`
`Scientific Background on the Nobel Prize in Physics 2014
`
`E F F I C I E N T B L U E L I G H T -E M I T T I N G D I O D E S L E A D I N G
`T O B R I G H T A N D E N E RG Y -S AV I N G W H I T E L I G H T S O U RC E S
`
`compiled by the Class for Physics of the Royal Swedish Academy of Sciences
`
`THE ROYAL SWEDISH ACADEMY OF SCIENCES has as its aim to promote the sciences and strengthen their influence in society.
`
`BOX 50005 (LILLA FRESCATIVÄGEN 4 A), SE-104 05 STOCKHOLM, SWEDEN
`TEL +46 8 673 95 00, INFO@KVA.SE HTTP://KVA.SE
`
`Nobel Prize® and the Nobel Prize® medal design mark
`are registrated trademarks of the Nobel Foundation
`TCL 1033, Page 1
`
`
`
`TCL 1033, Page 2
`TCL 1033, Page 2
`
`
`
`Efficient blue light-emitting diodes leading to bright
`and energy-saving white light sources
`
`
`Light-emitting diodes (LEDs) are narrow-band light sources based on semiconductor
`components, with wavelengths ranging from the infrared to the ultraviolet. The first
`LEDs were studied and constructed during the 1950s and 1960s in several laboratories.
`They emitted light at different wavelengths, from the infrared to the green. However,
`emitting blue light proved to be a difficult task, which took three more decades to
`achieve. It required the development of techniques for the growth of high-quality
`crystals as well as the ability to control p-doping of semiconductors with high bandgap,
`which was achieved with gallium-nitride (GaN) only at the end of the 1980s. The
`development of efficient blue LEDs also required the production of GaN-based alloys
`with different compositions and their integration into multilayer structures such as
`heterojunctions and quantum wells.
`
`The invention of efficient blue LEDs has led to white light sources for illumination.
`When exciting a phosphor material with a blue LED, light is emitted in the green and red
`spectral ranges, which, combined with the blue light, appears as white. Alternatively,
`multiple LEDs of complementary colours (red, green and blue) can be used together.
`Both of these technologies are used in today's high-efficiency white electroluminescent
`light sources. These light sources, with very long lifetimes, have begun to replace
`incandescent and fluorescent lamps for general lighting purposes. Since lighting
`represents 20-30% of our electrical energy consumption, and since these new white
`light sources require ten times less energy than ordinary light bulbs, the use of efficient
`blue LEDs leads to significant energy savings, of great benefit to mankind.
`
`
`This year’s Nobel Prize in Physics honours the inventors of efficient blue LEDs: I.
`Akasaki, H. Amano and S. Nakamura.
`
`Early history
`
`The first report of electrically generated light by emission from a solid-state device
`came from H.J. Round working at Marconi Electronics in 1907 [1]. He applied voltage
`across two contacts on a carborundum (SiC) crystal. At low voltages yellow light was
`observed, but more colours were emitted at higher voltages. Electroluminescence was
`also studied by O. Losev (1903-1942), a device physicist in the Soviet Union, who in the
`1920s and 1930s published several articles in international journals on
`electroluminescence from carborundum [2]. These developments took place prior to
`the formulation of the modern theory of electronic structure of solid-state materials.
`
`
`
`1
`TCL 1033, Page 3
`
`
`
`
`
`
`
`
`Fig. 1. Principle for light emission in a p-n junction. In a p-n junction biased with a forward voltage,
`electrons are injected from the n- to the p-side, and holes are injected in the opposite direction.
`Electrons recombine with holes and light is emitted (spontaneous emission). For efficient diodes it is
`important that the semi-conductors have direct bandgaps. LEDs with indirect bandgaps require
`phonon-assisted recombination, which limits the efficiency. The quantum efficiency of a LED is the
`ratio of the number of emitted photons to the number of electrons passing through the contact in a
`given time.
