How LED Lights Work — Science of Solid State Light

Walk into any home built in the last decade and you'll see them: tiny chips of semiconductor glowing brighter than the hot, glass-bulbed filaments they replaced. LEDs — light-emitting diodes — now produce the majority of the world's artificial light, slashing global lighting electricity demand by more than 40% compared with the incandescent era. But how does a small piece of crystal turn voltage directly into visible photons, with no heat-glowing wire in sight?

The answer sits at the intersection of quantum mechanics, materials science, and one of the most consequential Nobel Prizes of the 21st century.

The Core Idea: Electrons Falling, Photons Emitted

An LED — light-emitting diode — is a special kind of diode: a sandwich of two semiconductor layers joined at a p-n junction. This is the heart of how LEDs convert electrical current into photons.

Semiconductor doping basics:

• The n-type (negative) layer is doped with donor atoms (like phosphorus in silicon, or silicon in gallium nitride) that contribute extra electrons — negative charge carriers.
• The p-type (positive) layer is doped with acceptor atoms (like boron in silicon) that create "holes" — missing electrons that behave like positive charges and can move.

The recombination process: When you apply a forward voltage (bias) across the junction, a few things happen:

1. The electric field pushes electrons from the n-side toward the p-side, and holes from the p-side toward the n-side
2. At the junction, they meet and recombine
3. When an electron from the higher-energy conduction band drops down to fill a hole in the lower-energy valence band, it releases energy
4. In an LED, most of that energy comes out as a single photon — a particle of light

The energy released is equal to the bandgap of the semiconductor (the energy difference between conduction and valence bands). This is a quantum-mechanical process, not thermal — photon wavelength = 1,240 eV·nm / bandgap energy.

This process is called electroluminescence, and unlike a filament bulb it doesn't require heating anything to thousands of degrees. The light is produced directly by quantum transitions inside the crystal lattice, with essentially no wasted heat. That's why LEDs run cool and can last for decades without degradation.

Source: Schubert, E.F. (2018). Light-Emitting Diodes, 3rd ed. Cambridge University Press — foundational reference on LED physics and materials science.

Why LEDs Are So Ridiculously Efficient

An old-school incandescent bulb is, scientifically speaking, a tiny electric heater that happens to glow. It runs a tungsten filament up to ~2,700°C, and the resulting blackbody radiation includes a sliver of visible light. Only about 5% of the electrical energy becomes visible photons; the rest is infrared (heat). The physics is against it: blackbody radiation peaks in the infrared for anything below ~5,000°C.

LEDs invert that ratio entirely. Because photons are produced directly from electron-hole recombination at the bandgap energy, a well-designed white LED can convert 40–80% of input electrical energy into visible light (measured as wall-plug efficiency). The remainder is lost mostly to:

• Non-radiative recombination (phonon emission — vibrations in the crystal — instead of photons)
• Phosphor conversion losses (if using phosphor-based white LEDs)
• Series resistance and voltage losses in the driver circuit
• Not bulk heating (there's almost no heat generated)

This is a fundamental advantage: LEDs don't have to heat up to produce light. The quantum transitions happen at room temperature.

The practical numbers are striking:

• Incandescent bulb: ~15 lumens per watt (lm/W) of visible light, consumes ~60 W for a bright room light, runs at 2,700–2,900 K color temperature (warm white), lasts ~1,000–2,000 hours.
• Compact fluorescent (CFL): ~60 lm/W, ~15 W for equivalent brightness, 20–40% flicker at 100–120 Hz, contains mercury, lasts ~8,000–10,000 hours, warm-up time 30–60 seconds.
• Modern white LED: 100–200 lm/W, ~8–12 W for equivalent brightness, no flicker, no mercury, lasts 25,000–50,000 hours (L70 rating), instant on, available in 2,200–6,500 K color temperatures (warm to daylight).
• Laboratory LED (peak efficiency): 250+ lm/W demonstrated under lab conditions (Philips, 2015), but not in commercial products yet due to heat dissipation and cost challenges.

