Adapted from original text by Tom Jeltes, Cursor

Jeroen Bolk, incidentally, never thought about obtaining his doctorate when he joined NanoLab@TU/e as a technician in 2009. After graduating from the higher laboratory school – specializing in chemistry – he started to work at the NatLab and eventually ended up at NXP, the former semiconductor division of Philips in Nijmegen. But since he didn’t see enough opportunities for innovation there, he decided to make the switch to TU/e. ‘I had already learned much at that point about how you make electronic computer chips, but I’d never heard of photonic chips,’ he admits.

Within the Photonic Integration group, research had had just started into large-scale production of photonic chips. ‘Photonic chips were still in their infancy at that time. In 2010, in the framework of an NWO investment project, we received a wafer scanner from ASML, which can be used to accurately make patterns of 100 nanometers. The price of a new machine is ten million euros and it weighs thousands of kilos – we even had to adjust the floor of the lab for it.’

Normally, a wafer scanner works with wafers of silicon with a diameter between twenty and thirty centimeters, while the basic material for photonic chips consists
of wafers of indium phosphide of only three to four inches (7.5 to 10 cm). ‘The system couldn’t process such small wafers, which is why ASML eventually adjusted it in 2012 so that it could handle small wafers as well. That meant we had the only equipment in the world with which you could make such accurate patterns on indium phosphide wafers.’

Bolk, first in his capacity as technician, and later as a PhD researcher, spent the following years addressing the question of how you could actually make photonic chips with this unique machine. The patterns used in photonic circuits, incidentally, are fundamentally different from their electronic siblings, Bolk explains. ‘Angular patterns, for example, don’t pose a problem to electronics, but in order to transmit light you need rounder, smoother structures.’ For the complete photonic chip, ten to twenty of such patterns are applied on top of one another in layers. ‘This means you need a system that checks the thickness of the layers, whether the structures have the correct dimensions, and whether they are applied on top of one another in the correct way. You need an entire infrastructure for that, and I was responsible in first instance for the technical support.’

In the past years, Bolk proved that the wafer scanner can be of added value for a part of the patterns for optical chips, and that it even creates possibilities for developing photonic building blocks that were inconceivable in the past. ‘Among other things, we are now working on photonic chips that can be used as sensors to make the new generation of chip machines even better. I think that’s great.’ In May 2020 he obtained his PhD for his findings.