Lednetics.com

The first devices

In 1907, at the Marconi Company laboratories, Henry Joseph Round discovered the effect of photoluminescence using silicon carbide (SiC) and a radio component. Exactly twenty years later, Oleg Losev published a theory in Russia, where he claimed to have created the equivalent of a rudimentary LED; although the publication was also published in the German and British Empires, for decades there were no practical applications for his invention.

In 1936, at Marie Curie’s laboratories, physicist Georges Destriau obtained electroluminescence (which he himself called “Losev light”) by encapsulating zinc sulfide (ZnS) in a container in which he applied an alternating electric field. Three years later, the Hungarians Zoltán Bay and György Szigeti patented a silicon carbide device capable of emitting white or white light tending to yellow or green, depending on the impurity present. In 1951, Kurt Lehovec, Carl Accardo and Edward Jamgochian understood for the first time the real functioning of a diode capable of emitting light, using silicon carbide crystals and as an electrical source a battery and a pulse generator, comparing in the following two years the results obtained with others obtained by varying the type of crystals and their purity.

In 1955, Rubin Braunstein, of the Radio Corporation of America, obtained infrared light emission from gallium arsenide (GaAs) and other semiconductors, such as gallium antimonide (GaSb), indium phosphide (InP) and silicon germanide (SiGe), both at room temperature and 77 kelvin. Two years later, he demonstrated that rudimentary devices could be used to communicate over short distances; such devices would then be used in optical communications.

In September 1961, at Texas Instruments, James R. Biard and Gary Pittman discovered that a tunnel effect diode with gallium arsenide substrate was capable of emitting infrared light with a wavelength of 900 nanometers. In October, they demonstrated the effectiveness of communication between this diode and an electrically isolated photodetector. On August 8, 1962, Biard and Pittman applied for a patent entitled “Semiconductor radiant diode”: a diode with a p-n junction and diffuse zinc, with the cathode spaced to allow an efficient emission of infrared light when the device is in the so-called direct polarization. After receiving applications also from General Electric, Radio Corporation of America, IBM, Bell Laboratories and MIT Lincoln Laboratory, the U.S. patent office gave the two inventors the patent for the infrared light emitting diode in gallium arsenide, the first real LED for practical use. Immediately afterwards, Texas Instruments launched a project for their realization and, in October 1962, the company announced the commercial production of LEDs with gallium arsenide crystalline structure capable of emitting light with a wavelength of 890 nanometers.

The first emission LED in the visible spectrum was developed at General Electric by Nick Holonyak Jr. who published an article on December 1, 1962. Having obtained an LED with emission in the visible spectrum, then the realization for the first time of an electronic component capable of emitting light perceptible by humans, made Holonyak to the public eye as the “father” of the LED. In 1972, George Craford, a former student of Holonyak, created the first yellow light LED and improved the light output of red and red-orange LEDs by a factor of ten. Four years later, T. P. Pearsall created the first high efficiency and luminescence LED, resulting in new semiconductor compounds specifically suitable for fiber optic transmissions.

Mass Use

The first commercial LEDs were used to replace some incandescent and neon lamps, for seven-segment displays, optoisolators, expensive laboratory equipment first and then to switch to calculators, televisions, radios, phones and more. Both infrared and visible LEDs were still extremely expensive, however, in the order of two hundred dollars each, so they were used relatively little. Starting in 1968, Monsanto Company was the first in the world to start mass production of LEDs in the visible, using gallium, arsenic and phosphor to make red LEDs suitable as indicators (arrows, numbers, etc.). Later, other colors began to be available and LEDs began to appear on various other equipment and devices. In the seventies, LED devices were produced and marketed for less than five cents each. These diodes were made from semiconductor chips manufactured with the planar growth process devised by Jean Hoerni at Fairchild Semiconductor. The combination of this process with innovative encapsulation methods enabled Fairchild, under the leadership of optoelectronics pioneer Thomas Brandt, to significantly reduce production costs, opening the way for all other manufacturers.

The first LEDs had a metal casing similar to that used for transistors, with a glass lens for the passage of photons. Subsequently, they switched to transparent plastic casings, of various shapes and often with colors corresponding to the color of the light emitted. In the case of infrared LEDs, the tint can be applied to achieve the opposite effect, i.e. blocking the visible light output. Specific encapsulations are then designed for the efficient heat dissipation of high power LEDs.

