Raytheon Part Two

From Transistor to Spacistor

Semiconductor Research and Development at Raytheon

This is Part Two of From Transistor to Spacistor, Semiconductor Research and Development at Raytheon. For Part One click here.

First RF Transistors

The first objective for semiconductor companies developing radio frequency transistors was to achieve acceptable performance for a portable transistor radio. This required an oscillator/mixer stage that would operate to the top end of the broadcast band (1.6 Mhz) and an intermediate frequency amplifier (455 Khz).

Raytheon had viable RF transistors by mid 1954 and was developing prototype portable radios using them.

Uncharacteristically, Raytheon sought no publicity on the release of its CK760-762 range, unlike, for example, the launch of the CK722. It did have every reason to celebrate since these transistors were amongst the first RF transistors to be made available but they needed to reserve stock to support the planned production of their eight transistor portable, the 8TP.

The earliest notice of the new range found to date is a brief trade announcement in Radio & Television News in March 1955 which states that Raytheon had “recently announced the availability of three radio frequency fusion-alloy transistors.” [scan courtesy Joe Knight]

First journal advertising of these transistors was combined with the launch of the Raytheon 8TP transistor portable in May 1955.

In doing so Raytheon advertised these as “another Raytheon first! Genuine RF transistors that make possible this all transistor 455 kc IF portable.” This milestone refers to the fact that the first portable produced, the Regency TR-1 by Texas Instruments, used an intermediate frequency of 262Khz because Texas Instruments could not produce transistors that would amplify reliably at the more conventional 455 Khz.

The transistors were “completely interchangeable without selection of components” referring to the fact that the Texas Instruments transistors needed tailored neutralising components to suit each individual IF transistor.

Other features claimed were:

Successfully field tested for an entire year

Available in production quantities

Hermetically sealed

Equipped with standard military lead spacing

Made by the Raytheon perfected fusion alloy process that has already produced nearly two million transistors

Specifications for the transistors as at April 1955 were as follows. Later revisions lifted the specifications of these transistors.

Notably the CK759 was not included in the initial release and may have been introduced later as a means of utilising low specification transistors that were usable as IF amplifiers.

The CK760 was adopted as a universal RF transistor and initially was used in all stages of the 8TP receiver in its first configuration. [McGarrah]

Other than the above parameters, all the transistors in this range had identical specifications suggesting that they were made, tested and selected.

Raytheon research into improved RF performance focussed on methods to reliably reduce the base width of its transistors. Herman Nowak developed a method of obtaining alloy junction transistors with base width of only 1 mil by milling notches set at right angles on either side of the wafer and creating the base junction at the intersection of the notches. (This geometry gives a stronger wafer than if the notches were cut parallel). [Nowak 1954]

In 1956 George Freedman developed a method of controlling base width by starting with a wafer that included a PN junction and fusing an electrode containing a P impurity to an exposed N layer. Base width was controlled by monitoring electrical characteristics of the junctions as alloying proceeded. [Freedman 1956]

There is no evidence that Raytheon actually used either of these approaches. What is on record is that Raytheon used a conventional approach of reducing collector capacity and making the base width as narrow as possible. They did this by using smaller indium dots and reducing the wafer thickness. Frank Ducat recalled that this reduced collector capacity from typically 20 pF to 12 pF and improved alpha cut-off frequency from typically 4 MhZ to 13 Mhz. [Ward 2001]

Transistor Radio Market

Raytheon developed its transistor radios at its Chapel Street application laboratories. This work was led by Sheehan and Ivers who published a schematic for a similar 8 transistor set in Electronics magazine. [Reproduced in Carroll 1957]

The claim that the RF transistors had been under development for “12 months” is validated by the record of a presentation of a seven transistor portable to Massachusetts Governor, Christian Herter at the Raytheon Millionth Transistor celebration in the second half of 1954. (photo left from Electronics October 1954)

Production was undertaken by the Belmont Radio Company of Chicago who had a private label business making radios for Montgomery Ward, Western Auto, Federated and associated department stores using the Freshman, Goodyear, Wings, Imperial, Stark, Classique, Truetone and Wings brands. [Wolff 1984, Radiomusuem] Raytheon had acquired Belmont in 1944.

Raytheon launched the model 8TP portable transistor radio around May 1955 using CK760 transistors in its RF and IF stages. This was the second all transistor radio to market: the first being the Regency TR1 that had been released in November 1954 to take advantage of the Christmas rush. [Photo courtesy of Bob McGarrah] For more information on the 8TP see Bob McGarrah's site.

