Skip to content

Jean Hoerni and the Planar Process

Abstract

Jean Hoerni (1924–1997) was a Swiss theoretical physicist with two doctorates who became one of the “traitorous eight” founders of Fairchild Semiconductor in 1957. That December he wrote down an idea that ran against standard practice: leave the oxide layer on the silicon after diffusion so that it seals the transistor’s junctions. Developed in 1959 and sold from April 1960 as the 2N1613, the planar transistor was flat, protected and made from one side of the wafer, which is what let Robert Noyce wire several of them together on the same chip. Every integrated circuit since has been made by descendants of the process. Hoerni left Fairchild in 1961 and founded or ran three more semiconductor companies, the last of them Intersil, a pioneer of low-power CMOS chips for watches. He spent much of his fortune on schools in the Karakoram.

Geneva, Cambridge, Pasadena

Jean Amédée Hoerni was born in Geneva on 26 September 1924, the son of a bank manager. He took a degree in mathematics at the University of Geneva and then two doctorates in physics, the first in Geneva and the second at the University of Cambridge in 1952. The same year he moved to the United States as a research fellow at Caltech in Pasadena, where he worked with Linus Pauling.

In 1956 William Shockley, who had met him at Caltech, hired him for the new Shockley Semiconductor Laboratory in Mountain View. Hoerni was the group’s theorist. Like the others he grew tired of Shockley’s management, and on 18 September 1957 he was one of the eight who resigned together and founded Fairchild Semiconductor with a $1.38 million loan arranged through Sherman Fairchild (see Robert Noyce and Fairchild Semiconductor).

The Notebook Entry

Fairchild’s first products were silicon mesa transistors. In the standard process of the day, impurities were diffused into the silicon through openings in a layer of silicon dioxide, the oxide was then stripped off, and the transistor was etched into a raised plateau. That left the p-n junctions exposed at the edges of the plateau, where dust, moisture and loose particles could reach them.

On 1 December 1957, on pages 3 and 4 of his Fairchild patent notebook, Hoerni wrote down a “Method of protecting exposed p-n junctions at the surface of silicon transistors by oxide masking techniques”. His proposal was to keep the oxide: do the diffusions through it, leave it in place, and open only small windows for the metal contacts. The junctions would then meet the surface underneath a layer of glass. The idea sat in the notebook for more than a year while the new company worked to ship mesas. On 14 January 1959 he turned it into a patent disclosure, and in March 1959 he demonstrated a working planar transistor. The patent application went in on 1 May 1959 and produced US patents 3,025,589 (issued 20 March 1962) and 3,064,167.

Myth: Hoerni revived the planar idea to save Fairchild’s Minuteman contract

Many Fairchild histories say Hoerni dug the idea out of his notebook in response to a reliability crisis: in 1959 transistors Fairchild supplied to Autonetics for the Minuteman missile’s guidance computer failed, and shipments stopped. The dated notebooks show the opposite order. Gordon Moore’s tap tests found the cause, flakes of metal and “translucent crud” shorting the exposed mesa junctions, only in July 1959, at least six months after Hoerni’s January disclosure. Until the planar process took over, Fairchild kept shipping by mounting every transistor in a test circuit and pounding it with a hammer. What made Hoerni return to his idea in January 1959 is not recorded. See Myths and Misconceptions.

Planar

Fairchild introduced the planar transistor commercially as the 2N1613 in April 1960, and Hoerni presented it at the Electron Devices Meeting in Washington that October. It was more reliable than the mesa and soon replaced it. Fairchild licensed the process across the industry.

The second consequence mattered more. Because every region of a planar transistor lay in one flat surface and was made from the top of the wafer, other components could be made in the same surface and connected by a metal layer laid over the oxide and etched into wires. Robert Noyce wrote this down on 23 January 1959, nine days after Hoerni’s disclosure, and filed his integrated-circuit patent in July 1959; Fairchild had working planar integrated circuits by May 1960 (see The Integrated Circuit Revolution and Jack Kilby and the Integrated Circuit). Kilby’s first circuit at Texas Instruments, in 1958, was a sliver of germanium whose parts were joined by gold wires attached by hand. The Computer History Museum’s milestone page quotes one historian calling planar “the most important innovation in the history of the semiconductor industry.”

Hoerni’s notebook from those years also records a gold-doping technique that made silicon transistors fast enough to compete with germanium in high-speed computers, including the CDC 6600. In March 1960 he worried there that epitaxy, growing a thin crystal layer on the wafer, might make Fairchild’s work obsolete; instead epitaxy was combined with planar processing and raised its yields (see Semiconductor Fabrication).

Amelco, Union Carbide, Intersil

On 31 January 1961 Hoerni and Jay Last left Fairchild to run Amelco, the new microelectronics branch of Teledyne; Eugene Kleiner and Sheldon Roberts joined them a few weeks later, so that half the eight had left. Hoerni then moved to Union Carbide to set up its semiconductor business, Union Carbide Electronics (1963 in the National Inventors Hall of Fame’s account, 1964 in others). In 1967 he founded Intersil, financed largely by European investors. Intersil worked on low-voltage CMOS circuits and made chips for some of the first electronic watches and for calculators. The company was later bought by General Electric. Hoerni also consulted for Hughes Aircraft and Fujitsu. He was openly critical of the Swiss computer industry.

The Karakoram

Hoerni was a mountaineer and ski tourer of the old school, by the Heinz Nixdorf MuseumsForum’s account sleeping out in a bag stuffed with newspapers. In 1974, on a trek, he came to the Karakoram in northern Pakistan and kept returning. He set up a $1 million endowment to bring schooling and health care to the Balti villages there, paid $12,000 for a school in the village of Korphe that the climber Greg Mortenson had promised to build, and in 1996 founded the Central Asia Institute with Mortenson to build more.

He did not see what became of it. The Institute grew famous through Mortenson’s book Three Cups of Tea (2006), which tells the story of Hoerni’s first gift. In 2011 the book’s account and the charity’s finances were publicly challenged, and in April 2012 a Montana attorney general’s investigation found that Mortenson and the board had mismanaged the Institute; Mortenson agreed to repay it at least $1 million spent on personal expenses.

Recognition and Death

Hoerni received the Franklin Institute’s Edward Longstreth Medal in 1969 and the IEEE Computer Society’s McDowell Award in 1972. He died of myelofibrosis in Seattle on 12 January 1997, aged 72. His papers are at Stanford University. He was inducted into the National Inventors Hall of Fame in 2009 for patent 3,025,589.

📚 Sources