`
`
`
`
`
`The understanding of the physics of semiconductors and p-n junctions progressed
`during the 1940s, leading to the invention of the transistor at Bell Telephone
`Laboratories in the USA in 1947 (Nobel Prize 1956 to Shockley, Bardeen and Brattain).
`It became clear that a p-n junction could be an interesting device for light emission. In
`1951, K. Lehovec and co-workers of the Signal Corps Engineering Laboratory in the USA
`[3] used these ideas to explain the electroluminescence in SiC as resulting from the
`injection of carriers across a junction followed by radiative recombination of electrons
`and holes. However, the observed photon energy was less than the energy gap of SiC,
`and they suggested that radiative recombination was likely to occur due to impurities or
`lattice defects. In 1955, injection electroluminescence was shown in a number of III-V
`compounds [4, 5]. In 1955 and 1956, J.R. Haynes at Bell Telephone Laboratories
`demonstrated that electroluminescence observed in germanium and silicon was due to
`recombination of holes and electrons in a p-n junction [6] (see Fig. 1).
`
`Infrared LEDs
`
`Techniques to make efficient p-n junctions with GaAs were rapidly developed during
`the following years. GaAs was attractive because of its direct bandgap, enabling
`recombination of electrons and holes without involvement of phonons. The bandgap is
`1.4 eV corresponding to light in the infrared. In the summer of 1962, the observation of
`light emission from p-n-junctions was reported [7]. A few months later, laser emission
`in GaAs at liquid nitrogen temperature (77 K), was demonstrated independently and
`almost simultaneously by three research groups at General Electric, IBM and the MIT
`
`
`
`2
`TCL 1033, Page 4
`
`
`
`Lincoln Laboratory, in the U.S. [8-10]. It would be a few years, however, before laser
`diodes became widely used. Thanks to the development of heterostructures (Nobel
`Prize 2000 to Z.I. Alferov and H. Kroemer), and later quantum wells, allowing for a
`better confinement of the carriers while reducing the losses, laser diodes could operate
`continuously at room temperature, with applications in a large variety of areas.
`
`Visible LEDs
`
`Following early experiments at the end of the 1950s [11], progress in making efficient
`LEDs using GaP (indirect bandgap equal to 2.2 eV) was made in parallel by three
`research groups from Philips Central Laboratory in Germany (H.G. Grimmeiss), the
`Services Electronics Laboratories (SERL) in the UK (J.W. Allen) and Bell telephone
`laboratories in the USA (M. Gershenzon) [12-14]. They had different objectives, ranging
`from communication, lighting and television to indicator lamps for electronics and
`telephones. Using different dopants (e.g. Zn-O or N) at various concentrations, different
`wavelengths were generated ranging from red to green. By the late 1960s a number of
`manufacturers in different countries were making red and green LEDs based on GaP.
`
`Mixed crystals including Ga, As, and P (GaPxAs1-x) are interesting since the emission
`wavelength can be shorter than for GaAs, reaching the visible range while the bandgap
`is direct for x below 0.45. N. Holonyak Jr. and co-workers at the General Electric
`laboratory in the USA, began to work with GaPxAs1-x in the late 1950s, and succeeded in
`making p-n junctions and observing LED emission. Laser diode emission at 710 nm
`(red) was reported in 1962 [15].
`
`Early work on blue LEDs
`
`The step to the emission of blue light proved to be considerably more difficult. Early
`attempts with ZnSe and SiC, with high indirect bandgaps, did not lead to efficient light
`emission. The material that enabled the development of blue LEDs was GaN (Gallium
`Nitride).
`
`
`Gallium Nitride
`
`GaN is a semiconductor of the III-V class, with Wurtzite crystal structure. It can be
`grown on a substrate of sapphire (Al2O3) or SiC, despite the difference in lattice
`constants. GaN can be doped, e.g. with silicon to n-type and with magnesium to p-
`type. Unfortunately, doping interferes with the growth process so that the GaN
`becomes fragile. In general, defects in GaN crystals lead to good electron
`conductivity, i.e. the material is naturally of n-type. GaN has a direct bandgap of
`3.4 eV, corresponding to a wavelength in the ultraviolet.