The 8TP was an eight transistor set with a performance comparable to portable tube sets of the time. But it was more expensive at $79.95 than the four transistor Regency which was released at $49.95. The set was also released under the Western Auto Truetone brand in a different case. [McGarrah]

“The April and July 1955 issues of Consumer Reports separately put these two radios to the test, and concluded that Raytheon had every reason to call its 8TP the first serious transistor radio. The April '55 review of the Regency TR-1 found the $49.95 TR-1 to be a toy-like novelty which didn't come at a toy-like price, and stated that, "the consumer who has been waiting for transistor radios to appear would do well to await further developments before buying."The July '55 review of the Raytheon 8TP gave the set high marks: "The transistors in this set have not been used in an effort to build the smallest radio on the market, and good performance has not been sacrificed to attain this end." The 8TP series was ranked high in nearly all categories, "falling down only in sensitivity."” [Davidson 2007]

In order to promote the new portable to a public that had no exposure to transistors Raytheon gave away cards with a genuine dud transistor saying “Take this actual sample of an electronic miracle!” with an explanation of the benefits of transistors in the new set. [Photo courtesy of Joe Knight. Full size scan here]

Supplying the Market

Raytheon put little energy into promoting its RF transistors: this was because demand outstripped supply, a feature of the market in general. For example, Zenith produced its Royal 500, a seven transistor pocket sized set launched in November 1955, with four chassis types to accommodate four different transistor line-ups: one supplied by Raytheon (CK760 & CK759)) two supplied by Sylvania (2N94 and 2N194 & 2N193) and one by Texas Instruments (2N145). “This expedient was necessary to enable us to produce sufficient quantities by using transistors from many sources” according to its service manual. The majority of Royal 500 sets were equipped with Sylvania transistors which they sold to manufacturers at a lower price than did Raytheon. [McGarrah] More information on the Royal 500

Norman Krim confirms the problems of supplying RF transistors talking to Michael Woff: “Radio Shack was in Boston at the time, and they were able to get radios from Japan but they couldn't get transistors at all. Maybe that occurred a year or two later. They were after me for transistors.” [Wolff 1984]

By the end of 1955 the following transistor portables were on the market as shown in the table with their principal transistor supplier:

This shows that Raytheon did not have a market leadership role: in 1955 the only third party suppliers were Texas Instruments and Sylvania. Both these manufacturers had RF transistors in late 1954 before Raytheon announced its entrants.

Raytheon did better in 1956 and by the end of that year it was supplying Bulova, Dewald, Emerson and Hallicrafters. General Electric had a similar customer count. Texas stayed well ahead supplying Admiral, Columbia CBS, Halicrafters, Magnvox, Motorola, Sentinel and Westinghouse. In 1957 Texas picked up Bulova from Raytheon and Raytheon obtained Firestone.

Raytheon was active with its own range of transistor receivers having introduced the FM101A, T100, T150 and T2500.

Speaking to the success of Texas Instruments in the market place Roger Webster, who worked on the Regency TR-1, strongly endorsed the advantage of being first to have its transistors in the first transistor portable and the credibility it gave his company. “There was a radio design section that contacted all of the radio manufacturers – Warwick, Emerson, Zenith, you name it. We talked to all of them.” [Wolff 1985]

Silicon Semiconductors

The development of a Raytheon silicon transistor was initiated by Ivan Getting soon after he joined Raytheon in 1951. [Krim 2009]

Getting recalls “While the Receiving Tube Division was emphasizing commercial germanium junction transistors, the Research Division addressed the problem of silicon junction transistors... Knowing that BTL was heavily engaged in silicon research, I called Jim Fisk, now the Director of Research of BTL, for a technical exchange visit.” The request was declined but while Bell had declined to give Raytheon any assistance from its own silicon transistor programme, it did arrange for Raytheon to be supplied with pure silicon from their supplier.

At this time Raytheon’s licensors, Bell and General Electric had not developed their own silicon transistors. Of these two companies Bell was most advanced: Morris Tanebaum commenced a programme on silicon transistors in 1953 using at first the crystal pulling and double doping technique Bell had used in its early germanium junction transistors. This failed because the quality of the silicon used gave insufficient minority carrier lifetime even across the thinnest base widths they could create (as little as one mil). So Bell tried the rate growing technique invented by Robert Hall of General Electric [Hall 1952] which gave even thinner base widths eventually enabling a viable transistor to be produced early in 1954. [Tanebaum 2008] Bell also had an alloy junction silicon program which had produced transistors by mid 1954 [Kurshan 1954] and a year later had produced the first silicon double diffused transistor. [Tanenbaum 2008]

Unlike other companies, Raytheon started their silicon transistor development on two parallel tracks exploring the following approaches:

(1) The grown junction transistor

(2) The alloy junction transistor

The technology issues were formidable. Due to the high melting point of silicon (1410 C compared to germanium at 937C) producing high purity doped silicon was difficult. Inadequate purity meant poor minority carrier lifetime in the base region of transistors resulting in low alphas.

Silicon has a much lower coefficient of expansion than the metal alloys needed to make base connections or alloy junctions and this can result in stress cracking and unreliable contacts.

