10.3 Light-emitting diodes (LEDs)

Semiconductor materials for LEDs

The semiconductor material you choose for an LED determines almost everything about it: what color light it emits, how efficiently it operates, and how well it performs. That's because the material's electronic band structure dictates how electrons and holes recombine and whether that recombination produces useful light.

Direct vs indirect bandgap semiconductors

This distinction is fundamental to understanding why certain materials work for LEDs and others don't.

In a direct bandgap semiconductor (like GaAs or InP), the conduction band minimum and valence band maximum occur at the same crystal momentum (k-value). An electron can drop straight from the conduction band into the valence band and release its energy as a photon. This makes radiative recombination efficient and is exactly what you want in an LED.

In an indirect bandgap semiconductor (like Si or Ge), the conduction band minimum and valence band maximum sit at different k-values. For an electron to recombine with a hole, it needs a phonon (a lattice vibration) to supply the momentum difference. This two-particle process is far less probable, so most recombination events release heat instead of light. That's why silicon makes terrible LEDs despite being the dominant material in electronics.

Compound semiconductors in LEDs

Most LED materials are III-V compound semiconductors, formed by combining elements from groups III and V of the periodic table. Common examples:

• GaAs and AlGaAs: red and infrared emission
• InGaN: blue and green emission
• AlInGaP: red, orange, and yellow emission
• GaN: UV and blue emission

A major advantage of compound semiconductors is bandgap engineering. By adjusting the composition ratio of an alloy (for example, changing the aluminum fraction $x$ in $Al_xGa_{1-x}As$), you can continuously tune the bandgap energy and therefore the emission wavelength. This compositional flexibility is what allows LEDs to cover the full visible spectrum and beyond.

LED structure and design

The internal architecture of an LED determines how well it confines carriers, how efficiently they recombine, and how much light ultimately escapes. LED designs have evolved from simple PN junctions to sophisticated quantum well structures.

PN junction in LEDs

The simplest LED is a forward-biased PN junction.

• The P-type region has an excess of holes; the N-type region has an excess of electrons.
• At the junction, a depletion region forms with a built-in electric field.
• Under forward bias, the external voltage reduces the barrier, injecting electrons into the P-side and holes into the N-side.
• In the region near the junction, these injected carriers recombine. In a direct bandgap material, a significant fraction of this recombination is radiative, producing photons with energy close to the bandgap.

Simple PN junction LEDs work, but they suffer from poor carrier confinement. Injected carriers can diffuse away from the active region before recombining, which limits efficiency.

Heterojunction LEDs

Heterojunction LEDs solve the confinement problem by using layers of different semiconductor materials with different bandgaps.

A classic example is the AlGaAs/GaAs double heterojunction:

1. A thin GaAs layer (the active region, with a smaller bandgap) is sandwiched between two AlGaAs layers (with a larger bandgap).
2. The bandgap discontinuities at both interfaces create energy barriers that confine electrons and holes within the GaAs layer.
3. Carriers are forced to recombine in this narrow active region rather than diffusing away.

This confinement dramatically increases the radiative recombination rate and reduces carrier leakage, boosting efficiency well beyond what a simple homojunction can achieve.

Quantum well LEDs

Quantum well LEDs take heterojunction confinement a step further by making the active layer extremely thin, typically just a few nanometers.

• A thin layer of lower-bandgap material (e.g., InGaN) is sandwiched between wider-bandgap barriers (e.g., GaN).
• At this thickness, quantum confinement becomes significant: carrier energy levels become discrete rather than continuous, modifying the density of states and the emission characteristics.
• The emission wavelength can be precisely tuned by adjusting both the well thickness and the material composition.

Most modern high-performance LEDs use multiple quantum wells (MQWs), stacking several well/barrier pairs to increase the total active volume and light output. Quantum well designs offer better efficiency, narrower emission spectra (improved color purity), and greater wavelength tunability compared to bulk active regions.

Light emission in LEDs

Light emission in an LED comes down to one process: the radiative recombination of electrons and holes. Everything else in LED design exists to make this process happen as efficiently as possible.

Radiative recombination of carriers

When an LED is forward biased, electrons are injected into the active region from the N-side and holes from the P-side. When an electron in the conduction band recombines with a hole in the valence band, the energy difference is released as a photon.

Not all recombination produces light, though. Non-radiative recombination processes compete with photon emission:

• Defect-assisted (Shockley-Read-Hall) recombination: Carriers recombine through trap states associated with crystal defects, releasing energy as heat.
• Auger recombination: The recombination energy is transferred to a third carrier instead of producing a photon. This becomes significant at high carrier densities.