Source: U.S. Department of Energy (2024). "Lighting Facts: Efficacy and Performance of LEDs." Office of Energy Efficiency and Renewable Energy.

Color Comes from the Bandgap

The color of an LED's light is set by the bandgap of its semiconductor — the energy difference an electron crosses when it recombines. A bigger bandgap means a higher-energy (shorter-wavelength, bluer) photon. Smaller bandgap → redder, lower-energy, longer-wavelength photon.

The physics: When an electron recombines with a hole, the energy released goes into a photon. That photon's wavelength is determined by:

E (photon) = h × f = h × c / λ = bandgap energy

Where h is Planck's constant (~6.63 × 10^-34 J·s), c is the speed of light, and λ is wavelength. Conveniently, h × c ≈ 1,240 eV·nm, so:

λ (nm) ≈ 1,240 eV·nm / bandgap (eV)

Examples:

• A 1.4 eV bandgap material (silicon) emits at ~890 nm (infrared, not visible)
• A 2.0 eV bandgap material emits at ~620 nm (deep red)
• A 2.4 eV bandgap emits at ~515 nm (green)
• A 2.7 eV bandgap emits at ~460 nm (blue)
• A 3.5 eV bandgap emits at ~355 nm (UV, beyond human vision)

Semiconductor materials — different compounds have different bandgaps:

• GaAs (gallium arsenide) / AlGaAs → 1.4–2.2 eV → infrared and red (750–850 nm). Historically used for early LEDs and laser diodes. Still used in fiber-optic applications.
• GaP (gallium phosphide) / AlGaInP → 2.0–2.6 eV → red, orange, yellow, green (590–660 nm). Workhorse for 1980s–2000s colored indicators and traffic lights.
• InGaN (indium gallium nitride) / GaN (gallium nitride) → 2.4–3.5 eV → green, blue, violet (365–530 nm). The modern high-efficiency LED material; relatively new (1990s).

The historical problem: For decades (1960s–1990s), red, orange, yellow, and green LEDs existed at reasonable brightness. But blue LEDs were missing. Researchers tried for ~30 years to make efficient blue LEDs from various materials, but none worked well. Without blue, you cannot make white light (white is a mixture of red, green, and blue). Displays, backlights, and general lighting remained out of reach.

The Nobel-Winning Blue LED — A 30-Year Problem Solved

Without blue light, you cannot make white light efficiently. For ~30 years (1960s–1990s), researchers tried and failed to create an efficient blue LED. The problem: gallium nitride (GaN) seemed intractable to work with. It was incredibly hard to grow high-quality crystals, even harder to dope with p-type dopants (acceptors).

The breakthroughs:

• Isamu Akasaki and Hiroaki Amano (Meijo University) — in the early 1990s, they discovered that high-temperature GaN growth using an aluminum nitride (AlN) buffer layer could produce high-quality crystals. They also figured out p-type doping of GaN by irradiating Mg-doped GaN with electron beam, activating the dopants.
• Shuji Nakamura (Nichia Corporation) — independently developed a different approach using MOCVD (metal-organic chemical vapor deposition) and discovered that thermal annealing could activate p-type doping in GaN. He demonstrated the first practical blue LED in 1993 and white LEDs in the late 1990s.

Their work won the 2014 Nobel Prize in Physics — Akasaki, Amano, and Nakamura — with the Nobel committee citing it as "in the spirit of Alfred Nobel — for the greatest benefit to mankind."

Why such a big deal? Because once blue LEDs existed, white LEDs followed (using phosphor conversion or RGB mixing). And with efficient white LEDs, energy-efficient general lighting became possible for a planet of 8 billion people. The lighting sector alone now consumes ~15% of global electricity (down from ~20% before LEDs); the LED transition has saved hundreds of billions of dollars in energy costs and prevented 2+ gigatons of CO2 annually by 2025.