Raytheon Silicon Grown Junction Transistor

George Freedman, Walter Leverton, Fowler and Straub worked on the program but few details are available. One of the problems in making such transistors was to make a connection to the thin base layer without shorting the emitter or collector junctions. George Freedman at Raytheon patented a method of fusing a taut tungsten wire only 0.15 mil in diameter by coating it with a P-type solder and using electric heating of the wire to solder it to the base layer. [Freedman 1955]

By late 1953 Raytheon produced its first silicon transistors. Writing to the New York Times on 20^th January 1954, Ivan Getting referred to the problems of working with silicon noting “We here at Raytheon have had a large and very brilliant group working on these problems for a long time. I think we have substantially led the field in various steps to the point that during the last few months we have been making and testing silicon transistors. These have been tested at temperatures from 27 degrees C to 185 degrees C - the melting point of solder.”

In his biography Getting says that Raytheon first demonstrated a silicon transistor at the annual March IRE show in New York showing one operating in boiling water. This is confirmed in the edition of Electronics for March 1954 (including its picture, left) which reported that Raytheon had produced a grown junction NPN transistor with a base width of 1 mil and a gain of 40dB but that “Raytheon pointed out that quantity production was some time off. Silicon transistors are not expected even then to surplant germanium units.” [Electronics 1954] Getting attributed their success ahead of Bell due to the quality of the silicon each of its research team was using. “It seems that BTL had been receiving the purified silicon for well over a year. It was issued to the researchers in the order it was received. Thus the purified silicon that BTL was using was about a year old, whereas we received freshly refined silicon directly from BTL’s suppliers. Apparently, the purification process had been steadily improving and the older stuff contained traces of oxygen that substantially reduced electron mobility. In any case, the Raytheon Research Division was the first to produce operating high gain silicon junction transistors!” [Getting 1989]

Bad news was in store for the development team. On May 10, 1954 at the IRE national conference in Dayton, Ohio, Gordon Teal of Texas Instruments announced that his company had the silicon grown junction transistor in production. Recalling the astonishment created by his paper and announcement Teal mentions “One man from Raytheon put in a call to his executive vice president and was heard in the booth croaking hoarsely, 'They got the silicon transistor down in Texas!'” [Goldstein 1991] Norm Krim confirms this: Frank Ducat called him with the message no one at Raytheon wanted to hear. [Krim 2009] The Raytheon program on silicon grown junction transistors had been gazumped by Texas Instruments.

Following the announcement by Texas, Raytheon disclosed its work on a silicon grown junction NPN transistor the following month at the Conference on Semiconductor Device Research at Minneapolis June 28-30 1954 in a presentation by Leverton, Fowler and Straub entitled “Grown Junction NPN Transistors.” This followed a presentation by Texas Instruments on their transistor. An attendee at the conference, Herbert Nelson, taking notes for RCA wrote “the material contained in this paper was largely a repetition of that in the Texas Instrument paper though the transistors described apparently had not been developed to the same high degree of perfection.” [Kurshan 1954]

Raytheon Silicon Alloy Junction Transistor

Most companies followed the alloy junction approach since that had been so successfully pioneered in germanium. Hermann Statz, Sumner Wolsky, John Spanos and John Williams worked on the Raytheon silicon alloy transistor.

The principal problem needing solving was stress cracking between the silicon and either the alloying materials or the base tab connection due to differential thermal expansion exacerbated by the higher operating temperatures sought in silicon transistors. By the time Raytheon launched their silicon alloy transistors in 1956 there was a substantial body of published work, notably by RCA.

RCA developed a silicon alloy junction NPN transistor, the SX-152 from March 1954. [Knight 2008] This transistor was never commercialised. The work was presented by Herbert Nelson at the 1954 Minneapolis Conference and published. [Nelson 1954]

Most of their research related to alloy dot compositions that minimised thermal mismatch strains while permitting good wetting. At the same conference Hughes presented similar work on desirable alloy compositions to minimise thermal stresses and Bell also gave a presentation on their alloy junction transistor.

Sumner Wolsky and John Spanos at Raytheon developed two solutions to the problem of making strain free base tab connections for PNP and NPN transistors respectively. Their strategy was to use multiple materials so that strains were spread across more than one interface. [Wolsky 1954 Spanos 1956]

To make PNP transistors Raytheon assembled an N-type wafer, aluminium dots and the base tab in jigs designed to hold these components accurately in position while they were baked at temperatures between 800C and 1000C. Produced in this manner the aluminium dots resisted soldered connections due to an oxide layer on the aluminium. Late in the development program John Williams patented a method of soldering to aluminium dots using electrolytic etching of the oxide in a solder bath. [Williams 1957]

Raytheon’s first silicon transistor series were the CK790-793 series: they were all alloy junction PNP types produced from 1956 in the familiar Raytheon oval outline. In the Raytheon tradition of stand-out colors the silicon transistors were red. The tentative data sheet for the CK793 is dated January 1956 and the others followed through to April 1956. [Photos Courtesy of Joe Knight]

These transistors were expensive and intended for military applications (for example the CK793 was priced at $92 in the 1957 Federated catalog. [Ward 2002]

The mystery item pictured is the CK792 for which no data sheets appear to exist. The following data table is based on Raytheon data sheets dated 1956.