Reducing crystal defects through better epitaxial growth and optimizing carrier density in the active region are key strategies for maximizing the fraction of recombination events that produce light.

Emission wavelength and color

The wavelength of light emitted by an LED is set by the bandgap energy of the active region material. The relationship is:

$E = \frac{hc}{\lambda}$

where $E$ is the photon energy, $h$ is Planck's constant, $c$ is the speed of light, and $\lambda$ is the wavelength. A larger bandgap produces shorter-wavelength (higher-energy) photons; a smaller bandgap produces longer-wavelength photons.

Common material-color pairings:

White LEDs are typically made one of two ways:

• Phosphor conversion: A blue InGaN LED is coated with a yellow-emitting phosphor (commonly cerium-doped YAG). The combination of blue LED light and broad yellow phosphor emission appears white to the eye.
• RGB mixing: Separate red, green, and blue LEDs are combined. This approach offers tunable color temperature but is more complex.

Efficiency of light emission

LED efficiency is described at two levels:

• Internal quantum efficiency (IQE): The fraction of electron-hole recombination events in the active region that produce photons (rather than heat). High-quality materials with few defects can achieve IQE values above 90%.
• External quantum efficiency (EQE): The fraction of injected electrons that result in photons actually escaping the device. EQE is always lower than IQE because some generated photons are reabsorbed or trapped inside the chip by total internal reflection.

The relationship is: $EQE = IQE \times \eta_{extraction}$, where $\eta_{extraction}$ is the light extraction efficiency.

Strategies to improve EQE focus on getting more photons out of the chip:

• Surface texturing to randomize photon angles and reduce total internal reflection
• Photonic crystal patterns on the surface to redirect trapped light
• Transparent substrates to eliminate absorption losses
• Shaped chips (e.g., truncated inverted pyramids) to increase escape probability

Electrical characteristics of LEDs

Understanding the electrical behavior of LEDs is essential for designing proper drive circuits and getting the most light out of the device without damaging it.

Current-voltage relationship

Like any diode, an LED has a nonlinear I-V characteristic. Very little current flows until the applied voltage approaches the forward voltage (also called the turn-on voltage), after which current rises steeply.

The forward voltage is closely related to the bandgap of the semiconductor material. Typical values:

• Red (AlGaAs/AlInGaP): ~1.7–2.0 V
• Green (InGaN): ~2.2–3.0 V
• Blue (InGaN): ~2.8–3.5 V

The I-V curve follows the Shockley diode equation:

$I = I_s \left( e^{\frac{qV}{nkT}} - 1 \right)$

where $I_s$ is the reverse saturation current, $q$ is the electron charge, $V$ is the applied voltage, $n$ is the ideality factor (typically 1–2 for LEDs), $k$ is Boltzmann's constant, and $T$ is the absolute temperature.

Because of this exponential relationship, small changes in voltage produce large changes in current. That's why LEDs are almost always driven with a current source or a resistor in series, not a raw voltage source.

Series resistance effects

Real LEDs have a series resistance $R_s$ that comes from two sources: the bulk resistivity of the semiconductor layers and the contact resistance at the metal-semiconductor interfaces. At operating currents, this resistance causes a voltage drop $V = IR_s$ that doesn't contribute to light emission, reducing overall efficiency.

To minimize series resistance effects:

• Current spreading layers distribute current uniformly across the chip. Transparent conductive oxides like indium tin oxide (ITO) are commonly used, especially on the P-side of GaN-based LEDs where the P-type material has low conductivity.
• Optimized contact geometry ensures low-resistance ohmic contacts between the metal electrodes and the semiconductor.

Capacitance and high-frequency behavior

The PN junction of an LED stores charge in its depletion region, creating a junction capacitance. This capacitance limits how fast the LED can be switched on and off, which matters for applications like optical data communication.

The junction capacitance depends on:

• Junction area (larger area = more capacitance)
• Doping concentrations (higher doping = thinner depletion region = more capacitance)
• Applied voltage (forward bias increases stored charge)

For high-speed operation (e.g., visible light communication), the LED and its driver circuit need careful design. Techniques include pre-emphasis (overdriving the signal edges to speed up transitions), equalization in the receiver, and impedance matching to minimize reflections. Reducing the active area of the LED also helps by lowering the total capacitance, though this trades off against total light output.

Optical properties of LEDs

The optical properties of an LED determine how the generated light escapes the chip and how it's distributed in space. Even a chip with excellent IQE will perform poorly if most of the light gets trapped inside.

