The Early History of Gallium Nitride Research
Source/Type: White Papers/Technical Papers

Author: Edward A. Miller -- RCA Laboratories

January 31, 2005... In 1970 I was working at the David Sarnoff Research Center in Princeton, New Jersey. These Laboratories were named previously the R.C.A. Laboratories. I was a Research Chemist and my area of work, at the time, involved materials development, with novel and potentially valuable materials for developing light emitting diodes (LEDs) and solid state lasers.

Around September 1970 I was asked by my group head supervisor, David Richman, if I would take on the task of growing Gallium Nitride (GaN) for Dr. Jacques I Pankove, who was a staff member at the laboratories at the time. They needed someone to take over the work from Herb Maruska who had gone back to Graduate School. The semiconductor Gallium Nitride was a material that was of interest at the time because of its high band gap with the potential for developing a blue LED or laser. Dr. James J Tietjen, who was our Lab Director, envisioned using the GaN blue LEDs to make a flat panel television display, one which could be hung on the wall like a picture. The project that I would begin to work on was the gas phase growth of GaN as previously reported by Maruska and Tietjen.[1] Dr. Tietjen had started this project at RCA Labs in May 1968.

This program consisted of growing GaN, which was then a candidate semiconductor for blue LEDs or lasers because of its high band gap at about 3.4 eV, and which would require making p-type material. This blue emitter was greatly sought after because the other primary colors, red and green, had already been established with GaAsP and GaP:N. Blue was the one primary color that was missing an emitter material in order to be used in future full color flat panel displays.

Previous work at RCA Labs had established the method of growth by Vapor Phase Epitaxy (VPE) and had made some initial investigations of doping. The work entailed the growth of GaN in a horizontal quartz system originally designed for the growth of GaAs, using chlorides to transport vapors containing Ga, and a hydride gas (arsine) to transport the Group-V element. Zinc for p-type doping was transported as its own vapor from a vertical sidearm tube. To grow GaN, the arsine source was replaced with ammonia.

The VPE growth system consisted of a complex of quartz and pyrex tubes. The upstream section consisted of two parallel horizontal quartz tubes each about two feet long and one inch in diameter. They were joined to a single quartz tube, two inches in diameter and about two feet long. Films were deposited in the upstream end of this two inch diameter tube. In one of the narrower tubes, gallium metal (a liquid) would be placed in a quartz boat. At the entrance to this tube was a ground quartz joint coated with halocarbon stopcock grease, and through this joint hydrogen chloride gas and hydrogen were allowed to flow over the gallium. This reaction would form gallium chloride, a gas that can be transported by the hydrogen flow into the reaction zone. The other narrow tube also began with a ground joint, and served to admit gaseous anhydrous ammonia to the growth chamber, also with hydrogen as a carrier gas. When the two components ammonia and gallium chloride mixed in the growth chamber, GaN was formed.

Downstream of the growth zone, the two inch tube divided into two perpendicular tubes, with the original tube terminating at a large pyrex stopcock. The other sidearm served to conduct the effluent gases to the exhaust system. The stopcock created a loadlock. Beyond the stopcock, which was also lubricated with halocarbon stopcock grease, was a pyrex forechamber. This allowed samples to be inserted and removed from the growth chamber without opening that part of the reactor to the ambient. The forechamber received the substrate placed on a quartz sled attached to a true-bore quartz rod. The rod was positioned through a true-bore pyrex bearing that was filled with flowing inert gas. The rod could be inserted and withdrawn through the bearing.

The procedure was to place a sapphire substrate upon the sled at the front end of the true-bore rod. The stopcock was of course closed at this time. The true-bore gas bearing featured a ground glass joint which was then affixed to the forechamber. The forechamber would then be flushed with inert gas to remove all traces of air. If desired, it could also be flushed with hydrogen. Meanwhile, hydrogen was also flowing through the main growth tube.

The growth system was contained in a furnace with multiple zones. Each zone had an independent temperature control. Typically there were two zones for the dual one inch diameter inlet tubes, followed by a short mixing zone, and then the growth zone. There was also a sidearm for doping coming up out of the mixing zone, with its own temperature control. In the sidearm was located a small quartz bucket loaded with zinc metal. The bucket was held on the end of another true-bore quartz rod, which allowed the position of the bucket to be moved within the sidearm furnace. The zinc vapor and hence the delivery rate for zinc was determined by the temperature of the bucket. At first it was necessary to establish a calibration graph for zinc concentration in the grown film versus temperature. The zinc would be evaporated in a flow of hydrogen gas.