`
`
`
`
`
`
`
`3
`TCL 1033, Page 5
`
`
`
`
`
`
`
`
`
`Fig. 2.
`a) Growth of GaN on sapphire using an AlN layer [27].
`
`b) Resistivity of Mg doped GaN as a
`function of annealing temperature [32].
`
`
`
`
`Already at the end of the 1950s, the possibility of a new lighting technology using GaN,
`the bandgap of which had just been measured, was seriously considered at Philips
`Research Laboratories. H.G. Grimmeiss and H. Koelmans obtained efficient
`
`photoluminescence from GaN over a wide spectral range using different activators and
`a patent was filed [16]. However, at that time it was very difficult to grow GaN crystals.
`Only small crystals, forming a powder, could be produced, in which p-n junctions could
`not be created. The researchers at Philips decided to concentrate on GaP instead (see
`above).
`
`
`GaN crystals were more efficiently produced at the end of the 1960s by growing GaN on
`a substrate using the HVPE technique (Hydride Vapour Phase Epitaxy) [17]. A number
`of laboratories in the United States [18, 19], in Japan [20] and in Europe [21] studied the
`growth techniques and doping of GaN with the goal of developing blue LEDs, but
`material problems still seemed insurmountable. The surface roughness was not
`controlled, the HVPE-grown material was contaminated with transition metal
`impurities and p-doping was passivated due to the presence of hydrogen, forming
`complexes with acceptor dopants. The role of hydrogen was not understood at that
`time. J.I. Pankove, a leading scientist in the field, wrote in a review article from 1973
`[22]: "In spite of much progress in the study of GaN over the last two years, much remains
`to be done. The major goals in the technology of GaN should be: (1) the synthesis of strain-
`free single crystals, (2) the incorporation of a shallow acceptor in high concentrations" (to
`provide effective p-doping). Again the research effort was halted due to lack of progress.
`
`New growth techniques
`
`In the 1970s, new crystal growth techniques, MBE (Molecular Beam Epitaxy) [23] and
`MOVPE (Metalorganic Vapour Phase Epitaxy) [24] were developed. Efforts were made
`to adapt these techniques for growing GaN [25]. Isamu Akasaki began studying GaN as
`
`
`
`4
`TCL 1033, Page 6
`
`
`
`early as 1974, at the time working at the Matsushita Research Institute in Tokyo. In
`1981, he took up a professorship at Nagoya University and continued his research on
`GaN, together with Hiroshi Amano and other co-workers. It would take until 1986
`before GaN with high crystal quality and good optical properties could be produced
`with the MOVPE technique [26]. The breakthrough was the result of a long series of
`experiments and observations. A thin layer (30 nm) of polycrystalline AlN was first
`nucleated on a substrate of sapphire at low temperature (500 °C) and then heated up to
`the growth temperature of GaN (1000 °C) . During the heating process, the layer
`develops a texture of small crystallites with a preferred orientation on which GaN can
`be grown. The density of dislocations of the growing GaN crystal is first high, but
`decreases rapidly after a few m growth. A high quality surface could be obtained,
`which was very important to grow thin multilayer structures in the following steps of
`the LED development. In this way, high quality device-grade GaN was obtained for the
`first time (see Fig. 2a). GaN could also be produced with significantly lower background
`n-doping. Shuji Nakamura at Nichia Chemical Corporation, a small chemical company in
`Japan, later developed a similar method where AlN was replaced with a thin layer of
`GaN grown at low temperature [28].
`
`Doping of GaN
`
` A
`
` major problem for manufacturing p-n junctions was the difficulty to p-dope GaN in a
`controlled manner. At the end of the 1980s, Amano, Akasaki and co-workers made an
`important observation; they noted that when Zn-doped GaN was studied with a
`scanning electron microscope, it emitted more light [29], thus indicating better p-
`doping. In a similar way, when Mg-doped GaN was irradiated with low energy electrons,
`it resulted in better p-doping properties [30]. This was an important breakthrough and
`opened the way to p-n junctions in GaN.