Emission pattern and directionality

Most standard LEDs exhibit a Lambertian emission pattern, where the intensity varies as the cosine of the angle from the surface normal:

$I(\theta) = I_0 \cos(\theta)$

This produces a broad, hemispherical distribution with a wide viewing angle (about 120° full width at half maximum). For applications needing a narrow beam (spotlights, fiber optic coupling, automotive headlights), external optics are required:

• Reflectors redirect side-emitted light forward
• Lenses (molded into the package or added externally) focus the beam
• Photonic crystals patterned on the chip surface can shape the emission directionality at the source

Refractive index and light extraction

This is one of the biggest challenges in LED design. Typical III-V semiconductors have refractive indices around 2.5–3.5, while air has a refractive index of 1.0. This large mismatch means that light hitting the semiconductor-air interface at angles beyond the critical angle undergoes total internal reflection (TIR) and gets trapped inside the chip.

For a GaN LED ($n \approx 2.5$) emitting into air, the critical angle is only about 23°. A simple geometric calculation shows that only a small fraction of the generated light can escape through a flat surface.

Strategies to combat this:

• Encapsulation in epoxy or silicone ($n \approx 1.5$) reduces the refractive index step and increases the critical angle, letting more light out of the chip and into the encapsulant.
• Anti-reflection coatings minimize Fresnel reflection at the interface.
• Graded refractive index layers provide a smooth transition that reduces TIR.

Surface texturing for enhanced output

Surface texturing is one of the most effective ways to improve light extraction. Instead of a flat surface where trapped light bounces at the same angle repeatedly, a textured surface randomizes the propagation direction of reflected photons, giving them new chances to escape on subsequent bounces.

Common texturing approaches:

• Geometric patterns: Pyramids, cones, or hemispheres etched into the surface
• Photonic crystals: Periodic nanostructures that can both scatter light and create photonic bandgaps to redirect emission
• Random roughening: Created by wet or dry etching processes

The feature sizes are typically on the order of the emission wavelength (hundreds of nanometers) to effectively interact with the light. Fabrication methods include wet chemical etching, dry plasma etching, nanoimprint lithography, and self-assembled colloidal masks. The optimal texture design (shape, size, spacing, and arrangement) depends on the specific material system and target wavelength.

LED packaging and integration

The LED chip itself is just the starting point. Packaging turns a bare chip into a usable component by providing mechanical protection, electrical connections, thermal pathways, and optical shaping.

LED chip encapsulation

LED chips are encapsulated in a transparent material, typically epoxy (for low-cost, low-power LEDs) or silicone (for high-power LEDs that generate more heat).

The encapsulant serves multiple roles:

• Mechanical protection from moisture, dust, and physical damage
• Improved light extraction by reducing the refractive index mismatch at the chip surface (silicone has $n \approx 1.4–1.5$, much closer to the semiconductor than air)
• Optical shaping through the encapsulant geometry (dome, lens, or flat)

Silicone is preferred for high-power applications because it resists yellowing and thermal degradation far better than epoxy. The encapsulant shape matters too: a hemispherical dome ensures that light rays hit the encapsulant-air interface close to normal incidence, minimizing TIR at that second interface.

Thermal management in LEDs

LEDs are far more efficient than incandescent bulbs, but they still convert a significant fraction of input power to heat. This heat comes from non-radiative recombination in the active region and Joule heating ($I^2R$) in the series resistance. If the junction temperature rises too high, several problems occur:

• Light output drops (thermal droop)
• Emission wavelength shifts (redshift)
• Device lifetime decreases, sometimes dramatically

Thermal management strategies, in order of increasing complexity:

1. Heat sinks: Metal structures (aluminum or copper) with fins or other high-surface-area geometries that conduct heat away from the chip and dissipate it into the surrounding air.
2. Thermal interface materials (TIMs): Thermal pastes, pads, or phase-change materials that fill microscopic air gaps between the LED package and the heat sink, improving thermal contact.
3. Active cooling: Fans, heat pipes, or thermoelectric coolers for high-power LED systems where passive cooling isn't sufficient.

Proper thermal design is not optional for high-power LEDs. A well-designed thermal path from the junction to the ambient environment is just as important as the chip design itself.

LED arrays and displays

Integrating multiple LEDs into arrays enables applications from simple indicator panels to massive video displays.

Packaging approaches:

• Surface-mount device (SMD): Individual packaged LEDs are soldered onto a printed circuit board. This is the most common approach for general lighting and small displays.
• Chip-on-board (COB): Bare LED chips are mounted directly onto a substrate and wire-bonded, then covered with a shared phosphor/encapsulant layer. COB produces a uniform, high-density light source ideal for spotlights and downlights.
• Flip-chip: The LED chip is flipped upside down and bonded directly to the substrate through solder bumps, eliminating wire bonds. This enables higher packing density and better thermal performance.