In 1957 Raytheon developed its 2N327-330 series encapsulated in the low profile TO-30 outline. These can be related to the CK70-793 series but both the 2N and CK series were subject to revised specifications indicating that Raytheon was still developing its production. For example, in January 1958 it down rated the maximum operating voltages of the series while doubling maximum collector current. It also increased the can dimensions marginally. In 1960 registration of the series was cancelled. [JEDEC 1960 Photo courtesy Joe Knight]

In the absence for data on the CK792 there is no evidence for aligning it to the 2N329 other than that it fits the sequence.

1954 was the year of the silicon transistor: Texas Instruments had announced a “commercial” silicon transistor and RCA, Bell, Hughes and Raytheon had all disclosed their progress and the fact that they had prototype silicon transistors.

Silicon Power Transistors

Raytheon developed TO-3 style silicon power transistors although they do not appear to have commercialised them. In 1958 John Williams filed on a method of creating a base connection to a silicon diffused transistor. [Williams 1958] These transistors might have base widths of only 0.1mil where a conventional connection was impossible. His patent described a highly doped ring (P-type in the case of a NPN transistor) which on alloying would diffuse down to the P-type diffused base layer.

The photograph [courtesy Joe Knight] shows a close up of a TO-3 cased transistor using an annular base tab of the kind described in the patent.

This photo shows several Raytheon developmental silicon NPN TO-3 type Power Transistors, from about 1957. [Photo courtesy of Joe Knight] They include a DEV16 and a QC155.

Raytheon also developed a silicon power transistor similar to the Texas instruments 2N389/424 device that Texas released in July 1957. [JEDEC 1958]

This is a development type in the TO-53 outline. [Photo courtesy Joe Knight]

While this transistor is unidentified, Raytheon is listed as a second source supplier of the 2N389 and 2N424. [Sams 1960 cited by Joe Knight]

Raytheon had in-licensed its germanium transistor technology and had shown considerable corporate energy coupled with significant risk taking in order to get its transistors to market. It could not duplicate this approach in silicon because it needed to rely on its own original research. Norm Krim recalls pushing Raytheon’s Newton production engineers and Waltham R&D scientists for a good diffusion silicon transistor. “We then had many military customers (including Raytheon) asking for this product because it was good at much higher temperatures.” Raytheon did not make the investment in research to sustain the urgency sought by its management:

“Dr. Ivan Getting, VP- Engineering and Research, Rhodes Scholar, lost a battle with our Executive VP, Harold Geneen, over the Research Budget. Dr. Getting's diffused Silicon Transistor Budget was cut severely by Messrs Adams, President and Harold Geneen...Dr. Hermann Statz, Solid State Physicist part of the team told me he was present and devastated because he was close to success developing a diffused silicon transistor.” [Krim 2009]

The company failed to show the leadership it had maintained in respect of alloy junction technologies. Raytheon could not develop viable silicon technologies and was outclassed by the mesa and planar technologies before it had a viable product.

Junction Diodes

Junction diodes developed in parallel with junction transistors using the same technologies. The three forms developed in the early 1950s were the gold bonded diode, the alloy junction diode and the grown junction diode. Raytheon worked on all three.

The gold bonded diode was invented by Bell Laboratories and most famously commercialised by Transitron as well as many other companies. It was made by a derivative of the forming process used in the production of point-contact transistors. A gold wire doped with gallium or indium was brought into contact with N-type germanium and alloyed using a current pulse forming a P-N junction. This gave a lower noise diode with greater current and reverse voltage capacity at the expense of reduced frequency response. Raytheon made these and also silicon bonded diodes (but not using gold alloy). This picture shows an early silicon bonded type, a CK738 in a non-standard outline.

The alloy junction diode predates the alloy junction transistor and was first made by Robert Hall at General Electric. These gave much larger junction cross sections and were suitable for high current operation. They were made in germanium or silicon.

Grown junction diodes, like their counterpart, the grown junction transistor were difficult to make and reserved for specialised needs: they were characterised by low forward impedance, lower saturation currents and higher peak inverse voltage. For example, Raytheon made developmental silicon grown junction diodes capable of operation at 2000 volts. [Finnegan 1955]

This picture shows the common diode outlines in use in the mid 1950s. Raytheon made gold bonded germanium diodes in outlines A and D. Silicon bonded diodes were made in outline D. Silicon power rectifiers (CK775 and CK776) were stud mounted.

The Spacistor

Raytheon expected that the Spacistor would be the defining contribution by Raytheon to semiconductor technology and restore the leadership it had enjoyed in the early years. The development was led by Dr Hermann Statz. Others in the team included Robert Pucel, Conrad Lanza and Hans Shenkel. It was the most significant basic research project undertaken by Raytheon in the field of semiconductors. Overall the program extended from 1954 to 1961.

Statz published related work first in 1956 [Statz 1956] on a three terminal device he called a spacistor that used avalanche multiplication to amplify. A year later Statz published on a spacistor tetrode [Statz 1957] and signalled remarkable claims for this device in his patent application. “It is possible to obtain adequate frequency response characteristics even at frequencies ranging into the microwave region.” [Statz 1957]

The breakthrough was well reported in the technical and general media. The New York Times carried a substantial account by Pulitzer Prize winner William Laurence in his column Science in Review for July 21 1957.