Source: Nobel Prize Committee (2014). "The Nobel Prize in Physics 2014: For the invention of efficient blue light-emitting diodes which has enabled bright and energy-saving white light sources." Nobelprize.org.

Making White Light from a Semiconductor

You can't directly create a white-light photon from a single semiconductor bandgap; white requires a mixture of wavelengths. There are two main approaches:

1. Phosphor conversion (98% of commercial white LEDs) A high-power blue LED (typically InGaN, 450–460 nm) is mounted on a phosphor-coated dome or dome-shaped cavity.

• Some blue photons pass through the phosphor directly
• Others are absorbed by the phosphor (e.g., cerium-doped YAG — yttrium aluminum garnet) and re-emitted as broadband yellow light (500–700 nm)
• The mixture of passing blue + re-emitted yellow light looks white to the human eye (a process called the Stokes shift — the phosphor absorbs a high-energy photon and emits a lower-energy one, losing the difference as heat)

Advantages: Simple, cheap, one LED source, reliable. Disadvantages: Phosphor converts ~70–85% of absorbed photons to light (rest becomes heat); phosphor degrades over 30+ years, making LEDs slightly less blue-saturated with age; color quality depends on phosphor choice.

2. RGB mixing (specialized applications) Independent red, green, and blue LEDs are combined in a single fixture. A driver adjusts the current through each LED to create any color on the CIE color diagram.

Advantages: Full color control, better color rendering (CRI can exceed 95), no phosphor degradation. Disadvantages: Three separate LED chips, three separate drivers, more expensive, harder to tune, thermal management complexity.

Advanced approach (emerging, 2024–2026): Multi-phosphor systems add a red phosphor (e.g., KSF — K2SiF6:Mn4+) alongside yellow YAG. This boosts CRI above 95 and better replicates incandescent color rendering. These bulbs are more expensive (~$5–$10 vs $1–$2 for standard phosphor LEDs) but are increasingly common in high-end residential lighting.

Color Rendering Index (CRI): CRI measures how naturally colors appear under a light source (scale: 0–100, with 100 = perfect, like noon daylight or incandescent).

• Incandescent bulbs: CRI ≈ 100 (reference standard)
• Basic LED with standard YAG phosphor: CRI ≈ 70–80
• Higher-quality LED with multi-phosphor: CRI ≈ 90–95
• Humans can detect CRI difference above ~80 in side-by-side comparisons; below 80, skin tones look slightly off.

Source: Ohno, Y. (2000). "Color Rendering of Phosphor-Converted LEDs." Journal of the Illuminating Engineering Society, 29(1): 100–110. — Foundational paper on phosphor LED color science; still cited in 2026.

Comparison: Incandescent vs CFL vs LED

Property	Incandescent	CFL	LED
Efficacy (lm/W)	~15	~60	100–200
Energy to light	~5%	~20%	40–80%
Typical lifespan	1,000 h	8,000 h	25,000–50,000 h
Warm-up time	Instant	30–60 s	Instant
Contains mercury	No	Yes	No
Dimmable	Yes	Limited	Yes (with driver)
Cost (10-year total)	High	Medium	Lowest

Over a decade of use, an LED bulb typically saves 80–90% of the electricity cost of an equivalent incandescent, while lasting 25–50 times longer.

Efficiency Challenges and Thermal Management

Although LEDs are much more efficient than incandescents, there are real-world limits and challenges:

Droop (efficiency loss at high current): As you push more current through an LED to make it brighter, the efficiency actually decreases. This phenomenon, called droop, is a major research area. Causes:

• Current overflow (not all electrons recombine; some overshoot)
• Heating of the junction, which reduces bandgap and shifts emission wavelength
• Saturation of available states at the junction

For a high-power LED, efficiency might drop from 80% at 100 mA to 50% at 1 A. This is why high-power LED fixtures use multiple smaller LEDs rather than one massive one.

Phosphor losses (for white LEDs): The Stokes shift (absorbing a blue photon and emitting a yellow one) is fundamentally lossy. ~20–30% of the blue photon energy is lost as heat in the phosphor. This is a hard limit for phosphor-conversion white LEDs; you can't exceed ~50–60% wall-plug efficiency with this approach alone.

