The War of the Patents

Jack Kilby vs Robert Noyce

Robert Noyce’s invention of the monolithic IC, is enough to put him in the Computing Hall of Fame. This in turn, fuelled the PC revolution, and also gave Silicon Valley it’s more famous moniker, not to mention a certain Intel.

Noyce was one of the Traitorous 8, or the Fairchild 8, the men who left Shockley Semiconductor Labs in 1957, fed up with the Nobel Prize winner’s rather whimsical style of working and founded Fairchild Semiconductors.

Noyce’s major contribution would be two- the IC and later on Intel. I will be looking more at his work on the IC.

The Integrated Circuit

Though Robert Noyce was the one who got the patent for Integrated Circuit in 1961, he shares the credit with Jack Kilby. Matter of fact the late 1960s were marked by the Patent Wars, over the Integrated Circuit, with a whole lot of claimants trying to claim credit for it. It was finally narrowed down to 4- Noyce, Kilby, Kurt Lehovec and Jean Hoerni. And finally it became Noyce vs Kilby.

Though Jack Kilby was finally given the Nobel in Physics(2000) for his work on the invention of the IC, later day analysts felt that many inventors contributed in their own way to the invention of the IC, and Kilby cud not be getting sole credit. But if one looks at a very broad level, the 4 people who contributed the most to the development of the IC were Kilby who created the first prototype, Lehovec who electrically isolated components on a semiconductor crystal, Hoerni who came up with planar tech and Noyce himself.

But the major patent battle for IC during the late 60s was between Kilby and Noyce, both representing big companies, Texas Instruments and Fairchild. In effect it also became a fight between these 2 companies each claiming credit for the IC. The trigger for the debate was giving the Ballantine Medal to both Kilby and Noyce, in 1966 for development of IC. Many claimed that the prototypes invented by Kilby were not the actual ICs, and Noyce was not directly involved either in development of IC.

Noyce himself admitted later, he was only trying to resolve a production issue, not the invention of the IC itself. However the ongoing patent wars, between Texas Instruments and Fairchild, ensured Noyce ended up with the credit for the IC. Jack Kilby himself led the PR campaign for Texas Instruments, claiming credit for IC, challenging the patent for Noyce. On the other hand Noyce was not directly involved in the campaign, Gordon Moore was the one who led it.

This Noyce vs Kilby rivalry in Integrated Circuits, is on par with other legendary rivalries like Gates vs Ellison, Edison vs Tesla in the science and tech world. By the mid 70s, however Kilby and Noyce were considered the main inventors when it came to ICs. Finally during the 80s it was recognized that while Jack Kilby was the one who invented the IC, Robert Noyce was the one who developed and improved upon the original version. From then the 2-inventors theory has been widely accepted for Integrated Circuit.

The main cause for development of IC, was due to something called the "tyranny of numbers", a phenomenon by which the loss due to downtime in computing devices exceeded whatever benefits were expected from them. Take the ENIAC, it had more than 17,000 vaccum tubes, something went wrong, that meant you had to check each of these tubes, and that meant troubleshooting often took a whole lot of time. Just imagine inspecting something like this.

Invention of the transistor in 1947, reduced the size and power consumption, however the reliability still remained an issue in the electronic devices, add to that the rather tight packing, meant problems with connection and troubleshooting too. The actual concept of IC was propagated by British engineer Geoffrey Dumeer in a 1952 speech, where he claimed that electronic equipment could be placed in one solid block with layers for insulation, conduction, amplification. And this would remove the messy wiring.

While there were patents by Sidney Darlington and Bernard Oliver, it was Harwick Johnson who came up with the idea of combining transistors, resistors and capacitors on a single chip. However his idea was more theoretical in nature, but the design was used later.

In the meantime Bell Labs, General Electronics, IBM and RCA tried to come up with their own solution to the tyranny of numbers. Some of em experimented with thyristor based memory cells, however while it did work to an extent, large scale industrial production was not possible. Another very significant development that took place was using silicon over germanium for transistors. The advantage was that silicon could operate at much higher temperatures, was stronger and chemically inert insulator. Jean Hoerni proposed the planar technology in 1957.