“The spacistor, the Raytheon scientists assert, promises two major advantages over today’s best transistors. It will amplify, they predict, at frequencies of up to 10,000 megacycles (10 billion cycles), as much as 50 times higher than transistors. Also, because spacistors can be made from materials unsuited to transistors, they are expected to operate at temperatures as high as 500 degrees Centigrade.”

“The spacistor was described as “a new kind of semi-conductor amplifier devised to overcome the frequency limitations of the transistor by avoiding the slow diffusion of charge carriers (namely electrons) through the base region.” The spacistor utilises “very much higher field strengths” to accelerate the charge carriers so that their transit time is greatly shortened. This makes possible operation at much higher frequencies.” [Laurence 1957]

Here was an extraordinary new technology that would put Raytheon on the map and secure its future in semiconductors. This pleased the semiconductor group: however, the company breakthrough for 1957 featured on the cover of its annual report that year was the Hawk missile.

Spatz’s device was a tetrode. The drawing [Radio TV and Hobbies 1957] shows what is clearly an experimental device since two of its four electrodes are un-terminated. It had a single PN junction between its main electrodes which was reversed biased with relatively high voltages (100-200 volts) creating an enlarged space charge region either side of the junction. The other two electrodes were the injector and the modulator positioned closely together and within the space charge region. The modulating contact was a small alloyed strongly P-type region which was also reverse biased. The injector could be a tungsten point-contact or an alloyed strongly N-type contact.

The input signal was applied to the modulator: “The first experimental devices show a low frequency power gain at least as high as that obtained with our present-day transistors.” [Statz 1957] Raytheon wanted to be clear that their device was not a “transistor” and did not use the customary nomenclature for its electrodes.

The invention of the diffused transistor and its introduction by RCA in 1956 (the 2N247) gave radically improved high frequency performance by creating a base region with an electric field that improved the transit time of injected carriers. The field arose from the base doping profile which increased exponentially from the emitter side to the collector side of the base.

Statz’s objective was to achieve fields much higher than those in diffused transistors and thereby faster transit of injected carriers. Statz wrote: “Though the base regions of diffused transistors have a built in field, the total voltage drop across this region is limited, in principle, to at most one-half the energy gap of the semiconductor material used, In practice, this upper limited is never reached and in most devices corresponds approximately to 0.1 or at most to 0.2 volt.” But in the Spacistor high fields could be created and at the same time the thickness of the space charge region could be limited. “Therefore we have attempted to make amplifying devices which utilize these high fields to obtain very short transit times for charge carriers.” [Statz 1957]

In 1961 Jerry Lavine summarised progress he and his team (Rindner, Nost and Nelson) had made in developing a viable spacistor. [Lavine 1961] It is clear from this review that no commercial product was in sight. The problem with the modulator was that its surface oriented effect was limited (compared to the base in a junction transistor) and that it was difficult to design a practical device that had effective control of the injector current. A further practical difficulty was that both the injector and modulator had to be positioned within the space charge region which needed to be as small as possible. At this time where point contact injectors were being used they needed micromanipulators to position them accurately. Alternative constructions tried used planar transistor techniques and evaporated aluminium contacts.

“The important question of the maxium useful frequency has not, of course, been answered by the observations reported here.” Work to characterise Raytheon’s laboratory prototypes was performed at frequencies up to 25 Mhz. The very high input and output impedances of the devices made high frequency designs difficult. “If analog structures are to assume a useful role in solid state electronics, it will be at relatively modest frequencies where the high impedences of the structures may be used to advantage.”

In fact the spacistor had never shown performance any better than the average diffusion transistors of the mid 1950s. By the time Lavine reviewed the program in 1961 silicon planar transistors were the established high performers.

Finale

At the close of the decade Raytheon was looking for a strategy that would secure for the company a viable position in semiconductors going into the 1960s. Since the 1940s the success of the company had been in research and development of new technologies for the military. But this sector could not be relied on to sustain ongoing growth in a peace-time era. Raytheon weighed the prospect of commercial or non-military markets and committed itself to them. In 1959 Richard Krafve, Group Vice President, agreed with his executive team that Raytheon should enter volume production of transistors for commercial markets. Building of a new large scale dedicated facility at Lewiston Maine commenced. In 1960 Krafve became President and his entrancement with the future of semiconductors had greater license.

On the dawn of the 1960s some of the early entrants in the field of semiconductors realised that they were in no position to compete with the new silicon planar era and integrated circuits. Tran-Sil, Rheem and CBS put their semiconductor businesses up for sale at significant discounts and Raytheon took the bait. In 1961 it bought Tran-Sil’s silicon rectifier business for $1.25M in order to obtain a puny 1.3% of that market. In September that year it bought CBS’s Lowell Massachusetts’s silicon and circuit module plant.