As the growth system came on line, the products of GaN both undoped and doped with zinc would be passed on to Dr. Pankove for his evaluation. It was only a short time following my earliest growth run that the exciting news came over the phone from Dr. Pankove that he had observed for the first time electroluminescence in GaN.[2]

This news caused a modest amount of excitement in the Laboratories and I particularly recall the stimulating urgency of the work I was involved with. Now each new growth run was made to try to modify the temperature or a flow rate or some other variable to optimize a particular property of the GaN product. As time went by, new results were coming out of Dr. Pankove’s measurements and evaluations of the devices that he made from the GaN crystal films that I supplied. This was a very exciting time indeed!

Thick films of GaN were required to obtain free standing GaN wafers without the substrate attached for evaluation of the effect of stress in the film. On the thicker films an uneven thickness was obtained across the wafer. A quick room temperature test was designed and used to reveal the gas flow patterns within the growth tube. This was accomplished with the furnaces open to permit observations. By admitting both HCl and ammonia simultaneously into the growth tube (in the absense of any liquid gallium) at the typical flow rates for a run, a flow of white smoke was generated which showed us the gas flow patterns.

These tests established a need for better positioning of the substrate facing the gas flow. When this was accomplished by having the substrate facing the flow at a sloping angle, the growth of the thicker films became more uniform in thickness and satisfactory for testing the diodes.[3]

This period of time in 1971 and 1972 gave me what everyone who is involved with basic research would dream of, ie, a hectic, exciting productive effort to advance the frontier of science.

As 1973 arrived, new advances continued to be made with the GaN material.[4,5] Maruska and Pankove introduced Mg doping.[6] But each new advance became longer to achieve, as with any project as it matures, and obtaining a really bright blue LED became a harder goal to obtain.[7] The tide of interest began to wane at the Laboratories, and I had the feeling that we had better do something soon or support would dry up for GaN. While new results were being reported such as Li and Be doping,[8] the highly sought after quest for p-type doping was not accomplished.[9] But it seemed to us that if enough time were permitted, results would be forth coming.

I envied Bell Labs at this time because I had heard that they had a group of about 10 staff researchers working to make p-type GaN. I was just waiting for the "big shoe" to drop and read that they were successful. If only we could have enough time to do all of the various tests that we were thinking of, I was sure that we would succeed in beating them to the elusive goal.

It was at this time, in the summer of 1973, that I attended a Gordon Conference in New Hampshire featuring nitride research. At this meeting I was at a table one evening with the head of the Bell Labs Materials Group, and I did not hesitate to ask him about the GaN work at his laboratory. I was astonished at his answer. He said that they were shutting down the work on GaN and moving on to other more promising work, because they were unable to make p-type GaN. Well, I thought, so be it. Except for the fact that Dr. David Richman, my group head at the time, was also at the table and heard what was said, I do not think that I would have mentioned it again for some time.

As it turned out, our nitride work continued for about another 6 months. Herb Maruska returned from graduate school at the end of the year, expecting to rejoin the project. Then in January 1974, Dr. Tietjen called us into his office and explained that spending money on the GaN work with no results that could lead to a commercial product in the near term was exhausting his budget. "You guys have bled me dry with nothing useful to show for it," he said. Then he phased out the project. What else could he do?

A scientist is fortunate if at some time in his or her career, an exciting project comes along and the adrenalin rises. I had such a time with the GaN project and I will always remember it as a good time.
References

H. P. Maruska and J. J. Tietjen, "The preparation and properties of vapor-deposited single crystal GaN," Appl. Phys. Lett., 15, 327 (1969).
2. J. I. Pankove, E. A. Miller, D. Richman, J. E. Berkeyheiser, "Electroluminescence in GaN," J. Luminescence, 4, 63 (July 1971).
3. J. I. Pankove, E. A. Miller, J. E. Berkeyheiser, "GaN yellow light-emitting diodes," J. Luminescence, 6, 54 (Jan 1973).
4. J. I. Pankove, E. A. Miller, J. E. Berkeyheiser, "GaN electroluminescent diodes," RCA Review, 32, 383 (Sept 1971).
5. J. I. Pankove, E. A. Miller, J. E. Berkeyheiser, "GaN blue light-emitting diodes," J. Luminescence, 5, 84 (Mar 1972)
6. H.P. Maruska, D.A. Stevenson, J.I Pankove "Violet Luminescence of Mg-doped GaN," Appl. Phys. Lett., 22, 303 (1973).
7. J. I. Pankove, E. A. Miller, J. E. Berkeyheiser, "Electroluminescence in GaN," Proc. Of Luminescence Conf. Leningrad, Ferd Williams, editor, Plenum Publishing Corp, New York, pp. 426-430 (1972).
8. J. I. Pankove, M. T. Duffy, E. A. Miller, J. E. Berkeyheiser, "Luminescence of insulating Be-doped and Li-doped GaN," J. Luminescence, 8, 89 (Sept 1973).
9. J. I. Pankove, J. E. Berkeyheiser, E. A. Miller, "Properties of Zinc-doped GaN," J. Appl. Phys., 45, 1280 (Mar 1974).

 Return to Search Results
Start a New Search

*****