`
`
`The effect of electron irradiation was explained a few years later, in an article by
`Nakamura and co-workers [31]. Acceptors such as Mg or Zn form complexes with
`hydrogen and thus become passive. Electron beams dissociate these complexes and
`activate the acceptors. Nakamura showed that even a simple thermal treatment
`(annealing) leads to efficient activation of Mg acceptors. The effect of hydrogen on the
`neutralization of dopants was known from previous work using other materials by
`Pankove [32], G.F. Neumark Rothschild [33], and others.
`
`
` A
`
` crucial step in developing efficient blue LEDs was the growth and p-doping of alloys
`(AlGaN, InGaN), which are necessary in order to produce heterojunctions. Such
`heterojunctions were realized in the early 90s in both Akasaki’s and Nakamura's
`research groups [34, 35].
`
`
`
`5
`TCL 1033, Page 7
`
`
`
`
`
` Fig. 3. Structure of a blue LED with a double heterojunction InGaN/AlGaN. From [39]
`
`
`
`
`
`Double heterostructures and quantum wells
`
`
`The development of infrared LEDs and laser diodes had shown that heterojunctions and
`quantum wells were essential to achieve high efficiency. In such structures holes and
`
`electrons are injected in a small volume where recombination occurs more efficiently
`and with minimal losses. Akasaki and co-workers developed structures based on
`AlGaN/GaN [36, 37] while Nakamura with great success exploited the combinations
`InGaN/GaN and InGaN/AlGaN for producing heterojunctions, quantum wells and
`multiple quantum wells [38]. In 1994, Nakamura and co-workers achieved a quantum
`efficiency of 2.7% using a double heterojunction InGaN/AlGaN (see Fig. 3) [39]. With
`these important first steps, the path was cleared towards the development of efficient
`blue LEDs and their application was open. Both teams have continued to develop blue
`LEDs, aiming towards higher efficiency, versatility and applications. Blue laser emission
`based on GaN was observed in 1995-1996 by both groups [40, 41].
`
`Today’s efficient GaN-based LEDs result from a long series of breakthroughs in basic
`materials physics and crystal growth, in device physics with advanced heterostructure
`design, and in optical physics for the optimization of the light out-coupling. The
`historical development of blue, green, red and “white” LEDs is summarized in the
`picture below.
`
`
`
`
`
`
`
`
`6
`TCL 1033, Page 8
`
`
`
`Fig. 4. Historical evolution of commercial LEDs. From [42]. PC-White stands for phosphor
`converted white light, DH stands for double heterostructure. The wallplug efficiency is the
`ratio between emitted light power and supplied electrical power.
`
`Applications
`
`Illumination technology is presently going through a revolution, namely the transition
`from light bulbs and fluorescent tubes to LEDs. The light bulb, invented by Thomas
`Edison in 1879, has a low efficiency ≈16 lm/W representing approximately 4% energy
`efficiency from electricity into light. A lumen is a unit used to characterize the light flux,
`which takes into account the eye's spectral response. The fluorescent tube, containing
`mercury and invented by P. Cooper Hewitt in 1900, reaches an efficiency of 70 lm/W.
`White LEDs currently reach more than 300 lm/W, representing more than 50%
`wallplug efficiency.
`
`White LEDs used for lighting are often based on efficient blue LEDs that excite a
`phosphor so that the blue light is converted to white light. These high-quality LEDs with
`their very long lifetime (100 000 hours) are getting cheaper, and the market is currently
`exploding. A little further on in the future, three-colour LEDs may replace the
`combination of blue LED and phosphor for efficient lighting. This technology will allow
`for dynamic control of colour composition.
`
`Replacing light bulbs and fluorescent tubes with LEDs will lead to a drastic reduction of
`electricity requirements for lighting. Since 20-30% of the electricity consumed in
`industrial economies is used for lighting, considerable efforts are presently being
`devoted to replacing old lighting technologies with LEDs.