Rheem Semiconductors had just lost its battle with Fairchild over its unlicensed development of a 2N697 silicon mesa transistor using Fairchild’s intellectual property. The settlement forced Rheem to withdraw from this market and exit its only immediate hope for a profitable future. Raytheon wanted to join Silicon Valley so took the opportunity to buy Rheem Semiconductor Corporation and followed that with the purchase of another CBS plant: this time at Danvess, Massachusetts.

These companies had good reason to sell. Raytheon did not recognise that they had not kept pace with the development of semiconductors and were about to be enveloped in a wave of new technologies from young start-ups. As though to presage the doom of Raytheon’s semiconductor business, Norman Krim who launched Raytheon’s transistor business over ten years previously, found that when the company corporate headquarters was re-located to Lexington he was not included in its plans. Krim resigned and went to Radio Shack.

“Inspired by the great success of Texas Instruments Company, many firms had created new facilities and entered the field. The results were soon to become a classic of overcapacity, accompanied by rapidly falling prices in a fiercely competitive market. It was immediately prior to this stage that the electronics experts inside CBS and the Rheem Manufacturing Company longed to abandon the race and that Raytheon appeared as a purchaser. The experts had led Krafve toward two horses already slated to be scratched, though the reasons of their owners had not yet become generally discernible.” [Scott 1974]

In 1962 the company found that none of its commercial businesses were making money. Krafve had left after losing a Boardroom battle and Tom Philips had become Executive Vice-President. He had no affection for haemorrhaging businesses: “We stayed with germanium semiconductors too long. On the other hand, silicon was becoming an important material in planar technology.” [Scott 1974] In February 1962 Philips recommended the orderly withdrawal of Raytheon from semiconductors and tubes.

In 1964 Zenith announced the first hearing aid to use an integrated circuit, something it had had in development since 1961, soon after the first integrated circuits became commercial. [Sinclair 1964] Just as discrete transistors had revolutionised hearing aid design by displacing tubes, some ten years later these same discrete transistors were redundant. The signals were there to see from 1958 but Raytheon had been unable to respond to them. Initially integrated circuits were difficult to make in any yield and were far better suited to digital logic than linear amplifiers. This isolated Raytheon from the rapid attrition of its original market until it had closed down its semiconductor business.

Here are examples of waves of disruptive technologies with radical impacts on manufacturers and in turn their customers on a mere ten year cycle: In 1939 miniature tubes were introduced making the concept of a portable hearing aid viable for the first time. In 1953 transistors were rapidly displacing tubes with huge benefits as to portability and battery usage. Eleven years later hearing aids would transit from something the size of a packet of cigarettes to behind the ear technology thanks to integrated circuitry.

In this brief time span Raytheon became a star and undisputed leader in the battle for the market and then withdrew from semiconductors; one outcome being as improbable as the other.

Disruptive technologies such as semiconductors created remarkable opportunities and profound difficulties for manufacturers and stakeholders in their industry. Innovation this fast meant that investments in research and plant became redundant before the technology had a chance to mature. Intellectual capital was lost. In semiconductors the final advent of maturity meant vastly improved production technologies leading to savage price competition and price reductions by unprecedented orders of magnitude. And these issues were potent internationally: significant on the world stage to the extent that international leadership was changing hands.

Lynn points out that while the transistor was invented in the United States, by 1960 Japan was producing more transistors than the United States. But the United States surged ahead as silicon replaced germanium in discrete devices and integrated circuits (IC) became important. The first IC calculator was developed in the United States but within a few years Japan controlled markets for this and many other solid state consumer products. By the mid 1980s Japan was again ahead on semiconductor production but in the 1990s it lost ground to Taiwan and Korea for commodity semiconductors and the United States for sophisticated new products. [Lynn 2000]

But within the major cycles of innovation such as the transition from discrete transistors to ICs technology chaos reigned as companies leapfrogged one another with new developments resulting in dead-ends for some and spectacular kudos for others. Technology breakthroughs were announced at scientific meetings straight off the lab bench and prior to any production development. Texas Instruments chose to announce that their silicon grown junction transistor was in “commercial” production on the strength of pilot quantities. Bell sought to represent their point-contact transistor as being ready for commercial applications: something never achieved. Philco found their delicately etched surface barrier transistors were being challenged by diffusion technologies before Philco could make their own product reliably.

An example of the changing lead on technologies is shown in the attached table that plots the development of radio frequency transistors over time:

Raytheon endeavoured to find its pathway with the virtue of persistence. Although germanium alloy junction transistors were technically obsolete by the end of the 1950s they were still being produced in huge volumes (but low price) all through the 1960s. But hindsight indicates Raytheon invested too much energy in their silicon alloy junction equivalent. It could not have anticipated the rapid successes of the Silicon Valley start-up that produced the planar transistor. Investing in the spacistor with the promise of 10Ghz performance was clearly something strongly encouraged by Raytheon’s success in microwave, radar and guidance systems. The spacistor never delivered and was abandoned around the time Sylvania broke the 1Ghz barrier with a silicon planar transistor.

