mask, a stencil of the circuit, and shrinking the image onto a silicon wafer coated with light-sensitive resist. The same wafer goes back under the light dozens of times, one layer after another, each lined up with the last to within a few nanometres. Smaller lines mean more transistors, so the whole industry is a fight against one limit:

$CD=k_1\frac{\lambda }{NA}$

• CD: the smallest line the machine can print (critical dimension)
• λ: the wavelength of the light: shorter light, finer lines
• NA: numerical aperture: how wide a cone of light the lens gathers; water under the lens pushes it past 1
• k₁: everything else, from tricks of illumination and masks; physics stops it at 0.25

ASML climbed this ladder twice over: first by raising NA (bigger lenses, then water), then by cutting λ from 193 nm to 13.5 nm with extreme ultraviolet.

He studied electrical engineering at HTS Arnhem and physics at the University of Twente, and walked into ASML in 1984, the year it began. Its first machine leaked oil. One of his first jobs was alignment: how the machine finds where the last layer sits so the next lands on top. His method, based on phase gratings, became the basis of how ASML machines line up layers.

He led the PAS 5500, ASML's first commercial hit, then drove the three bets that made the company: TWINSCAN's two wafer stages, immersion lithography and EUV. He ran ASML's technology for some thirty years, became President and CTO in 2013 and retired in April 2024. ASML lists 38 US patents in his name.

Before a wafer is exposed it has to be mapped: its height across the surface and the exact position of its alignment marks. On a single stage the lens waits while that happens. TWINSCAN gives the machine two wafer tables: one is mapped while the other is under the lens, then they swap. The lens almost never waits.

$\text{wafers per hour}=\frac{3600}{\max (t_m,t_e)+t_s}$

• tₘ: seconds to measure one wafer
• tₑ: seconds to expose it
• tₛ: seconds to swap the tables
• max: only the slower of the two jobs counts, because they run at once

Van Hout also joined in 1984 and served ASML on and off from then. He joined the Board of Management in 2009 as Chief Marketing Officer.

In 2013 he became Chief Program Officer for EUV. The physics had been shown in labs; the job was to make it a product chipmakers would bet their factories on, with enough light, uptime and suppliers to run around the clock. Bits&Chips summed up his term as bringing EUV “to a level where EUV was widely accepted”. From 2018 he was Chief Strategy Officer, in charge of ASML's strategic suppliers, then retired on 29 April 2021.

No lamp or laser shines at 13.5 nm with enough power, so ASML's machines make the light from plasma. About 50 000 tiny tin droplets, each around 25 micrometres across, fly through a vacuum every second. A carbon-dioxide laser hits each one twice: a weak pre-pulse squashes it into a thin pancake, then the main pulse heats it to plasma that shines in extreme ultraviolet. Only about 5 % of the laser energy comes out as useful light.

EUV is swallowed by air, glass and lenses, so the whole path is in vacuum and made of mirrors. Each mirror is a stack of about 50 molybdenum–silicon layer pairs. Every boundary reflects a little, and when the pair thickness is half the wavelength all the small reflections add up in step:

$2d\cos \theta =\lambda$

• d: thickness of one molybdenum + silicon pair, about 7 nm
• θ: the angle the light makes with the mirror's perpendicular
• λ: 13.5 nm, the wavelength

Even then each mirror keeps only about 70 %. After a collector, the illuminator, the mask (itself a mirror) and the projection optics, a few per cent of the light is left. That is why the tin source has to be so bright.

$\text{light left}=R^n$

• R: how much one mirror reflects, about 0.7
• n: how many mirrors the light bounces off

Van den Brink: honorary doctorate, University of Amsterdam (2012); Knight of the Order of the Netherlands Lion and IEEE Cledo Brunetti Award (2014); IEEE Robert N. Noyce Medal (2015); imec Lifetime of Innovation Award (2019). ASML counts him among its Fellows.