Documente Academic
Documente Profesional
Documente Cultură
In the space of only a month, history was made four decades ago when four groups of researchers
independently developed and demonstrated their own versions of the injection laser. The author describes
how the experiments laid the foundation for the vast array of materials and technologies used in the
fabrication of modern compound semiconductor alloy devices.
A
little over 40 years ago, on Sept. versions of the injection laser. The efforts lasers were fabricated by zinc (Zn) diffu-
16, 1962, Gunther Fenner, a of the other three groups that also suc- sion into conventional (and commer-
member of the team headed by ceeded in making a semiconductor laser cially available) n-type gallium-arsenide
Robert N. Hall at the General Electric in late 196211 were led by: Nick Holonyak (GaAs) “bulk” crystals, one of the “first”
Research Development Center in Jr.4-6 at General Electric’s Syracuse, N.Y., laser diodes (Holonyak’s) was created
Schenectady, N.Y., operated the first facility; Marshall I. Nathan7, 8 at IBM from a small piece of single-crystal
semiconductor diode laser.1-3 Within Research Laboratory, Yorktown Heights, ternary gallium-arsenide (GaAsP) alloy
about 30 days, workers in three other N.Y.; and Robert Rediker 9, 10 at MIT material grown by vapor-phase trans-
laboratories in the United States had Lincoln Laboratory, Lexington Mass. port, making it the first alloy compound
independently demonstrated their own While three of these early p-n junction semiconductor device to assume
© 2003 IEEE. Reprinted with permission from the Feb. 2003 issue of IEEE LEOS Newsletter.
DE (eV)
about the semiconductor laser In DOT
theoretical papers on semi-
problem for some time prior to P GaAs conductor lasers written by
N GaAs
the June 1962 SSDRC confer- 0.001 Au PLATED KOVAR
Benjamin Lax, Herb Zeiger
ence; Dumke, in particular, had and Bill Krag in 1959.31, 32 As
calculated the requirements for noted, above, Lincoln Labs’
population inversion in a semi- 10-4 10-3 10-2 10-1 1 10 work was devoted to the
Current (amps)
conductor.27 The first sentence development of efficient
of Dumke’s paper (submitted emitters and included several
April 3, 1962), reads: “Since the initial Figure 2. Spectral linewidth vs. current for a key contributors to the theory and the
operation of the ruby maser, there has GaAs diode made at IBM and operated at experimental characterization of the
been considerable speculation concern- 77 K. The diode did not have a Fabry-Pérot recombination radiation from GaAs p-n
geometry so cavity modes were not
ing the possibility of observing maser observed. [From Nathan, IEEE J. Quant.
junctions. In 1962, this effort—centered
action in semiconductors such as Ge and Electron. QE-23, 679 (1987).] in Rediker’s group—included Bob Keyes,
Si.” After rejecting the possibility of laser Bill Krag, Ted Quist, Al McWhorter, Herb
operation in the indirect semiconductor Zeiger and others. The search for laser
Ge due to excessive free-carrier absorption, was submitted to Applied Physics Letters operation from GaAs diodes fabricated at
near the end of this paper, he concluded: (received on Oct. 6, 1962) and published Lincoln Labs met with success on Oct. 12,
“At present, it is not clear whether or not on Nov. 1, 1962.7 Later, Rick Dill and 1962,33 with laser operation first con-
one would obtain anything like typical Dick Rutz developed a process for for- firmed by the observation of “filaments”
maser action from a device utilizing mation of the optical cavity by cleaving in the near-field pattern of a GaAs diode
direct transitions as in GaAs.” 27 all four sides of the laser diode.28 when examined with an infrared con-
After hearing the news of the 1962 The early Lincoln Labs GaAs program verter.9 The Lincoln Labs device
SSDRC reports on efficient GaAs was quite small; it involved two technical employed a Fabry-Pérot cavity having
diodes, the IBM team was spurred to an staff members and a technician. The polished parallel facets, as suggested ear-
increased level of activity. Several mem- team’s work on GaAs was initiated in lier by Zeiger.34 The paper in which
bers of the newly expanded laser team 1958 to study the possibility of making these laser results were described was
dedicated their efforts to making diodes III-V high-speed microwave devices.10, 29 submitted to the editor of Applied Physics
and analyzing their performance. Gordon The III-V program was driven by the Letters on Oct. 23, 1962. It was published
Lasher considered what might theoreti- vision of Robert Rediker, who champi- (after changes) on Dec. 1, 1962. A photo-
cally be observed in the stimulated emis- oned the study of GaAs when most other graph of one of the early Lincoln Labs’
sion spectrum of a GaAs diode. Marshall groups were concentrating on silicon. GaAs laser diodes is shown in Fig. 3.