Thermal management: Even 20–40% waste heat is significant. A 10 W LED bulb dissipates 2–4 W as heat. That heat must escape through the base and into the room, or the junction temperature will rise, reducing efficiency and lifetime. Poor thermal design (bad heat sink, plastic bases) causes LEDs to dim and fail early despite the chip being capable of 50,000 hours.

2026 research frontiers:

• Perovskite LEDs: Lab devices (2024–2025) show promise for even higher efficiency and tunable color, but stability is poor (perovskites degrade in moisture).
• Micro-LEDs: Tiny LEDs (<100 micrometers) for high-density displays (AR glasses, smartwatches). More efficient per unit area, but manufacturing yield is currently too low for cost-effective mass production.
• Laser diodes as light sources: Some researchers are exploring blue laser diodes + phosphor (instead of blue LEDs + phosphor) for even higher brightness and efficiency, but lasers are directional (wasteful for diffuse room lighting).

Source: Pust, P., et al. (2015). "Narrow-band red-emitting Sr[LiAl3N4]:Eu2+ as a next-generation LED-phosphor material." Nature Materials, 14(5): 486–493.

OLEDs: The Organic Alternative

If LEDs are inorganic crystals, OLEDs (organic light-emitting diodes) use thin films of carbon-based molecules (polymers or small molecules) as the light-emitting material. Structurally, they work the same way (organic material in a p-n junction, electron-hole recombination produces photons), but the application is fundamentally different.

Key differences:

• Self-emissive: Each OLED pixel is its own light source. No backlight needed (unlike LCD displays).
• Blacks are true black: The pixel is simply turned off, so no light is emitted. Contrast ratio is infinite (not the "dark gray" of backlit displays).
• Thin and flexible: OLED films are just micrometers thick, so they can be printed on flexible plastic substrates.

Advantages: Superior image quality (infinite contrast), thin/flexible form factor, power-efficient for dark images (since off pixels use zero power).

Disadvantages: Shorter lifetime (8,000–20,000 hours before brightness drops significantly, vs 50,000+ for LEDs), blue OLED degradation (blue organic emitters are chemically unstable), higher cost, burn-in risk (if a static image is displayed for hours, that pixel dims permanently).

OLEDs currently dominate high-end smartphones (iPhone, Samsung Galaxy, Google Pixel) and premium TVs (LG OLED TVs are the standard for film producers). But they're not yet suitable for general lighting because of cost and lifetime. The same GaN semiconductor physics powers both traditional LEDs and emerging perovskite LED displays.

Why This Matters Beyond Your Living Room

Lighting once consumed roughly 20% of global electricity (as of 2015). The LED transition is one of the rare technological shifts that simultaneously cuts cost, energy use, and carbon emissions. The numbers are staggering:

• Global energy savings: LED adoption prevented ~630 megatons of CO2 emissions in 2023 alone (IEA, 2024), equivalent to taking ~200 million cars off the road for a year.
• Economic savings: Households and businesses save ~$40 billion annually on lighting electricity costs.
• Material science spillover: The GaN semiconductor technology pioneered for blue LEDs now powers:
  • Power electronics: GaN transistors in fast chargers (USB-C 65 W+ chargers use GaN for efficiency)
  • Radio frequency: Military radar and 5G infrastructure rely on GaN amplifiers
  • Renewable energy: Grid-scale solar inverters and EV chargers use GaN for efficiency
  • LED grow-lights: Vertical farms and indoor agriculture rely on precisely tuned LED spectra (red + far-red for photosynthesis)
  • UV-C disinfection: Hospital-grade sterilization using 255–280 nm UV LEDs (discovered after blue LEDs proved feasible)
  • LiFi (light-fidelity): Experimental wireless communication using modulated LED light (not yet commercially viable, but the physics works)

For more on capturing solar energy with the same semiconductor physics in reverse, see our explainer on how solar panels work. And to understand the electricity grid that powers all those bulbs, see how electricity grids work.

