The Man Behind the Microchip Page 10
The meeting was the typical technical conference, with multiple sessions running simultaneously, luncheons in the grand ballroom of the stately William Penn hotel, conference-sponsored scenic drives and sightseeing, and a full “ladies program” of get-acquainted coffees, fashion shows, and shopping excursions for the wives. Shockley did not allow himself to be distracted from his primary mission. He spent his time attending lectures, assessing candidates, and grilling professors at the University of Pittsburgh about their most promising students. He wrote page after page of impressions in a small green memoranda book that fit in his pocket.
On October 10, Shockley noted, “Noyce—Philco; has talked sense about surface transistor.” Bob Noyce had presented a paper on “Observations of Channel Formation on N- and P-Type Semiconductors” at the last Electrochemical Society conference. The subject of his talk this time is unknown—the semiconductor session was assembled so quickly that the paper abstracts were not included in the main program—but back at Philco, Noyce was investigating the well-known “punch-through” problem in which the electric field extends through the whole width of the base, which makes the base disappear electrically and causes the transistor to malfunction. His ten-minute talk at the October conference likely concerned this topic. Shockley quickly decided that Noyce was the only scientist worth pursuing at Philco.16
Three months passed before Shockley—busy finding a building, outfitting a lab, pursuing other recruits, and finalizing arrangements with Beckman—called Noyce in Philadelphia. By that time, the famous physicist knew what sort of work would appeal to the young researchers he wanted to attract. Shockley Semiconductor would conduct the finest scientific research, he promised, with an aim towards building a saleable, profitable product. A job at the company would be an opportunity to work not only with him, but also with the finest rising men in the business, on one of the most exciting technologies to appear in the field of physics for decades, the transistor. (“Is your future brighter in another electronics job?” he reminded himself to ask.) It was also a chance to come to California.17
Certainly, the company’s location appealed to Bob Noyce. From his earliest years of shoveling snow through Iowa’s frigid winters, he had fantasized about the Golden State’s towering mountains and sunny beaches. “All Iowans think California is heaven,” he once said. His brother Don’s recent move to Berkeley for a job in the university’s Chemistry Department only added to the region’s attractions. Noyce frankly told Shockley that he had been considering leaving Philco and would love to move to California—especially if it meant a chance to engage in basic research again. Privately, he swore that he would not miss an opportunity to work for William Shockley: “Getting that job,” he told himself, meant he would “definitely be playing in the big leagues.” The call from Shockley had also gotten his competitive juices flowing. “I sort of wanted to see if I could stand up in that league of competition,” he later admitted.18
Betty Noyce was less excited. However wonderful the professional opportunity for Bob, the essential feature of the Shockley job in Betty’s mind was its location in California. Her family’s ties to New England ran deep. She had a network of friends there. Her parents were only a few hours away. The prospect of moving two children under the age of 18 months to the other side of the country, where she knew no one, did not appeal to her. Noyce’s promise that they would buy a nice house in the San Francisco Bay Area, which at the time was more affordable than Philadelphia, swayed her—but not entirely. She agreed only to a two-year trial period in California. He promised that if she was not happy at the end of two years, they would head back east.19
Before Shockley would bring Noyce to California for an interview, he asked him to report to the offices of McMurry-Hamstra, a New York psychological testing firm, for evaluation. Noyce spent an entire day in Manhattan, where he completed a standard IQ test, described ink blots, and played word association games. The testers also asked a question specially devised by Shockley to gauge creative thinking: 127 people enter an elimination tennis tournament. Since it’s an odd number, one player must draw a bye in the first round. How many matches must be played to determine a winner?20
The standard approach to solving this question would be to chug through a multistep process of division and addition. (Pair the 126 people into 63 matches. Add to those 63 winners the player who drew the bye, and then split this group into 32 new matches, which would yield 16 winners, etc.) Shockley was looking for a much simpler sort of solution, one that Noyce gave him. There is only one winner, so 126 people have to be eliminated. Since a person can only be eliminated through a match, it must take 126 matches to come up with a winner. Q.E.D.