Nathan concentrated on studying the Rediker had visited professor Heinrich
photoluminescence and electrolumines- Welker in Erlangen, Germany, in 1958 to Do alloys work?
cence from various GaAs structures oper- learn more about GaAs and related mate- It is interesting to note that in 1960-62,
ated under pulsed conditions at low rials. While working at Siemens, Welker many in the field of semiconductors were
temperatures. While the IBM team real- had been the first person in the West to convinced that an alloy was not worth
ized that for a diode laser, spectral line identify the “intermetallic” III-V materi- pursuing and, that if such materials were
narrowing should occur and that the als as semiconductors. to be successfully produced, random
light output versus current should be The Lincoln team developed a diffu- alloy “disorder” would render them use-
superlinear, because they could not figure sion technique for producing p-n junc- less because of the stochastic nature of
out how to provide proper “cavity mir- tions. The team’s laser effort had, in fact, the distribution of atoms in the lattice.35
rors” for the diodes, no provision was evolved from the work of Rediker’s group There were also different ideas about how
made in this early work for optical on the comparison of the electrolumines- to produce such alloys. One, for example,
feedback.8 On Sept. 29, 1962, Nathan cence characteristics of GaAs diffused (which has not succeeded even to date)
observed narrowing of the electrolumi- and alloy diodes. Keyes was invited to involved the diffusion of P into GaAs to
nescence to a FWHM of ~3 nm. A few join the team since “he owned the spec- create GaAsP layers. Epitaxial growth of
days later, he found that the linewidth trometer.”30 As noted above, the work semiconductors was still a very new con-
measured for this first successful diode of Keyes and Quist on the study of the cept and while the creation of alloyed
was the limit of the spectrometer: recombination radiation from diffused p-n junctions using metal “performs”
~ 0.2 nm. The GaAs diodes showing GaAs p-n junctions led to the observation was a well-known process, it lacked the
“stimulated emission” first reported by and report of high internal quantum control necessary to create true alloy
the IBM group did not employ an optical efficiencies, which “sparked” the post- semiconductors.
cavity.7,8 The IBM paper describing the SSDRC efforts at many research labs Holonyak had been working on
group’s first stimulated emission results to develop a semiconductor laser. The the growth of ternary GaAsP alloy
GaAsP lasers for sale, becoming 6. N. Holonyak, IEEE J. Sel. Top. Quant. Electron. 6,
1190 (2000).
the first company to offer such 7. M. I. Nathan, W. P. Dumke, G. Burns, F. H. Dill,
devices commercially.39 Since the Jr., and G. Lasher, Appl. Phys. Lett. 1, 62 (1962).
Nathan’s paper was received Oct. 6, 1962.
devices only operated pulsed at low Wire bond 8. M. I. Nathan, IEEE J. Quant. Electron. QE-23,
Laser diode 679 (1987).
temperatures, they were obviously
Contact 9. T. M. Quist, R. H. Rediker, R. J. Keyes, W. E. Krag,
useful only for research purposes or B. Lax, A. L. McWhorter, and H. J. Zeiger, Appl. Phys.
