High Purity Germanium

Several companies have been involved in growing or trying to grow high purity germanium, but few have had success at it, notably Canberra, ORTEC, and Umicore. In recent years there have been diverse attempts by some involved in underground scientific research to first: learn how to grow good HPGe crystals, and second: build and staff an underground production facility. All this is intended to reduce the cosmogenic activation of the crystals whilst in process. Few if any of the people who ever succeeded in growing good Ge crystals above ground have supported or involved themselves in these efforts.

A brief summary of HPGe crystal growing endeavors follows.

General Electric

The lithium-drifting or compensation process was necessary only because germanium was not available with sufficient purity to make thick or big detectors. The net electrically active impurity concentration must be in the order of 1X10¹⁰ atoms/cc to deplete a few centimeters of germanium with reasonable (say 5kV) reverse bias . The impurity level in drift-grade germanium is about 10¹³ atoms/cc which is, after all, 99.99999% pure was about as good as germanium got before the work of Bob Hall at General Electric’s Research Laboratory in Schenectady, N.Y.. Hall perfected zone refining and Zcrochalsky crystal growth techniques which yielded small crystals having 10¹⁰ to 10¹¹ impurity atoms/cc in the early 1970s. By the mid-1970s GE had created a commercial product line in the Space Technology Products Division near Philadelphia and Dr. Peter Hewka set about to make HPGe crystals for the detector industry. 

The first HPGe crystals sold by GE were suitable for planar detectors only because they were small and Ge(Li) detectors had become relatively large volume devices by that time. The material sold for about $20/gm which was about 10 times the going rate for drift-grade material. PGT was the first company to use HPGe on a large scale. They employed an Israeli scientist, Micha Harchol, who is credited with developing processes for the production of both planar and coaxial devices. PGT production soared at this time because ORTEC and Canberra were both behind in device technology.

The GE HPGe business was always unpredictable- except for the “summer problem” when they routinely lost control of purity every year during the summer months. Apparently this problem was never solved.  Good GE Germanium exhibited great charge collection although it was not better that that of good Ge(Li) detectors. This must have been disappointing to some but more disappointing was the appearance of serious charge collection problems due to Cu-related defects and dislocations (crystallography).

ORTEC

ORTEC had a very successful drift-grade crystal growing operation which began in 1969. By the mid-1970s they working on HPGe but according to Rex Trammel, who headed the effort, they did not succeed in the first few years. Dr. Larry Darken, who earned a Ph.D in solid state physics at Notre Dame in 1973, was hired by ORTEC in 1974 to work on the development of HPGe. According to Larry, ORTEC had been trying to grow HPGe crystals in their Ge(Li) pullers but the impurity levels were 2-3 orders of magnitude too high. By 1974 Bob Hanson at LBL, had some success in growing HPGe crystals of up to 4 cm in diameter. Larry visited LBL under their liberal technology transfer policy to learn more about their program. ORTEC management expected that this infusion of knowhow would result in significant progress within two years or less, but good results did not appear until 1977 after a new puller was commissioned and alterations to the LBL processes were implemented. By 1979 ORTEC considered the process to be stable and Larry joined the Solid State Division of ORNL, remaining a consultant to ORTEC. In 1981 ORTEC developed boron implanted contacts for Ge detectors. They had after all, designed and built an Ion Implanter and even sold a couple of them. This was one of the several technological achievements that ORTEC failed to pursue for one reason or another in those years. Apparently they believed that these contacts would put them in such an unchallengeable position in the detector business, that they did not need the consulting services from Larry anymore.  Ironically Canberra found that making detectors using ion implantation to be one of the easier challenges we ever faced.

Princeton Gamma-Tech

PGT began a low-budget HPGe crystal growing effort around 1974. Joe Baicker boasted that they would do this on a shoe-string budget relative to others. They had no experience in drift-grade material but figured to avoid that stage completely. They eventually grew some drift-grade material when HPGe did not materialize and continued this effort until 1989. In the last few years they had Peter Hewka on the staff and they grew some good HPGe crystals but were never competitive.

Tennelec

In 1982 GE sold the HPGe business to The Nucleus/Tennelec  which was owned by David Coffey. GE had refused to negotiate with Canberra because they said they did not want the business to fall into the hands of a detector manufacturer which might lead to denial of material to others. We got a call from Jim Moughianis, who ran Tennelec, assuring us that they would honor this agreement, and they did so if for a short while. Tennelec had earlier bought John Walter’s little silicon surface barrier detector business, Waltec, and John again found himself in the HPGe business. The business was moved to Oak Ridge, and Peter Hewka stayed around for a while but he did not want to move to Oak Ridge. Eventually he left Tennelec and joined PGT. In 1983 Tennelec turned to Larry Darken to resurrect the operation.