References

Aspray W 1991 Virginia Powell Strong, an oral history conducted in 1991 by William Aspray, IEEE History Center, Rutgers University, New Brunswick, NJ, USA.

Armstrong D Miniature Transistor Hearing Aid Radio-Electronics Magazine - December, 1953 40-1 Reproduced by R McGarrah at

http://transistorhistory.50webs.com/maico.html

Bardeen J Brattain W 1948 The Transistor, A Semiconductor Triode Phys Rev 74 230-31

Bell 1952 Transistor Technology Vol 1 and 2 July 1952

Blais M US Patent 2882464 Transistor Assemblies Continuation of Serial 324094 filed December 4 1952 patented April 14 1959

Braun E, Macdonald S 1982 Revolution in miniature: the history and impact of semiconductor electronics re-explored in an updated and revised second edition Cambridge University Press, 1982

Carroll J 1957 Transistor Circuits and Applications McGraw Hill Book Company Inc New York

Davidson R 2007 1955 Raytheon 8TP Series

http://www.geocities.com/soho/atrium/1031/trans/US/1azRay8TP.html

Ducat 1953 Radio Electronics Engineering magazine for September 1953 Cited by Bob McGarrah http://transistorhistory.50webs.com/xstrhist.html

Earls A 2005 Raytheon Company: The First Sixty Years Arcadia Publishing

Electronics 1954 Grown Silicon Transistors Appear Electronics 27, March p8

Finnegan F 1955 Junction Diodes – Features and Applications IRE Trans Electron Devices 2 51-62

Fortune Magazine 1953 The Year of the Transistor March 1953

Freedman G 1955 US Patent 2948836 Electrode Connections to Semiconductive Bodies Filed 30^th March 1955 granted 9^th August 1960

Freedman G 1956 US Patent 2916408 Fabrication of Junction Transistors filed March 29 1956 patented December 8 1959

Getting I 1989 All in a Lifetime: Science in the Defense of Democracy Vantage Press New York

Goldstein A 1991 Gordon K. Teal, Electrical Engineer, an oral history conducted in 1991 by Andrew Goldstein, IEEE History Center, Rutgers University, New Brunswick, NJ, USA.

Goldstein H 2003 The Irresistible Transistor Spectrum, IEEE 40, 3 42-7 http://www.spectrum.ieee.org/print/5638

Hall R 1952 P-N Junctions Produced by Growth Rate Variation Phys Rev 88 139

JEDEC 1956 Joint Electron Device Engineering Council Release 1671 Filed June 25 1956

JEDEC 1958 Joint Electron Tube Engineering Council Release 2277 September 22 1958

JEDEC 1960 Joint Electron Device Engineering Council Release 1960-1960B First filed July 1 1957

Knight J 2008 The First RCA Experimental, Developmental and Production Transistors Tube Collector 10 No 4 7-11 Also on this site RCA Developmental Transistors

Krim N 2009 email to J Ward with comments on this article September and December 2009

Kurshan J et al 1954 Report on the Conference of Semiconductor Devices Research Minneapolis MN June 28-30 1954 RCA Technical Report PEM-452

Laurence W 1957 New Electronic Device, the ‘Spacistor,’ is Introduced as a Revolutionary Discovery New York Times July 21 1957

Lavine J Rindner W Nost B Nelson R 1961 The Spacistor IRE Trans on Electron Devices 8, 4 252-64

Levy I 1953 US Patent 2827598 Method of Encasing a Transistor and Structure Thereof filed March 19 1953 patented March 18 1958

Lynn L 2000 Technology Competition Policies and the Semiconductor Industries of Japan and the United States: A Fifty-Year Retrospective IEEE Trans Engineering Management, 47, 2 200-210

McGarrah R 8TP Series Transistor Radio (1955)

http://transistorhistory.50webs.com/ray8tp.html

McGarrah R Transistor History 101 http://transistorhistory.50webs.com/xstrhist.html

McGarrah R 1999 Interview with Frank Ducat

http://transistorhistory.50webs.com/xstrhist.html

McGarrah Zenith Royal 500 Seven Transistor Radio (1955)

http://transistorhistory.50webs.com/royal500.html

Melliar-Smith C 1998 The Transistor: An Invention Becomes Big Business Proc IEEE 86 1 86-110

Nebeker 1995 Ivan Getting An Interview Conducted by Frederik Nebeker Center for the History of Electrical Engineering February 25, 1995

Nelson H 1954 A Silicon NPN Junction Transistor by the Alloy Process Transistors 1 Radio Corporation of America 1956 172-81

Nowak H 1954 US Patent 2846626 Junction Transistors and Methods of Forming Them filed July 28 1954 patented August 5 1958

Parziale E 1956 US Patent 2955242 Hermetically Sealed Power Transistors filed November 27 1956 patented October 4 1960

Powers D 1951 The Present Status of Transistors and Transistor Applications Internal memo October 1, 1951 Bell Telephone Laboratories cited by Bob McGarrah

http://transistorhistory.50webs.com/weco1951.html

Radio TV and Hobbies 1957 Spacistor-New Semi-Conductor December 1957

Radiomuseum http://www.radiomuseum.org/

Raytheon 1948 Germanium Crystal Triode CK703 Data Sheet, Special Tubes Division, Revision One November 3 1948