Lett. 1, 91 (1962). Quist’s paper was received Oct.
special defense applications. The Red laser 23, 1962 and in final form Nov. 5, 1962.
price for one of Hall’s IR-emitting light 10. R. H. Rediker, IEEE J. Quant. Electron. QE-23,
692 (1987).
GaAs laser diodes was initially 11. For details of this work, see the related Special
$1,600—a price which had been Issue Papers in the IEEE J. Quant. Electron. QE-23,
June 1987, Refs. 3, 5, 8, and 10.
somewhat arbitrarily set by the GE Scattered light 12. Goryunova described III-V materials as semicon-
marketing group at 10 times that 0.2 mm ductors for the first time in 1950. In her Ph.D. thesis,
completed in 1951 at Leningrad State University, she
of a currently available Texas Heat sink indicated that III-V zinc-blend compounds are semi-
conductors. Her work was not published outside of
Instruments IR incoherent “LED.” the U.S.S.R. until much later due to the Cold War.
The price was later reduced to 13. R. J. Keyes and T. M. Quist, unpublished paper
$800.39 The “visible red” GaAsP laser Figure 4. Photograph of one of Holonyak’s presented at the Solid State Device Research
Conference, Durham, N.H., July 1962.
first GaAsP injection lasers. This is the first
diode was initially priced at $3,200, then 14. R. J. Keyes and T. M. Quist, Proc. IRE 50, 1822 (1962).
direct photograph of a laser diode made
reduced to $1,600.39 This pricing deci- 15. R. J. Keyes, T. M. Quist, R. H. Rediker, M. J. Hudson, C. R.
using its own photon emission as a light Grant, and J. W. Meyer, Electron. 36, 39 (1963).
sion was based on the fact that GE’s source. The color film is overexposed in the 16. N. Holonyak, Jr., Am. J. Phys., 68, 864 (2000).
marketing managers decided that region of the facet where laser operation is 17. The IBM work was sponsored by the US Army Electronics
Research and Development Laboratory, Ft. Monmouth,
Holonyak’s visible laser was “twice as occurring. [From N. Holonyak Jr., IEEE J. N.J., under Contract DA 36-039-SC-90711). See Ref. 7.
valuable” as Hall’s IR-emitting laser Quantum Electron. QE-23, 684-91 (1987).] 18. Some of this early work is discussed briefly in R. D.
Dupuis, IEEE J. Quant. Electron. QE-23, 651 (1987).
device. Incoherent GaAsP visible LEDs 19. D. N. Nasledov, A. A. Rogachev, S. M. Rivkin, and B. V.
were also offered for sale. Interestingly, Tsarenkov, Sov. Phys. Sol. State. 4, 782 (1962).
20. J. L. Bromberg, The Laser History Project.
40 years later, GaAsP red-emitting LEDs
21. M. G. A. Bernard and G. Duraffourg, Phs. Stat. Sol. 1,
are still sold. While some researchers did heterojunction quantum-well DVD 699 (1961).
not appreciate the fundamental distinc- lasers, for example, are currently grown 22. T. Maiman, Nature 187, 493 (1960).
23. See T. Maiman, in “The Laser Odyssey,” Laser Press,
tion between “infrared” GaAs lasers and primarily by metal-organic chemical Blaine, Wash., 2000.
“visible” GaAsP lasers, Holonyak was vapor deposition and cost considerably 24. John von Neumann carried out the first documented
theoretical treatment of a semiconductor laser in 1953.
acutely aware of the significance of visi- less than $1 in packaged form. It is also This paper is reproduced in J. von Neumann, IEEE J.
ble-spectrum LEDs and lasers, a feeling interesting to note that GaAsP LEDs, Quant. Electron. QE-23, 658 (1987).
25. See the reproduction of his unpublished 1953 paper in J.
shared by others working in the field.40 closely related to those first diffused von Neumann, IEEE J. Quant. Electron. QE-23, 659
These first visible injection lasers made GaAsP diodes, are still produced com- (1987).