Shockley required this extensive testing of every one of his top-level recruits. Few of them thought it strange. Indeed, several had undergone similar aptitude testing for other employers, and one had even taken a college course on “Testing Evaluation for Employees.” What did strike the young men as unusual, at least in retrospect, was the overt attempt to gauge personality. As one of them put it, “They spent, I think, too much effort on whether I liked my mother or not.” At one point, the testers showed the recruit a line drawing of a man with clenched fists standing at the foot of a bed in which a woman lay with her eyes shut. How would you describe this scene, the testers wanted to know. Had the man just killed the woman? Was he worried that she was ill? Was she sleeping? Was he angry at her?21
Why Shockley required this psychological evaluation is a bit of an open question. He had recently fallen in love with a nurse who taught at a psychiatric facility, and this relationship deepened his already keen interest in the workings of the human mind. Perhaps he felt that if there were a way to gain insight into his employees’ personalities and access to their innermost thoughts, why not do it? Such knowledge might help him match scientists for personal compatibility and maintain harmony in the lab.
Noyce completed the tests to the satisfaction of the McMurry-Hamstra evaluators, and quickly accepted Shockley’s offer to fly him and Betty to San Francisco for an on-site interview. On February 23, 1956, Bob and Betty Noyce left the frigid East Coast on a red-eye flight to San Francisco. They touched down at 6 AM to one of the Bay Area’s beautiful Indian summer days. Bob insisted that they find a house near the lab before lunch—“First things first,” he would later recall—and only then did he go to his interview. He got the job.22
SHOCKLEY ASSEMBLED a team quickly. Vic Jones, a young PhD from Berkeley who specialized in plasma and nuclear physics, moved across the Bay in March. A Welshman with bushy eyebrows and boyish good looks, Jones was so genial and unassuming that the regulars at his favorite pub had no idea he was a scientist with a PhD. Shockley hired Jay Last, a rangy 26 year old who looked too young to drive, out of the new solid state physics program at MIT before he even completed his doctorate. Campus rumor held that no one built thinner crystals than Last, who was very proud of his “damn steady hands.” A call to Lawrence Livermore Labs led Shockley to Gordon Moore, a quiet 27-year-old physical chemist at the Johns Hopkins Applied Physics Lab who had been born in a small farm town on the Northern California coast and raised only a few miles north of Palo Alto. Dean Knapic hired two young men whom he worked with in production at Western Electric: Julius Blank, a portly mechanical engineer with a heavy New York accent who thought the job might be “an adventure”; and Eugene Kleiner, a well-born Viennese refugee from Nazi terror and an expert tool builder.23
A professor at MIT referred Shockley to a PhD metallurgist named Sheldon Roberts, who had developed an interest in silicon and wanted to leave his job at Dow Chemical. When Jean Hoerni, a Swiss theoretical physicist with two doctorates, called Shockley looking for a job at Bell Labs, Shockley lured him to his own company. Vic Grinich, tall and thin with curly hair he wore longer than the fashionable buzz cuts, responded to a want ad that Shockley had written in code and published in a scientific journal to screen out insufficiently intelligent applicants. Grinich came from the Stanfo
rd Research Institute, where he had tried to use grown-junction transistors to drive a color television. (He managed to get the video portion functioning, but never got the audio to work.)24
Only a handful of the roughly 20 men who came to Shockley Semiconductor in the first half of 1956 had passed their thirtieth birthdays. A few had worked in private corporations, but most had either recently received their doctorates or had been employed exclusively in academic or government labs. Almost no one had worked directly with semiconductors, which were still considered esoteric devices. Many of the researchers, such as Gordon Moore—who had been so frustrated by his government-funded research that he once calculated the taxpayers’ cost-per-word he published—were drawn to Shockley by the prospect of “actually making a product and selling it.” The chance to live in California excited them all. In every case, however, the biggest attraction was Shockley himself.25