`
`7
`TCL 1033, Page 9
`
`
`
`Today, GaN-based LEDs provide the dominant technology for back-illuminated liquid
`crystal displays in many mobile phones, tablets, laptops, computer monitors, TV
`screens, etc. Blue and UV-emitting GaN diode lasers are also used in high-density DVDs,
`which has advanced the technology for storing music, pictures and movies. Future
`application may include the use of UV-emitting AlGaN/GaN LEDs for water purification,
`as UV light destroys the DNA of bacteria, viruses and microorganisms. In countries with
`insufficient or non-existent electricity grids, the electricity from solar panels stored in
`batteries during daylight, powers white LEDs at night. There, we witness a direct
`transition from kerosene lamps to white LEDs.
`
`References
`
`1. H.J. Round, Electr. World, 49, 308 (1907).
`2. O.V. Losev, Telegrafi ya i Telefoniya bez Provodov 44, 485 (1927); Phil. Mag. 6, 1024
`(1928); Compt. Rendu Acad. Sci. 39, 363 (1940); USSR patent 12191 (1929).
`3. K. Lehovec, C.A. Accardo & E. Jamgochian, Phys. Rev. 83, 603 (1951).
`4. G.A. Wolff, R.A. Hebert & J.D. Broder, Phys. Rev. 100, 1144 (1955).
`5. R. Braunstein, Phys. Rev. 99, 1892 (1955).
`6. J.R. Haynes, Phys. Rev. 98, 1866 (1955); J.R. Haynes & W.C. Westphal, Phys. Rev. 101,
`1676 (1956).
`7. J.I. Pankove, Phys. Rev. Lett. 9, 283-285 (1962); J.I. Pankove & J.E. Berkeyheiser, Proc.
`IRE, 50, 1976-1977 (1962); R.J. Keyes & T.M. Quist, Proc. IRE, 50, 1822-1823
`(1962).
`8. R.N. Hall, G.E. Fenner, J.D. Kingsley, T.J. Soltys & R.O. Carlson, Phys. Rev. Lett. 9, 366–
`368 (1962).
`9. M.I. Nathan, W.P. Dumke, G. Burns, F.H. Dill & G. Lasher, Appl. Phys. Lett. 1, 62–63
`(1962).
`10. T.M. Quist, R.H. Rediker et al., Appl. Phys. Lett. 1, 91 (1962).
`11. G.A. Wolff, R.A. Hebert & J.D. Broder, Phys. Rev. 100, 1144 (1955); D.A. Holt, G.F.
`Alfrey & C.S. Wigglins, Nature 181, 109 (1958).
`12. H.G. Grimmeiss & H. Koelmans, Phys. Rev. 123, 1939 (1961); H.G. Grimmeiss & H.
`Scholz, Phys. Lett. 8, 233 (1964).
`13. J. Starkiewicz & J.W. Allen, J. Phys. Chem. Solids 23, 881 (1962).
`14. M. Gershenzon & R.M. Mikulyak, J. Appl. Phys. 32, 1338 (1961).
`15. N. Holonyak & S.F. Bevacqua, Appl. Phys. Lett. 1, 82 (1962).
`16. H.G. Grimmeiss and H. Koelmans, Z. f. Naturforsch. 14a, 264 (1959); 15, 799 (1960);
`H.G. Grimmeiss, H. Koelmans & I.B. Maak, German patent, DBP 1 077 330 (1960).
`17. H.P. Maruska & J.J. Tietjen, Appl. Phys. Lett. 15, 327 (1969).
`18. R. Dingle, D.D. Sell, S.E. Stokowski & M. Ilegems, Phys. Rev. B 4, 1211 (1971).
`19. J.I Pankove, E.A. Miller, D. Richman & J.E. Berkeyheiser, J. Lumin.4, 63, (1971); 8, 89
`(1973); H.P. Maruska, D.A. Stevenson and J.I. Pankove, Appl. Phys. Lett. 22, 303
`(1973).