26. J. Black, H. Lockwood, and S. Mayburg, Postdeadline
of the ternary alloy GaAsP foreshadowed mercially. And the diode laser has become Paper P14, presented at the American Physical Society
meeting, Baltimore, Md., March 28, 1962.
the future when virtually all semiconduc- the dominant form in the commercial
27. W. P. Dumke, Phys. Rev. 127, 1559 (1962).
tor lasers would employ alloy materials. laser market, with millions produced 28. F. H. Dill and R. F. Rutz, US Patent 3247576 (filed Oct.
In addition, the GaAsP first compound each year. We enjoy a multitude of bene- 30, 1962, issued Apr. 26, 1966).
29. R. H. Rediker, private communication, Sept. 2002.
semiconductor alloy light emitters were fits from the early research and devel-
30. R. H. Rediker, private communication, Dec. 4, 2002.
the earliest progenitors of the now ubiq- opment of semiconductor diode lasers, 31. B. Lax, in “Quantum Electron., A Symposium,” C. H.
uitous LED and have spawned related multi-element semiconductor alloys Townes, Ed. New York; Columbia University, 1960, p.
428.
devices in a variety of III-V ternary and and their derivative heterojunctions. 32. H. J. Zeiger and W. E. Krag, Quarterly Progress Report on
quaternary materials systems, leading to The future seems to hold an even greater Solid State Research, Lincoln Laboratory, MIT, Oct. 1959,
p. 41.
the continued development of an “ulti- variety of applications for these devices 33. R. H. Rediker, person communication, Nov. 24, 2002.
mate light source”—the high-efficiency and their progeny. 34. The MIT Lincoln Labs’ team reviewed Hall’s Phys. Rev.
Lett. GaAs laser paper describing the optical cavity
injection luminescence source available formed by polishing so they knew that this approach
Russell D. Dupuis (russell.dupuis.@ece.gatech.edu) would work.
today in the form of advanced “high-
holds the Steve W. Chaddick Chair in electro- 35. N. Holonyak, Jr., private communication, September
brightness” LED products in the InAlGaP optics in the School of Electrical and 2002.
and InAlGaN alloy material systems. Computer Engineering at Georgia
Member Institute of Technology, Atlanta, Ga.
36. Holonyak had forgotten about the infrared converting
“snooper scope” that was used by Hall’s group at GE, as
Today, over 40 years after the first well as by Rediker’s group at Lincoln Labs, to observe the
emission patterns from their IR-emitting diode lasers.
demonstration of the injection laser in References 37. Air Force contract AF 19 (604)-6623.
September 1962 and the first demonstra- 1. Robert N. Hall, private communication, October 2002. 38. However, Hall’s GaAs laser work was funded entirely by
tion of a compound semiconductor 2. R. N. Hall, G. E. Fenner, J. D. Kingsley, T. J. Soltys, and R. GE Internal Research and Development funds and all the
technical notes and data were recorded in notebooks
O. Carlson, Phys. Rev. Lett. 9, 366 (1962). The editor of
alloy device in October of the same year, Phys. Rev. received this paper Sept. 24, 1962.
separate from those used for the Air Force contract
research. (R. N. Hall, private communication, Nov.
advances in materials and device design 3. R. N. Hall, IEEE J. Quant. Electron. QE-23, 674 (1987). 2002).
have made the injection laser an essential 4. N. Holonyak, Jr. and S. F. Bevacqua, Appl. Phys. Lett. 1, 39. Allied Industrial Electronics Catalog No. 650, Chicago,
82 (1962). Holonyak’s paper was received by the editor Ill., 1965, p. 77.
device in many important systems. High- of Appl. Phys. Lett. Oct. 17, 1962. 40. See the article by Harlan Manchester, “Light of Hope—
performance red-emitting InAlGaP 5. N. Holonyak, IEEE J. Quant. Electron. QE-23, 684 (1987). Or Terror,” Readers Digest, p. 97 (Feb. 1963).