Raytheon 1952 Germanium Transistor CK716 Data Sheet, Special Tubes Division, Revision Three February 8 1952

Raytheon 1954 Raytheon Industrial Tube Characteristics Raytheon Manufacturing Company

Raytheon 1956 Technical Information CK793 Tentative Data January 3 1956

Raytheon 1956 Technical Information CK790 Tentative Data April 2 1956

Raytheon 1956 Technical Information CK791 Tentative Data April 2 1956

Raytheon News April 1953 Reproduced in Earls A 2005

Raytheon 1955 Transistor Applications Volume I Raytheon Manufacturing Company

Raytheon 1957 Transistor Applications Volume II Raytheon Manufacturing Company

Riordan M, Hoddeson L 1997 Crystal Fire W W Norton & Company

RTMA 1951 Radio Television Manufacturers Association Release 964 April 18 1951

Saby J 1952 Fused Impurity PNP Junction Transistors Proc IRE 40 1358-60

Sams 1960 Transistor Substitution Handbook 3rd Ed Howard W Sams & Company

Sardella J 1953 US Patent 2720617 Transistor Packages filed November 2 1953 patented October 11 1955

Sardella J 1955 US Patent 2888736 Transistor Packages filed March 31 1955 patented June 2 1959

Scott O 1974 The Creative Ordeal: The Story of Raytheon Atheneum, New York: 1974

Sheckler A 2004 The Beginning of the Semiconductor World; General Electric’s Role Proc IEEE Conference on the History of Electronics

Shockely 1948 US Patent 2569347 issued September 25th 1951 2569347 filed June 26th 1948

Shockley W Sparks M Teal G 1951 P-N Junction Transistors Phys Rev 83 151-62

Shockely W 1976 The Path to the Conception of the Junction Transistor IEEE Trans on Electron Devices ED-23 July 597-620

Sinclair J Druz W 1964 A Microlithic Hearing Aid Amplifier IEEE Trans on Audio 12 6 118-20

Spanos J 1956 US Patent 2982893 Electrical Connections to Semiconductor Bodies filed 16^th November 1956 patented 2^nd May 1961

Stokes J W 1982 70 Years of Radio Tubes and Valves Vestal Press Ltd New York 1982

Statz H Pucel R 1956 The Spacistor, A New Class of High Frequency Semiconductor Devices Proc IRE 45 3 317-24

Statz H Pucel R Lanza C 1957 High-Frequency Semiconductor Spacistor Tetrodes Proc IRE 45 11 1475-83

Statz H Pucel R 1957 US Patent 3192400 Semiconductor Devices Utilizing Injection of Current Carriers into Space Charge Regions filed July 15 1957 as Serial 672046 and patented on June 29 1965

Tanebaum M 2008 First-Hand: Beginning of the Silicon Age IEEE Global History Network August 2008

http://www.ieeeghn.org/wiki/index.php/First-Hand:Beginning_of_the_Silicon_Age

Tilton J 1971 International diffusion of technology: The case of semiconductors. Washington, DC: Brookings Institution.

Torrey 1948 H Whitmer C Crystal Rectifiers 15 MIT Radiation Laboratory Series. McGraw-Hill, New York, 1948

Toute la Radio 1954 Revue Critique de la Presse Mondiale: Transistors Jonction au Silicium 186 June 1954

Ward J 1997 email from F Ducat http://www.ck722museum.com/page12.html

Ward J 1998 Email from N Krim http://www.ck722museum.com/page12.html

Ward J 2001 The Story of the CK722 1^st ed

Ward J 2002 Raytheon CK793

http://semiconductormuseum.com/PhotoGallery/PhotoGallery_CK793.htm

Ward J 2003 The CK722 Classic Germanium Transistor Museum

http://www.ck722museum.com/page6.html

Ward J 2008 Raytheon CK705 Germanium Crystal Diode

http://semiconductormuseum.com/PhotoGallery/PhotoGallery_Raytheon_CK705_GermaniumDiode.htm

Williams J 1953 US Patent 2862470 Transistor Mold Assemblies filed November 19 1953 patented December 2 1958

Williams J US Patent 3089219 Transistor Assembly and Method filed October 19 1953 patented May 14 1963

Williams J 1957 US Patent 2857321 Method of Soldering to Aluminium or Other Material having Surface Oxide filed 15^th March 1957 patented October 21^st 1958

Wolsky S 1954 US Patent 2879457 Ohmic Semiconductor Contact filed 28^th October 1954 patented 24^th March 1959

Wolff M 1984 Norman B. Krim, an oral history conducted in 1984 by Michael Wolff, IEEE History Center, Rutgers University, New Brunswick, NJ, USA.

Wolf M 1985 Roger Webster, an oral history conducted in 1985 by Michael Wolff, IEEE History Center, New Brunswick, NJ, USA