Tennelec had serious quality problems in their first years. Canberra used a significant amount on Tennelec material in 1984 and 1985 but much of it had to be returned for credit. This led to an argument with Jim Moughianis, their fiery President, and brought the relationship to a standstill for more than a year. In 1987 and 1988 Tennelec quality improved dramatically and we started buying material again. This eventually led to the production of “Big Boy”, the world’s first 100% Ge Detector which was finished on October 4, 1988. Collaboration on Big Boy began as a challenge to ORTEC, who announced months earlier that they were putting on a contest for customers to guess when and by whom the first 100% Ge detector would be made. They promised to announce the winner of the contest and the upcoming IEEE-NSS Conference. Clearly they believed that they would be the producer of this detector but Orren and Larry made other plans. In record time, Tennelec  grew the big crystal and Canberra quickly turned it into a detector just a week or two before the NSS. Rather than waiting, we sent out an announcement of this accomplishment to some 5000 customers on our mail list days before the NSS. Big Boy measured some 114% in Relative Efficiency with energy resolution of around 2.1 KeV(FWHM). Tennelec’s fiscal performance was outstanding in 1988 at which time they were sold to UEI, the British parent of Link Systems. It was at this time that quality again hit the skids continuing through most of 1989.

In the early 1990s Tennelec approached Canberra with a proposal to form a joint venture to capitalize on HPGe crystal growing. They offered Canberra a 50% partnership in return for an exclusive supply agreement and a modest investment to expand the facility.  We turned down this offer but Tennelec gradually won more and more of the Canberra business. Finally we went to MHO, knowing that we were their biggest customer by far, and asked for terms commensurate with our importance as their dominant customer. They refused to understand the seriousness of our demands so, in 1995, we made some such agreement with Tennelec and they became our major supplier. This worked out quite well for both companies and continued until we bought Tennelec from their new owner, Oxford Instruments, in 1999.

We did not buy the Tennelec  building, however.  Oxford was asking far too much for this property under the assumption that we would not have the temerity to move such a critical operation as crystal growing. After much thought, however, we decided to take the risk and bought a nicer building just across town at a fraction of the price Oxford was asking. We mounted a campaign to build finished inventory and within a year we were established in the new facility with hardly any interruption in crystal supply.

Metallurgie-Hoboken Overpelt (MHO) now Umicore

MHO’s effort to develop HPGe was headed up by Dr. Walter Schoenmaekers. As always MHO took an industrial approach to the problem and put a lot of engineering effort and facilities expense to work on it. Progress was slow. However, and Walter left to join Canberra at a time when production and quality were still sporadic. Later, under the direction of Marc Van Sande, MHO began to do DLTS measurements on every crystal to check for copper levels. This led to a very high acceptance rate (>95%) on shipped crystals. MHO eventually bought two huge crystal pullers (from Ferromagnetics) and this led them into the forefront in crystal growing, even if their customer base was reduced by the defection of ORTEC and Canberra.

Before MHO began DLTS measurements and quality improved, a large percentage of crystals were returned for credit. Orren Tench and Emery Olcott visited the plant to discuss this along with pricing issues. After the obligatory grand tour and a meeting with Walter Schoenmaekers and some of his associates we went to lunch at the executive dining room where we met with Mr. Bossuyt who was the plant manager, the first time we had met him. Bossuyt described HPGe as “stuff we ship to customers only to have it shipped back to us”. Early in this luncheon I explained our need for greater quantities, better quality, and lower prices to enable us to better compete with ORTEC. His reaction, to say the least was emphatic. Slamming his hand down on the table with a mighty force, he said “If you expect lower prices, the answer is Nyet, Nyet, Nyet!”. Bossuyt’s associates almost went into shock at this display of emotion and, after finishing a rather tense meal, Emery and Orren left with little hope that MHO would be a reliable, competitive supplier in future. This, perhaps more than any other reason, made us receptive to an investment in HPGe crystal growing when it came along later.

Canberra Semiconductor nee Berkeley Germanium (BGC)

Around 1980 we had a visit from and odd couple from California who were looking for customers and personal investors for a new HPGe crystal growing operation to be known as Berkeley Germanium Corporation. Georg Mraz, the guru, was a Czech scientist who came to LBL as a visiting scientist and then defected. His partner, Mary Williams, who for some strange reason thought it was important for me to know that she and Georg were sharing a hotel room on this trip, was in real estate in California and had arranged the financing for this venture. We gave them encouragement but declined to invest in the company. Predictably, they ran out of money in about a year and we again had a chance to enter the picture.