`20. M. Sano & M. Aoki, Jpn. J. Appl. Phys. 15, 1943 (1976).
`
`
`
`8
`TCL 1033, Page 10
`
`
`
`21. H. G. Grimmeiss & B. Monemar, J. Appl. Phys. 41, 4054 (1970); B. Monemar, Phys. Rev.
`B 10, 676 (1974).
`22. J.I Pankove, J. Lumin.7, 114 (1973).
`23. A. Y. Cho & J. R. Arthur, Prog. Solid State Chem. 10, 157 (1975).
`24. H.M. Manasevit, F.M. Erdman & W. I. Simpson, J. Electrochem. Soc. 118, 1864 (1971).
`
`25. S. Yoshida, S. Misawa & S. Gonda, Appl. Phys. Lett. 42, 427 (1983).
`
`26. H. Amano, N. Sawaki, I. Akasaki & Y. Toyoda, Appl. Phys. Lett. 48, 353 (1986).
`27. K. Hiramatsu et al., J. Crystal Growth 115, 628 (1991).
`28. S. Nakamura, Jpn. J. Appl. Phys. 30, L1705 (1991); S. Nakamura, M. Senoh, & T. Mukai,
`Jpn. J. Appl. Phys. 30, L1998 (1991).
`29. H. Amano, I. Akasaki, T. Kozawa, K. Hiramatsu, N. Sawaki, K. Ikeda & Y. Ishii, J. Lumin.
`40 &41, 121 (1988).
`30. H. Amano, M. Kito, K. Hiramatsu, & I. Akasaki, Jpn. J. Appl. Phys. 28, L2112 (1989).
`31. S. Nakamura, N. Iwasa, M. Senoh, & T. Mukai, Jpn. J. Appl. Phys. 31, 1258 (1992); S.
`Nakamura, T. Mukai, M. Senoh & N. Iwasa, Jpn. J. Appl. Phys. 31, L139 (1992).
`32. J.I. Pankove, D.E. Carlson, J.E. Berkeyheiser & R.O. Wance, Phys. Rev. Lett. 51, 2224
`(1983).
`33. G.F. Neumark Rothschild, US patent 5252499 (1988).
`34. H. Murakami, T. Asahi, H. Amano, K. Hiramatsu, N. Sawaki & I. Akasaki, J. Crystal
`Growth 115, 648 (1991).
`35. S. Nakamura & T. Mukai, Jpn. J. Appl. Phys. 31, L1457 (1992).
`36. K. Itoh, T. Kawamoto, H. Amano, K. Hiramatsu & I. Akasaki, Jpn. J. Appl. Phys. 30,
`1924 (1991).
`37. I. Akasaki, H. Amano, K. Itoh, N. Koide & K. Manabe, Int. Phys. Conf. Ser. 129, 851
`(1992).
`38. S. Nakamura, M. Senoh, & T. Mukai, Jpn. J. Appl. Phys. 32, L8 (1993); S. Nakamura et
`al., J. Appl. Phys. 74, 3911 (1993).
`39. S. Nakamura, T. Mukai & M. Senoh, Appl. Phys. Lett. 64, 1687 (1994).
`40. I. Akasaki, H. Amano, S. Sota, H. Sakai, T. Tanaka & M. Koike, Jpn. J. Appl. Phys. 34,
`L1517 (1995).
`41. S. Nakamura et al., Jpn. J. Appl. Phys. 35, L74 (1996).
`42. M.H. Crawford et al., “Toward Smart and Ultra-Efficient Materials, Devices, Lamps
`and Systems”.
`
`Further sources on the history of LEDs are
`E. Fred Schubert: Light Emitting Diodes, 2nd edition, (2006).
`H.G. Grimmeiss and J.W. Allen, J. Non-crystalline Solids 352, 871 (2006).
`S. Nakamura and M.R. Krames, Proc. IEEE 101, 2211 (2013).
`R.D. Dupuis and M.R. Krames, J. Lightwave Tech. 26, 1154 (2008).
`
`9
`TCL 1033, Page 11
`
`