On this page
The LED bulb in your room started with a problem that defeated research labs for three decades: making a bright blue light-emitting diode. Red and green LEDs had existed for a long time, but without blue you cannot mix white light. This post covers how Isamu Akasaki, Hiroshi Amano and Shuji Nakamura solved it, and the Class 12 semiconductor physics that explains why blue was so hard.
Who are the three laureates?
The 2014 Nobel Prize in Physics press release gives the prize to all three "for the invention of efficient blue light-emitting diodes which has enabled bright and energy-saving white light sources". It lists:
Isamu Akasaki, born 1929 in Chiran, Japan, PhD 1964 from Nagoya University.
Hiroshi Amano, born 1960 in Hamamatsu, Japan, PhD 1989 from Nagoya University.
Shuji Nakamura, born 1954 in Ikata, Japan, PhD 1994 from the University of Tokushima, and at the University of California, Santa Barbara at the time of the prize.
Akasaki worked with Amano, then a PhD student, at Nagoya University. Nakamura worked at Nichia Chemicals, a small company in Tokushima.
Why was the blue LED so hard?
An LED is a p-n junction. When a voltage pushes electrons and holes together in the active layer, they recombine and give out light. The Nobel Committee's popular science background explains that the wavelength depends entirely on the semiconductor, and blue light can only be produced in some materials.
Gallium nitride (GaN) was a strong candidate, but two problems seemed hopeless. Nobody could grow GaN crystals of high enough quality, and it was almost impossible to make p-type GaN. Many researchers switched to zinc selenide. All three laureates stuck with gallium nitride.
How did they solve it?
The popular background gives the steps:
Better crystals. In 1986 Akasaki and Amano grew high-quality GaN by putting a layer of aluminium nitride on a sapphire base first. Nakamura began work in 1988 and two years later found his own route: a thin GaN layer grown at low temperature, then further layers at higher temperature.
A p-type layer. At the end of the 1980s Akasaki and Amano noticed, by chance, that their material glowed more brightly when examined in a scanning electron microscope. The electron beam was improving the p-type layer. Nakamura worked out why: hydrogen was blocking it, and the beam removed the hydrogen. In 1992 he found a simpler, cheaper method, heating the material.
First bright blue LEDs. In 1992 Akasaki and Amano presented their first diode emitting bright blue light. The Scientific Background adds that in 1994 Nakamura and co-workers reached a quantum efficiency of 2.7% using a layered structure of indium gallium nitride and aluminium gallium nitride.
The Committee writes that the laureates built their own equipment and carried out thousands of experiments, and that most of the time they failed.
A worked example: band gap and colour
LEDs are in the JEE (Main) 2026 syllabus, which lists the I-V characteristics of an LED in Unit 19, though the rationalised NCERT Class 12 semiconductor chapter now stops at diodes as rectifiers.
A photon's energy is E = hc/λ. Using the NIST value of h, the speed of light and 1 eV = 1.602 × 10⁻¹⁹ J, hc ≈ 1240 eV nm, so E (in eV) ≈ 1240 ÷ λ (in nm).
Blue light at 450 nm: E ≈ 1240 ÷ 450 ≈ 2.76 eV.
Red light at 650 nm: E ≈ 1240 ÷ 650 ≈ 1.91 eV.
Pure GaN: the Scientific Background gives its band gap as 3.4 eV. Then λ ≈ 1240 ÷ 3.4 ≈ 365 nm, which is ultraviolet, as the document states.
So blue needs a larger band gap than red, and pure GaN overshoots into the ultraviolet. Alloys such as indium gallium nitride, which the laureates developed, bring the light into the visible range.
Why does this matter for lighting?
The press release gives these figures from 2014: the best white LEDs had reached just over 300 lumens per watt, compared with 16 for ordinary bulbs and close to 70 for fluorescent lamps. LEDs can last up to 100,000 hours, against 1,000 for incandescent bulbs and 10,000 for fluorescent lights. White light is made either by using blue light to make a phosphor glow red and green, or by combining red, green and blue LEDs.
As a rough check: an 800-lumen lamp at 16 lumens per watt needs 800 ÷ 16 = 50 W. At 70 lumens per watt it needs about 11 W. The 300 lm/W figure was a record, so an everyday LED lamp giving 800 lumens will draw more than 800 ÷ 300 ≈ 2.7 W.
Common mistakes students make
Using λ in metres with 1240. The shortcut E ≈ 1240/λ needs λ in nanometres and gives eV.
Thinking colour comes from coloured plastic. In an LED, the band gap sets the colour.
Forgetting that LEDs need forward bias. No forward current, no recombination, no light. Our post on p-n junction diode characteristics covers the I-V curve.
For the wider topic, see our guide to semiconductors and logic gates. If band gaps and recombination still feel abstract, the Fermi Concept Library covers every concept in physics, each with its own lesson.
What can students take from the way these three worked?
Do not drop a good idea because others have. Gallium nitride was out of fashion. They kept going.
Notice the odd result. A brighter glow under an electron microscope was the clue to the p-type problem.
Expect to fail often. Thousands of experiments, most of them failures, came before success.
Take the colour of any LED around you, look up its wavelength on the packet or datasheet, and work out its photon energy with E ≈ 1240/λ.
