The tech engineering pioneer with Irish roots

While researching her family’s history, Lynda Connolly, Inclusion Project Manager at Intel Ireland, uncovered an intriguing story that takes us from Ireland to the United States and leads to the development of a critical process in transistor design.

In 1911, Thomas (Tom) Kerwin moved from the small town of Mountrath in Co. Laois to the United States. Tom made the City of Quincy, Massachusetts his home. He took advantage of the opportunities available to him, joining the U.S. Navy during WWI and later forging a career as a police officer. Tom married Elizabeth (Liz) O’Driscoll in 1925. Like Tom, Liz had made the move from Ireland to Massachusetts, having grown up in Rathmore, Baltimore, Co. Cork. The couple had four children: Thomas, William, Elizabeth, and Robert.

Tom and Liz’s youngest child, Robert, or Bob as he is known, was raised in Quincy, Massachusetts attending the Massachusetts Fields School, and North Quincy High School. In 1954, he earned a Bachelor of Science from Boston College and was awarded a Master of Science from MIT in 1958. In 1964, Bob completed a PhD in Chemistry at the University of Pittsburgh, serving as a Junior Fellow in Polymer Science at the Mellon Institute. At that time Bob joined AT&T’s Bell Telephone Laboratories as a Member of Technical Staff in the Electronic Component Processes Laboratory in Murray Hill, New Jersey.

At Bell Labs in early 1966, Bob’s team was tasked with finding a solution to two major challenges that persisted in transistor development. First, it was difficult to maintain the surface cleanliness of the silicon wafer necessary to prevent current leakage and deterioration of the gate structure in MOS devices. Second, the precise alignment of the gate structure to the underlying source and drain diffused areas was prone to tool and operator induced errors in the photolithographic process. The group proposed to solve the cleanliness problem through the use of a dual dielectric, silicon nitride over thin silicon oxide, which could be deposited early in the processing sequence while the pristine wafer surface was clean. Differential etching steps would have to be developed to make this possible. Bob proposed to solve the alignment problem by inverting the process steps. Instead of attempting to perfectly align the gate structure with the edges of the diffused source and drain areas, put the gate in first and let it define and thus be aligned with the edges of the source and drain through the diffusion process. This would require that the gate structure be capable of withstanding the high temperatures of the diffusion process, in excess of 900 degrees centigrade.

Amongst Bob’s colleagues at this initial group meeting were those with years of experience in studying solid state diffusion processes and their susceptibility to contamination by impurities. They were reluctant to allow him to experiment with gate electrodes with high melting point metals such as vanadium and titanium in their diffusion furnaces. So, he suggested the use of deposited polycrystalline silicon as the gate electrode, betting that it would be made conductive enough by the diffusion itself and it wouldn’t contaminate silicon wafers. The team went ahead and developed the optimum evaporation times and temperatures to deposit polycrystalline silicon, and recipes for differential etching of silicon nitride and silicon dioxide and produced the first MOS transistors with the self-aligned silicon gates within four months. Bob and his colleagues Don Klein and Jack Sarace wrote up the process details and filed a patent application in January 1967 and presented a paper on this at a technical symposium in August of that year. The new process provided a distinct improvement in production yields.

Bob’s innovation in circuit technology created an enduring legacy. According to the Computer History Museum, “Within five years silicon-gate MOS had become the industry standard process for new IC product development replacing bipolar technology in all but the highest speed applications.” In 1980, Intel co-founder Gordon Moore swore an affidavit acknowledging the significant contributions that Bob and colleagues made to silicon gate technology stating that it was “an important part of Intel’s strategy when the company started in 1968, and it has subsequently been responsible for Intel’s growth and success.”

Bob met Gordon in person in 1998 at a symposium at Murray Hill celebrating the 50th anniversary of the invention of the transistor. During a coffee break they discussed the self-aligned gate and its contribution to the success of Intel. In accordance with a consent decree, AT&T was required to make all its patented innovations available to others for reasonable licensing fees. The objective of the decree was to share patents for the purpose of advancing the state of the art rather than for market control. Intel and many other companies took advantage of these licensing opportunities to foster their success in the semiconductor industry. On that day in 1998, long after the founding of Intel, Gordon Moore offered to pour Bob a coffee from the urn as thanks for his innovation. Remembering this moment fondly, Bob says, “[Gordon Moore] was right [to offer a coffee as recompense.] Under the patent rules and the consent decree, it was perfectly right that Intel and other companies that were founded on the basis of AT&T patents did what they did. It was the purpose of the consent decree and it worked very well.” Bob’s self-aligned gate process is still used by Intel across our manufacturing sites.

Bob, along with Donald L.  Klein and John C. Sarace, was the recipient of the Institute of Electrical and Electronics Engineers’ Jack A. Morton Award in 1994, which later became the Andrew S. Grove Award. Bob was credited “for pioneering work and the basic patent on the self-aligned silicon-gate process, a key element in fabrication of very large scale integrated circuits.” Willard S. Boyle, who brought the problem to Bob and his team at Bell Labs, went on to win the Nobel Prize in Physics in 2009 with George E. Smith for inventing the charge-coupled device.

Bob Kerwin’s parents left Ireland over 100 years ago in search of new opportunities. Reflecting on Bob’s story and the transformation of Ireland in the last thirty years, Eamonn Sinnott, General Manager of Intel Ireland, said “The impact of FDI and Ireland being a launchpad for so many companies that are seeking to grow internationally has been a tremendous tide that has lifted all boats here. If people want to participate in the type of leading-edge manufacturing that Intel represents, or work with top MedTech or BioPharma companies, they are all here. It’s a great wonder to see that evolution in one generation. And it is remarkable to learn about somebody like Bob Kerwin, someone with deep Irish roots, that has enabled that.” Opening in 1989, Intel was one of the first of such organisations to establish operations in Ireland. Today, Intel Ireland employs over 4,900 people and contributes more than €1.08 billion to the Irish economy annually. At the first Global Irish Civic Forum held in Dublin in 2015, President Michael D. Higgins spoke of a global Irish family:

As is often stated, emigration has been a defining characteristic of the Irish people. While some often focus on the large numbers who left Ireland during and following the dark days of the Famine, and it is true that it represented a mass exodus of our people from these shores, over the generations it has been our propensity to be a migratory people.

Ireland’s reputation is built both on the people who live here and those who travel to every corner of the world to succeed in their given calling. Bob Kerwin and his family are testament to this. In recognition of Bob’s contribution to science, Intel has established a new award in Bob’s name to be conferred as part of the Intel Ireland Mini Scientist program from 2021 onwards.