This photograph of a crystal being grown was taken at CSI around 1982

This time we took over the company leaving the investors with a royalty deal that ratcheted down with time and with additional investment. Of course the guru of Canberra mergers and acquisitions, Emery Olcott put this deal together. Although it was far from the most complex Olcott deal we ever saw, the added elements of steamy personal relationships, the assortment of investors, and the personalities of the part time and full time work force, combined to make it very interesting. One of the people who gave credibility to the effort was Scott Hubbard, also from LBL, who had been on board first part time and then full time. The fact that several scientist from LBL had invested in BGC gave us some confidence in Mraz, who otherwise might not have elicited much confidence, if for his appearance alone.

Mraz was a tall thin man with a strong accent, a low voice, jet black hair, and a goatee-like beard. The striking facial feature was one stark-white eyebrow on his dark complexion. It was difficult to look him in the face while carrying on a conversation without cracking up. We never got an explanation from George himself, but Scott Hubbard much later told me that Georg claimed that his eyebrow had turned white abruptly after a frightening experience (nightmare?). Perhaps we should have heard this story before we took over BGC- to avoid a frightening experience of our own.

Two of the LBL investors, Don Landis, the well-known analog circuit designer, and Norm Madden, a detector systems expert of some repute, were both involved in helping BGC technically. Madden was, at one time, a part time if not full time employee of the company. Another investor, Phil Wood, who had a finance background, was working there and Canberra kept him on part time for a period to keep the books. Phil’s father, incidentally, was the founder and owner of Troybilt, maker of the famous garden tiller. 

Besides all the technical problems and personnel problems, some people at LBL who had no vested interest in the venture were miffed at BGC for using the Berkeley name. We fixed this by re-naming the company, Canberra Semiconductor, Inc.(CSI) and putting Scott Hubbard at the helm. Scott cultivated relations with LBL and for a time the people in Fred Goulding’s group began to treat Canberra with something less than disdain- quite a new experience for us.

The novel initiative at BGC had been the purchase (lease) of a custom-designed crystal puller from Leybold-Heraeus. Before everybody had used homemade pullers or converted old silicon pullers to grow Zcrochalsky Ge crystals. The L-H machine cost about $200k, an enormous at the time and the industry was eager to see what it could do. By the time Canberra took over any early euphoria was long over. Everyone sat back to try to see “the forest from the trees” and it became clear that miracles were not happening. We had to fix up the facilities and the process both of which had been thrown together in absolute haste. This took time and money. Some of the money went to Mraz in the form of a third mortgage on a house to live in so he could move his bed from the CSI premises.  Yes this was a 24-7 job for Mraz before Canberra arrived. By some miracle Georg eventually pain off this mortgage in full.

After resuming operations with Scott and GM and Georg as Production Manager the next three years were like a soap opera, with optimism and disappointment alternating almost on a day-to-day basis. Crystals did get grown, of course, and many passed inspection for physical perfection and impurity concentration. These were sent to Meriden, for processing into detectors where we spent an enormous effort to make detectors from them. The net result of all this is that we made and sold some of the worst HPGe detectors in history. Our specialty was cheap, and it can be said that our customers got what they paid for. What an impact this must have had on our reputation, however!

This Ge crystal from Canberra Semiconductor (CSI) was unusable (impurity concentration?) but looked pretty and was used for display.

As an example of the extremes we endured to move these detectors, we had found that we could check for radio-iodine at nuclear plants with good NaI(Tl) detector or with bad Ge detectors. Ivor Priess, at RPI, was engaged by our Systems Division to do the applications work for a Radio-Iodine Analyzer based on the S-10 portable MCA and a portable Ge detector. We were fighting ORTEC for some systems at Duke Power long before our unit was ready when John Smalling, our Sales Engineer there, convinced management that ORTEC was offering portable Ge detectors for $1900 each. We eventually got the order at a price not much higher than this. Later we found out that $1900 was ORTEC’s price for the portable cryostat option alone. Eventually this customer was soured by the inappropriate application of these Ge detectors for emergency response and no doubt we paid doubly for this venture in lost business later.

After quite some time without much success, we asked Walter Schoenmaekers, who had joined Canberra to make Ge detectors for the European market, to visit CSI and give us the benefit of his experience and knowhow in Ge crystal growing. After a day or so at CSI, his advice: “Start over!

Regardless of our effort to reach stable, repeatable results, our process was never in control and we did not maintain decent quality or production rates. Finally, in 1985, we closed CSI and sold off the equipment. The bottom line losses in this venture amounted to about $2,000,000. The detector business made enough profit during this period to absorb these losses.