Shockley planned to unify this hodgepodge of expertise under his own technical leadership. He made it clear to his young team that he would assign them to projects of his own choosing to be done as he directed. “You would be able to discuss your way, your thoughts about how you might do it,” recalls one former employee, “but at the end of the time, [Shockley’s] view was that unless your thoughts were better than his, you’d do it his way.”26
SHOCKLEY DID HIS BEST for the young men. He flew them and their wives to California at a time when flying was still relatively uncommon. He put them up at the nicest hotel in town, Rickey’s, a wood-paneled beauty of an inn with sumptuous landscaping. He hired them at salaries of better than $800 per month—substantially higher than they had been making in their other jobs. He recommended a real estate agent—his aunt—who could help the men and their young families settle into their new homes. He even apologized for the weather when an unusual cold snap meant that several recruits, who arrived in March with bathing suits in hand, instead found icicles on the fountains outside their hotel room doors.27
A critical mass of employees had arrived by mid-April, and Shockley arranged a welcoming party. Noyce, then in his last weeks at Philco, was determined to attend, even though it meant driving across the country in his four-year-old Chevy, its back seat covered with suitcases. As was often the case with him, he ran behind schedule—so far behind in fact, that he had only gotten as far as Salt Lake City by the morning of the festivities. It was raining when he left Utah and positively pouring when he got to the Bay Area. One of the windshield wipers had given out, and he had smoked without stopping to keep himself awake. By the time he found the party, at 10 PM, the celebration was well under way. His appearance made an indelible impression on another recruit:
He hadn’t shaved, he looked like he’d been living in his suit for a week—and he was thirsty. There was a big goddamn bowl of martinis on the table there. Noyce picks up the goddamn bowl, and starts drinking [from] it. Then he passes out. I said to myself, “this is going to be a whole lot of fun.”28
Meanwhile, Shockley was feeling no pain himself. He spent the latter part of the evening dancing the tango with a rose in his teeth.29
SHOCKLEY MADE TWO KEY TECHNICAL DECISIONS when he started his company. First, he would build transistors from silicon, rather than from germanium, which was then the preferred semiconductor substrate. Shockley had been an early advocate of silicon, dashing off a letter outlining its benefits within weeks of learning, in March 1955, that researchers at Bell Labs had successfully grown silicon crystals as pure as germanium crystals. Although its higher melting point makes silicon more reactive and therefore harder to work with than germanium, silicon (the basic ingredient in sand) is the second-most abundant element on earth, after oxygen. Even more important, unlike germanium devices, which often leaked and had a debilitating tendency to malfunction at high temperatures or in high humidity, silicon devices would function reliably in almost any environment, hot or cold, wet or dry. Since the world’s largest potential transistor buyer, the Department of Defense, was willing to pay top dollar for stable equipment, silicon transistors seemed to promise financial success.30
The second critical decision was to build transistors using the new doping process called diffusion, recently developed at Bell Labs. In the diffusion process, a semiconductor is cooked in a furnace containing appropriate impurities (called “dopants”) that then seep into the silicon in much the same way that hickory flavor seeps into meat cooked in a barbecue pit. Diffusion resulted in the best-defined P- and N-regions of any method then available, which in turn meant that diffused transistors ought to be faster and capable of operating at a higher frequency than other devices. The process was important and new enough that Shockley had sent Noyce and Gordon Moore to a Bell Labs seminar to learn more about it.31
The company set to work in a converted Quonset hut at 391 South San Antonio Road, in the heart of an industrial district-cum-shopping center—the nearest neighbor was a Sears store—roughly five miles south of Stanford and footsteps across the Palo Alto border into Mountain View. The building, which one observer likened to an auto-parts warehouse, had insufficient gas and power for a laboratory’s needs, and privately Shockley’s new employees worried about whether or not their boss would outfit the lab appropriately. Their concerns proved unfounded. Shockley bought good scientific apparatus and tools, though he certainly did not bother with any sort of investment in aesthetics, since he planned to move the offices and labs to the Stanford Industrial Park, where Beckman was building a facility, in the fall of 1956.32
