How a strange insect put Philips on the litho path

Compared to the US, Europe was another planet in the decades after World War II, in both economic and technological terms. Countries like England and France were putting money into defense technology, but not the billions the Americans were investing. Large industrial companies like Philips and Siemens primarily served consumers and other companies. There was barely room for tiny semiconductor startups; they lacked both the capital and the customers.

In America, it was mostly the Department of Defense that was stimulating research into the miniaturization of electronics. US companies had emerged from the war at full speed and were continuing to profit from a generous government budget for the development of military electronic applications. Modules filled with vacuum tubes and later transistors formed the new ‘brains’ in weapons systems. The big difference with Europe: inventive small companies were also raking in defense contracts.

Famous symposium

In the 1950s and 1960s, researchers from AT&T’s Bell Labs and the Philips Natuurkundig Laboratorium (Natlab) formed close ties. Hajo Meyer, who joined the Natlab in 1950 and rose to director, said in an interview that the first time he visited Bell Labs was with his boss, Piet Haaijman. “We apparently left a good impression there because after our trip, a collaboration was established and we had regular visitors from Bell.”

Haaijman and a team of Natlab researchers made the crucial trip in April 1952. In Murray Hill, New Jersey, the Dutch learned all there was to know about the fabrication and use of transistors. They were joined by other researchers from Europe, representing Telefunken, Siemens, GEC and Ericsson. The future inventor of the integrated circuit, Jack Kilby of Texas Instruments, was also among the visitors. In 1952, the world of semiconductor electronics was still small and orderly.

It was the first time AT&T had opened its doors. The American telecom company’s success had always been based on filing patents and leveraging its technology expertise. It had always responded ruthlessly to patent infringements. But after inventing the transistor, the telecom titan took an amazing turn.

AT&T had a good reason for its about-face. Since 1949, it had been under fire from the US Department of Justice wanting to limit its monopoly by splitting off Western Electric. Countless small companies were eager to start making the transistors that were already beginning to penetrate AT&T’s switchboards. These corporate parasites were fully prepared to suck up Bell’s semiconductor expertise free of charge. AT&T had no desire to play Goliath against dozens of Davids.

And so in the spring of 1952 at the Transistor Technology Symposium, Bell Lab’s researchers were telling twenty-five American and ten foreign companies all the ins and outs of transistor technology. The price was reasonable. For 25,000 dollars – a down payment on future licensing fees – Philips and the other participants could learn everything about manufacturing the superior junction transistor patented by William Shockley in 1948.

The participants were given detailed documentation to take home, and with “Ma Bell’s Cookbook” by his side, Haaijman soon mastered the art. By the end of 1952, the Natlab was already making hundreds of junction transistors. Not much later, it was already developing its own improvements. Leo Tummers and Piet Jochems acquired lab-wide fame as the co-inventors of the pushed-out base transistor, an electronic component launched by Philips around 1958 that became a major commercial success.

From my conversation with Hajo Meyer, I deduce that Philips’ interest was particularly aroused by the possibilities that transistors brought to telephony. AT&T was already making switchboards with the new electronic components, and in 1952, on their visit to Bell Labs, Haaijman and Meyer were accompanied by a director of Philips’ telephony business from Hilversum – it’s not clear whether he also went with them to the famous April 1952 transistor symposium.

Photorepeater

In 1962, Haaijman brought back a chip from a fact-finding trip to the United States: a single slice of semiconductor crystal onto which various electronic components had been soldered and integrated. No one knows where he got it from. There was a rumor at the Natlab that he secretly swiped it when no one was looking, but the true story has been lost to history. The souvenir created quite a stir. The chip’s importance was evident: the Americans had made a giant leap forward in reducing the cost of electronics.

Together with managing director Evert Verwey and group leader Leo Tummers, Haaijman set the course for semiconductor research at the Natlab from the 1950s through the 1970s. They realized they had missed the boat, and several researchers were ordered to drop everything and throw themselves at the new technology. Tummers permanently installed a microscope with Haaijman’s chip on it in his group’s wing of the building. It served as a statement, there to be seen by anyone who still had doubts about the future.

“Look at this,” Tummers told Frits Klostermann, a young researcher that had just arrived at the Natlab. “If they did it, we can do it, too.” Klostermann peered through the microscope and saw, for the first time, what a chip looked like: some kind of strange insect, glittering in all the colors of the rainbow.

Tummers asked Klostermann to put a shift register on an integrated circuit. At the time, a shift register consisted of separate elements soldered together: transistors, resistors, diodes and capacitors. Klostermann’s task was to integrate all those components and connections into one small piece of silicon. In other words, to turn them into a chip.

Klostermann’s biggest problem was the photographic equipment to produce masks to lithographically print tiny patterns onto silicon wafers. He could have the masks made in the US, but that would have taken months. He decided to develop an imaging device himself. In a few months, the young engineer created his first configuration. “A simple projection system has been constructed, which seems to produce photographic images of sufficient quality,” he wrote in his report on the final quarter of 1962.

However, because Klostermann’s bosses had other priorities, his efforts got delayed. Ultimately, in 1967, he did produce a Photorepeater. This instrument contained much of the technology needed to produce a wafer stepper, an effort that started at the beginning of the 1970s.