The advent of blue LEDs

The first blue-violet LED was made with gallium nitride (GaN) doped with magnesium at Stanford University in 1972 by Herb Maruska and Wally Rhines, PhD students in materials science and engineering. The previous year, a former colleague from Maruska, Jacques Pankive, together with Ed Miller, at the Radio Corporation of America, obtained for the first time blue electroluminescence through gallium nitride but with zinc doping: from it, then, they obtained the first gallium nitride diode to emit green light. In 1974, Maruska, Rhines and Professor David Stevenson received the patent for their invention. In the 1970s, no practical use could be found for gallium nitride diodes doped with magnesium, and research slowed down, only to return decades later with the development of blue LEDs and laser diodes.

In August 1989, Cree was the first company to market blue silicon carbide LEDs, so with an indirect prohibited band that makes the device very inefficient. Also in the late 1980s, key milestones in the epitaxial growth of gallium nitride with doping of acceptors brought optoelectronic devices into the modern era. On this basis, in 1991 Theodore Moustakas, of Boston University, devised a method for the production of high luminescence blue LEDs through a two-step process, obtaining a patent six years later.

In 1993, with a growth process similar to that of Moustakas, Shuji Nakamura, from Nichia, also produced a high luminescence blue LED. Both Moustakas and Nakamura received a patent and this generated confusion as to who was the real inventor of the blue gallium nitride LED, in fact Moustakas discovered his method first but his registration at the patent office was later than Nakamura’s one. The possibility of producing blue LEDs on an industrial scale opened up the development of new technologies and applications in the following decades, so much so that Nakamura received the Millennium Technology Award in 2006 and in 2014, together with Hiroshi Amano and Isamu Akasaki, the Nobel Prize for physics.

In parallel, in Nagoya, Isamu Akasaki and Hiroshi Amano himself worked on the development of a method to grow gallium nitride on a sapphire substrate, doped with acceptors, and on demonstrating the increased efficiency of LEDs made with this process. In 1995, at the University of Cardiff, Alberto Barbieri studied the efficiency and reliability of high luminescence LEDs with a structure formed by layers of aluminum phosphide, gallium and indium (AlGaInP) and gallium arsenide (GaAs), with a “transparent contact” that is a transparent film of indium tin oxide (also known as ITO, Indium tin oxide).

Between 2001 and 2002, methods of growth of gallium nitride on silicon were successfully demonstrated and, in January 2012, Osram found a way to produce industrial quantity LEDs in indium gallium nitride (InGaN) grown on silicon substrates. At least until 2017, manufacturers used silicon carbide substrates, although the most common one remained sapphire because it has properties very similar to gallium nitride, which reduces the formation of defects in its crystalline structure during growth.

At the end of the decade, Samsung and the University of Cambridge carried out research on gallium nitride LEDs grown on silicon substrate, initially followed by Toshiba, but then stopped the research. Some opted for epitaxial growth by nano-printing lithography, while others opted for multi-layer growth to reduce the differences between crystalline structures and thermal expansion rate, in an attempt to avoid chip breakage at high temperatures, decrease heat emission and increase light efficiency.

White LEDs and use in lighting

White light can be produced by using different colored LEDs together: one red, one green and one blue; however, the color quality will be low because only three narrow bands of the visible spectrum are used. A better method is to use a high efficiency blue LED, using the properties of phosphor to produce white light. In these devices, when the blue LED light hits an overlying layer of phosphor, doped with yttrium garnet, aluminum (YAG) and cerium (Y3Al5O12:Ce), it produces a yellow fluorescent light: the overall effect of blue and yellow light has a very wide bandwidth and is therefore perceived as white light by the human eye, with a higher color rendering index than white obtained by combining red, green and blue LEDs.

As with its predecessors, the first white LED was expensive and inefficient. However, the improvement in light output grew exponentially: the latest developments and research have been carried out by Japanese, Korean and Chinese companies, such as Panasonic, Nichia, Samsung and Kingsun. The trend of this growth is defined by Haitz’s Law (pictured), which takes its name from Roland Haitz.

The light emission and efficiency of blue-violet LEDs grew and at the same time the cost of the devices fell, allowing to produce white LEDs with relatively high power, potentially suitable to replace traditional lighting.

In the decade of the two thousand, experimental white LEDs produced 303 lumens per watt of electrical current input, with a life span of up to one hundred thousand hours, although those on the market stopped at 223 lumens per watt. Compared to an incandescent lamp, it was therefore obtained a substantial increase in electrical efficiency at the same price and, sometimes, at an even lower cost.