The entirety of Shockley Semiconductor Laboratories was housed in a single room. Lab tables ringed the perimeter, and a big desk was planted smack in the middle. Here sat the business manager, who in the beginning was Shockley himself—tangible proof that business stood at the heart of the operation. One corner of the hut served as a modest machine shop for building models and equipment. Semiconductor manufacture was far too new to have generated standard, off-the-shelf equipment.33
FROM THE MOMENT he stepped through the doorway of the Quonset hut, Bob Noyce was in a position of leadership at the lab, simply because he, unlike most of the other new hires, was an experienced transistor researcher. His work at Philco and his long-standing interest in transistors gave him an understanding of the most current theories about semiconductors and practical experience in working with the devices. He had an almost intuitive notion of how things should work in a semiconductor lab. One Shockley employee recalled Noyce building a clean vacuum pump without drawings—blowing the glass himself. Noyce also had clear ideas about how the technology could potentially develop in the future. “In his mind, he could see where it was all going,” observed another co-worker.34
Within weeks of his arrival at Shockley, Noyce was heading up a team that included six members of the senior scientific staff, all of them PhDs. He also helped with recruiting, set salaries for incoming technical employees, and escorted prominent visitors through the lab. When Shockley wanted to estimate how many crystals the company should be able to process per month, he asked Noyce about Philco’s run rate. When he wanted suggestions on technical journals the company should receive, he asked Noyce to gather the information. Shockley also noted several suggestions from Noyce about the diffusion process in his early days at the company.35
To his peers, Noyce was approachable in a laid-back way. “He was somebody you went over to and said, ‘What do you think about this?’” recalled Vic Jones. “He had a very quiet leadership style, a gee-whiz-aw-shucks-farm-boy approach that was very attractive.” Noyce taught without condescension. For example, rather than simply handing orders to the technician in charge of photoresists—the light-sensitive liquid used in the process of transferring a semiconductor pattern onto a wafer—Noyce was likely to chat with him about the types of lenses that would be used in conjunction with the resists. Such information would not only result in better photoresists, it would teach the resists man a little bit about a part of the development process that perhaps he had not known before. Many of Noyce�
��s peers felt that they learned more from him than they did from Shockley.36
The only other employee with knowledge and experience equivalent to Noyce was Smoot Horsley, the first man hired. Horsley was a decade older than the rest of the staff, and while they came to work in shirtsleeves—Noyce once even showed up in shorts—Horsley preferred the thin dark ties and white shirts he had worn for his mission work for the Mormon church. Horsley was intensely loyal to Shockley, and the young scientists saw him as an extension of the boss, not as a peer. Behind his back, they called him “Smooth Horsley.”
Noyce was also quite close to Shockley. The men would swim together and occasionally go for drinks. Shockley’s new wife, the psychological nurse, had become friendly with Betty and regularly stopped by the Noyces’ home to chat or to play with the children while their mother ran errands or worked on the creative writing she still enjoyed. Several employees thought that Noyce’s was the only opinion in the lab that mattered to their boss, who viewed the Iowan, with his quick mind and love of performing, as a youthful incarnation of himself.37
SHOCKLEY SET THE YOUNG MEN TO WORK according to their specific fields of expertise. Sheldon Roberts organized an analytical laboratory so he could better understand the properties of silicon. Others joined him in an effort to grow pure crystals. Julius Blank and Eugene Kleiner worked with Dean Knapic, lathing and jigging many of the lab’s rigs and measuring devices in-house. (They sent their simplest designs to local machinists.) They also built the crystal puller used to grow the silicon ingots. Jean Hoerni calculated diffusion curves, theorizing for how long, and at what temperatures and concentrations, different impurities should be diffused into the surface of the semiconductor. Gordon Moore empirically tested Hoerni’s theories in furnaces he had helped to build.