There were 3 main factors hindering the devlopment of ICs, one was Integration of different components in one semiconductor crystal, another was isolating the components and finally connecting the components without using electrical wires. Jack Kilby began working for Texas Instruments since 1958, he already had lot of experience in radio, was a WWII veteran. He was the one who came up with the 3 rules of integration, tat would be the cornerstone for ICs'.

Kilby's 3 laws of integration were

1) Only thing a semiconductor company can produce well is semiconductors

2) All circuit elements can be made of a semiconductor

3) All circuit components can be formed on one semiconductor crystal.

Jack Kilby began assembling the IC prototype in August 1958 and on September 12 he presented it , a single transistor oscillator, an improvement on Johnson's 1953 patent. One week later he came up with the 2nd protoype with 2 transistors.

Though Kilby's design managed the integration part, the isolation and interconnection still needed to be worked upon. Wires were still used for interconnecting and components were separated by cutting grooves on chips. Texas Instruments however went along with Kilby's idea to military customers, most of whom rejected it, however the US Air Force, felt it was suitable for their molecular electronics program and ordered for the prototypes.

In April 1960, Texas Instruments announced Multivibrator #502 as the world's first IC on market, and set a price of around 300 USD. This had two transistors, four diodes, six resistors and two capacitors, the traditional discrete circuitry.

However isolation was still an issue and the solution to this was found by Kurt Lehovec, working with Sprague. As per Lehovec, a p-n junction had a high impedance to electric current, so insulation cud be achieved between two components by using large number of p-n junctions. Lehovec came up with a device which was a linear structure, divided into isolated n-type cells by p-n junctions, conductivity type was determined by the crystal's pulling speed. Though the management showed no interest, he filed a patent in 1959 and then left US.

What Robert Noyce essentially did was to work upon the designs of Kilby and Lehovec, and reinvent it using the planar process described by Jean Hoerni. He described his hypothetical solution to Ed Keonijan, who developed the onboard computer for Atlas rocket. Noyce's contribution to IC was two fold, he improved upon Lehovec's design, the first layer being a silicon chip, passivated by an oxide layer. Impurities wud be diffused to create low resistance wells, and planar devices placed in em. This wud create a 2-D structure. The other major contribution of Noyce would be solving the interconnection problem in ICs. This was what prevented mass production of ICs, what he proposed was preserving the oxide layer, tat separated the chip and metallization layer.

The first planar IC was developed at Fairchild on May 26, 1960, by a group formed by Noyce himself, and headed by Jay Last. An improvement on this design came in August using the p-n junction isolation and finally the first operational device came out in September. Fairchild however did not show much interest in the work of the group, and their Marketing VP, felt that the entire IC project was a waste of resources, one of the many blunders that led to it's downfall.

In 1961, the "traitorous 8" left Fairchild, that included Noyce and Moore, Last and Hoerni who had worked on the IC too left it, and joined Amelco. While other team members David Allison, Lionel Katner formed Signetics. It was a major blow to Fairchild. Trying to come to terms with the departure of it's leading technologists, Fairchild tried comming up with with their first commercial IC series Micrologic. Texas Instruments had already come up with a series of planar ICs for space satellites by then.

Fairchild designed the ICs for the onboard computers of the Apollo spacecraft, though manufacture was by Raytheon, while Texas Instruments developed for the Minuteman ballistic missiles. In a way NASA, created a non military market for the ICs. By 1964 both Fairchild and Texas Instruments, replaced the resistor-transitor logic of ICs with diode transistor logic to overcome electromagnetic interference. While Fairchild had the larger numbers, during 1961-65, TI was ahead in the revenue, had 32% if market share.

One reason why Noyce's IC design was better than Kilby's was use of silicon far more resistant to higher temperatures than germanium. Noyce succesfully combined Lehovec's isolation principle, that allowed each transistor to operate independently. Later on Noyce co founded Intel with Gordon Moore, and soon Fairchild lost the race, as Intel drove the changes, and with addition of Andrew Grove, went on to become a giant, led to foundation of Silicon Valley, that is a story for another day.