July 1998 - ChipScale Review

July 1998

eMail the Editor

A Retrospective with Fred Kulicke and Albert Soffa, Two Industry Pioneers

A moment in history

Editor's Note: This year marks the 50th anniversary of the invention of the transistor by Jack Kilby of Texas Instruments and the late Dr. Robert Noyce of Intel. Not coincidentally, it's been more than 40 years since Fred Kulicke and Albert "Buddy" Soffa founded the company that still carries their name and has become the largest manufacturer of chip assembly equipment in the world.

Fred Kulicke (standing left) and Albert Soffa (with paper) demonstrate an early K&S wirebonder at a 1966 U.S. Commerce Department-sponsored trade show in Frankfurt, Germany. At the microscope is Helmut Seier, then K&S vice president of European Operations. On the right are two unidentified Commerce Department Officials.

Retired from Kulicke & Soffa Industries Inc., Willow Grove, Pa., since the early 1990s, Fred Kulicke and Albert Soffa are now enjoying the fruits of their labor- Fred through his painting and Albert by supporting infant industries in Israel through his work in the creation of "technology incubators." The interview that follows is a compilation of several interviews done at different times by Leon Oboler, Contributing Editor and a former Kulicke & Soffa employee, especially for Chip Scale Review.

Leon: How did you begin manufacturing semiconductor equipment?

Albert: The first contact with anything in what was to become the semiconductor industry was in 1957, when we were called in by a salesman who was selling us clutches. He said, "Come to Reading, Pa.,-they're having problems with clutches." At the time, we were an engineering jobshop and had lots of experience designing and making machines, although we had stopped doing special machines and had a product - automated machinery for breweries. We had done some work making machine drives for Bell Labs at Western Electric in Reading. The engineers there were enamored of us because we solved their problems. It turned out that they were making germanium diodes and the mesa transistors that preceded planar transistors. The experimental line was making around 10 per day manually. Western Electric asked us to look at its factory and figure out what we could do to help automate the process. In 1957 we had heard about transistors but didn't really know what they were. The Japanese were already selling radios that were called 'The Transistor,' but we'd never paid much attention to how they were made. Originally both the Japanese and Western Electric used women who would pick up a wire by hand and weld it to the transistors by thermocompression (to the die) and spot welding (to the posts). For Western Electric, we did a flowline of all the machinery that would be necessary to make transistors in 1957. After working out the process, we said, "Let us do what we do best," and that turned out to be wire and die bonding. Die bonding was eutectic. Operators would pick up a germanium chip with tweezers and rub it on a piece of gold leaf that had been heated to around 600 degrees in a (TO-5)can. The resulting alloy created a strong mechanical and electrical connection. The cans required special inconel molds. We worked with others to drill them manually. Ultimately we designed an automated machine that used the Western Electric process, but did it in series.

Fred: The biggest problem with wire bonding was the wire. They didn't know how to draw gold wire down thin enough and they couldn't make any wires more than a few inches long. The initial solution was the Walston process which was invented in the last century. They would encase a thin gold rod in a silver tube. The multi-layered rod would then be swaged down to the diameter they needed. After etching away the silver, the gold wire would be around 0.5 mil in diameter. Later, a jeweler was finally able to develop a process for drawing very fine wire. The key was in realizing that you couldn't do it too fast. The jeweler was able to make longer wires and also sold wire on spools which made the wire bonding process faster because you didn't have to spend as much time picking up loose wires.

LEON: How did the business develop?

Fred: Our first machines were actually sold to RCA, who was just getting into the business. They were doing a job for the Air Force under a Western Electric license.

Albert: The government required Bell/Western Electric to license its process to any company that wanted to manufacturer transistors. Since we had developed the machines for the Western Electric process they sent us any licensee that came in (during 1959-60). We had special access to any Western Electric plant, and Bell had licensed everybody in the world.

Fred: [laugh] We had an unnatural lock on the market ... but it was a very small market.

Albert: Initially we made eutectic die bonders and thermocompression wire bonders. In 1959 we made a mask aligner. This aligner used a metal mask to direct-deposit material onto the stripe on the germanium wafer in a vacuum chamber. Eventually we made everything from cleanroom boxes to saws, but we stayed out of the chemical processes.

Fred: The quick changes in the industry meant we had to do R&D on processes before we were sure we needed them.

Albert: Fred even experimented with creating planar leads, such as those used in tape-automated bonding or the newer chip-scale packages, back in the early 1960s. No one ever used them until National introduced TAB.

Fred: Ball bonding came in when silicon replaced germanium as the primary wafer material. We needed to come up with a way that would not crack the fragile silicon die. It turned out that the answer was ultrasonic scrubbing-which was developed by a company in West Chester, Pa., named Sonibond. It's funny how things come together. When we first got into bonding at Bell Labs, no one knew what was actually going on. We knew soldering, but all Bell used was a wire and a heated surface. I realized later that this was similar to the blacksmith hammer-welding I'd learned in the artillery: clean the materials, heat them to plasticity and lay them together with pressure. Who could have imagined that such an old process would be the basis of manufacturing in the 21st Century?

LEON: Is there anything you would like to say in summarizing your experiences?

Albert: I think it was a very interesting dynamic in the beginning to see how companies bred clusters of businesses. People would leave and set up small machine shops and then get together or grow. A good example was Hughes on the West Coast that resulted in a group of companies around Stanford University. Within weeks, the venture capitalists would fund companies-like Intel and Fairchild-all from developments around Hughes. K&S was a big part of that because people would call us and ask to buy whatever we had because they had to have a showroom for investors. Things were moving really fast. I guess they still are.

Fred: I think what impressed me was that everything we had to do was new. That gives you real satisfaction.

(For more history and information about Kulicke & Soffa today, see the Web site at www.kns.com.)

Chip Scale Review o 7291 Coronado Drive, Suite 8 o San Jose, CA 95129 o Email: editor@chipscalereview.com

Features, 98/07/27, 05/13/99, ID=9807/kulickesoffa1

© 1998 ChipScale REVIEW