Key design considerations for LED arrays and displays:

• Pixel pitch (center-to-center distance between pixels) determines the resolution and minimum viewing distance
• Uniformity requires careful binning of LEDs by brightness and color, plus calibration
• Color mixing can be achieved with multiple LED colors per pixel or with diffuser layers
• Driver circuitry must handle the current requirements of many LEDs while maintaining uniform brightness across the array

LED performance and applications

Efficiency and luminous efficacy

LED efficiency is quantified in several ways:

• External quantum efficiency (EQE): Photons emitted per electron injected. Combines IQE and extraction efficiency.
• Wall-plug efficiency (WPE): Total optical output power divided by total electrical input power. This is the most practical efficiency metric because it accounts for the voltage drop across the device.
• Luminous efficacy: Measured in lumens per watt (lm/W), this weights the optical output by the human eye's spectral sensitivity (the photopic response curve). A perfectly efficient green LED at 555 nm would have the highest possible luminous efficacy of 683 lm/W.

Modern commercial white LEDs routinely exceed 200 lm/W in laboratory settings, with production devices typically in the 150–200 lm/W range. For comparison, incandescent bulbs achieve about 15 lm/W and fluorescent tubes about 80–100 lm/W.

Reliability and lifetime of LEDs

LED lifetime is typically defined as the time for light output to fall to 70% of its initial value (designated L70). Most quality LEDs are rated for 50,000 hours or more at L70, which translates to over 5 years of continuous operation.

Factors that degrade LEDs over time:

• Junction temperature: The single biggest factor. Every 10°C increase in junction temperature roughly halves the lifetime.
• Humidity: Moisture can corrode contacts and degrade encapsulant materials.
• Electrical overstress: Excessive current or voltage spikes damage the active region.
• Material degradation: Gradual darkening of encapsulants, growth of crystal defects, and electromigration of contact metals.

Manufacturers use accelerated life testing to predict long-term reliability. LEDs are operated at elevated temperatures and currents to speed up degradation mechanisms, and the results are extrapolated to normal operating conditions using established models (like the Arrhenius equation for temperature-dependent degradation).

LEDs in lighting and displays

Lighting applications:

LEDs have largely replaced incandescent and fluorescent sources in residential, commercial, and industrial settings. Beyond energy savings, LEDs enable features that traditional sources can't match: instant on/off, easy dimming, tunable color temperature (warm white to cool white), and compact form factors. Smart lighting systems use LEDs with wireless controls for automated dimming, color adjustment, and occupancy-based switching.

Display applications:

• LCD backlighting: LEDs replaced cold-cathode fluorescent lamps (CCFLs) as the backlight source in LCD panels, enabling thinner displays, better color gamut, higher contrast through local dimming, and lower power consumption.
• Direct-view LED displays: Large-format displays (billboards, stadium screens) use arrays of discrete LEDs as the pixels themselves.
• Micro-LED displays: An emerging technology using arrays of micron-scale LED chips as individual pixels. Micro-LEDs promise higher brightness, wider color gamut, faster response times, and better energy efficiency than both LCD and OLED displays, though manufacturing challenges remain.

Specialized LED applications

LEDs have expanded well beyond lighting and displays into fields that exploit their specific spectral, switching, and size characteristics.

Horticulture: LED grow lights can be tuned to emit at the specific wavelengths that plant photosynthetic pigments absorb most efficiently (primarily red ~660 nm and blue ~450 nm). This spectral precision means less wasted energy compared to broadband light sources, and growers can adjust the light recipe for different growth stages.

Medical and biomedical: Blue LEDs (around 460 nm) are the standard treatment for neonatal jaundice, breaking down bilirubin in the skin. Red and near-infrared LEDs are used in photobiomodulation therapy for wound healing and pain management. In optogenetics research, LEDs activate light-sensitive proteins (opsins) in genetically modified neurons, allowing precise control of neural circuits.

UV applications: UV-C LEDs (around 265 nm) are used for water purification and surface sterilization, offering a mercury-free alternative to traditional UV lamps. UV-A LEDs cure adhesives, coatings, and dental composites.

Visible light communication (VLC): Because LEDs can be modulated at high frequencies (MHz range), they can transmit data wirelessly through modulated light. VLC systems, sometimes called Li-Fi, offer high bandwidth, inherent security (light doesn't pass through walls), and freedom from radio frequency interference. This is particularly useful in environments like hospitals and aircraft cabins where RF communication may be restricted.