Frequently Asked Questions

Q: Why do LEDs need a driver or resistor?
A: LEDs are current-driven devices with a steep current-voltage curve. Below ~1.8–2.0 V, virtually no current flows. Above that, current surges exponentially with tiny voltage increases. Without a current-limiting driver or resistor, a small voltage rise can push 100+ mA through the LED instantly, overheating the junction and destroying it in milliseconds. This is why LED strings have integrated drivers/resistors and why connecting an LED directly to a 5 V power supply without current limiting will destroy it.

Q: Do LEDs really last 25,000 hours?
A: The LED chip itself can, yes. But the rated lifespan (e.g., "25,000 hours" on packaging) usually refers to L70 — the time until light output drops to 70% of the original brightness. This is when most users notice the bulb has dimmed. Real-world life depends heavily on thermal management; poorly cooled fixtures (bad heat sinks, plastic bases, sealed rooms with no airflow) fail much sooner (5,000–10,000 hours). Well-designed fixtures with proper heat sinks often exceed the rated life.

Q: Are LEDs harmful to the eyes?
A: Standard indoor LED lighting at recommended brightness levels (60–100 lux for home use, 300–500 lux for offices) is not harmful. However, there's growing research on blue-light exposure, especially in the evening when it suppresses melatonin and disrupts sleep. Staring directly at bright LED chips (or the sun through a magnifying glass) is harmful to vision and should be avoided. Bare high-power LEDs (>50 W) can cause eye damage. That said, a properly diffused LED bulb is no more harmful than a CFL or incandescent of equivalent brightness.

Q: Why are some LED bulbs flickery?
A: Flicker happens when the LED driver poorly rectifies AC current. In North America, AC alternates at 60 Hz; in Europe/Asia, 50 Hz. A cheap driver rectifies this AC to DC but doesn't smooth it well, so the LED actually turns fully on/off 100–120 times per second (for 60 Hz, it's two rectifications per cycle). The human eye can't consciously perceive flicker above ~60–90 Hz, but some people report subtle discomfort or headaches. Quality LED drivers use higher-frequency switching or better capacitive smoothing to eliminate flicker entirely.

Q: Can LEDs work in cold temperatures?
A: Yes, and they work better in the cold. Semiconductor efficiency increases at lower temperatures because thermal noise is reduced and recombination becomes more efficient. This is why LEDs dominate outdoor lighting in cold climates, freezer lighting, and aerospace. CFLs, by contrast, are sluggish in the cold (slow to reach full brightness) and can fail below ~–10°C.

Q: Why do blue LEDs look so bright compared with red ones at the same wattage?
A: Two effects: (1) The human eye's luminosity function peaks around 555 nm (green-yellow) — the eye is ~200 times more sensitive to green than to deep red at the same photon count. (2) Blue LEDs, through phosphor conversion to white, deliver light across the visible spectrum (including that peak-sensitivity green region), so they appear much brighter per watt. A 1 W red LED looks quite dim; a 1 W blue LED (converted to white) looks as bright as a 5–10 W incandescent.

The Takeaway

LEDs aren't magic — they're quantum mechanics in action. A semiconductor junction, biased to push electrons and holes into proximity, releases energy as photons when they recombine. Choose the right material (bandgap), and you can emit any color. Mix phosphors or combine red-green-blue chips, and you get white light. Efficiency is the key: no heating required, just quantum transitions. The blue LED breakthrough unlocked this technology for civilization; the green LED transition today is saving humanity billions in energy costs and megatons in carbon emissions. Understanding how LEDs work gives insight into solar cells, laser diodes, power electronics, and the entire semiconductor industry.

---

Read next:

• How Solar Panels Work — Photovoltaic Effect Explained
• Semiconductors Explained — The Science Behind Modern Electronics
• How Electricity Grids Work — Power Generation to Your Home