Simulating a GaAs Infrared LED at 870 nm

1. Introduction: Why GaAs emits infrared light

An LED operates in the opposite direction to a photodiode: rather than absorbing photons to generate a current, it uses an injected electrical current to generate photons. Under forward bias, electrons and holes enter the active semiconductor region. When they recombine radiatively, the energy released produces light with a wavelength determined primarily by the semiconductor bandgap.

\[ E_{\mathrm{ph}}=\frac{hc}{\lambda}\simeq E_g, \qquad \lambda\simeq\frac{1240\ \mathrm{eV\,nm}}{E_g\,[\mathrm{eV}]}. \]

Using a GaAs bandgap of approximately 1.42 eV gives a wavelength near 870 nm. The spectrum of a real LED is broader than a single wavelength because carriers occupy a distribution of energies; its precise peak and width depend on temperature, band structure and recombination physics.

The active GaAs region is surrounded by Al_0.3Ga_0.7As, which has a larger bandgap. The resulting heterojunction band offsets help confine injected carriers within the GaAs, where they can recombine and emit light. Unlike a laser diode, this example describes spontaneous emission: there is no optical cavity gain or lasing threshold.

2. Opening the preset and inspecting the LED structure

Click New simulation in the File ribbon, open GaAs demos and select GaAs infrared LED (870 nm). Choose a destination folder and save the example. OghmaNano opens the complete semiconductor stack in the Device structure tab; there is no need to construct the layers before running your first simulation.

2.1 Device layers and doping

Open the Layer editor from Device structure to inspect the semiconductor layers. The reference example uses the following four-layer structure, viewed from the top emission surface to the bottom contact:

The central GaAs active layer has a smaller bandgap than the two AlGaAs layers. Their conduction- and valence-band offsets help retain electrons and holes in the active region. The n⁺ GaAs backing provides a contact region but can absorb light propagating towards the back of the device; optical extraction is therefore a separate consideration from carrier confinement.

To inspect the electrical structure, open Doping/Ions from the Electrical ribbon. The Electrical parameters button in the main Device structure tab opens the material-parameter editor. OghmaNano can obtain alloy-dependent bandgaps, electron affinities, mobilities and recombination coefficients from the Lua electrical materials database. The AlGaAs alloy fraction is x = 0.30 in the Al_xGa_1−xAs layers.

3. Running the LED and examining its J–V characteristic

The preset is configured for an electrical forward-bias sweep from 0 to 1.7 V, with the optical illumination disabled: this is an electrically driven LED, not a photodiode. Click Run simulation in the File ribbon or press F9. The Terminal tab shows the solver progress as the applied voltage increases.

After the calculation, select Output and double-click jv.csv to plot current density against applied voltage. At low forward bias, injection is small. Once the applied voltage reduces the junction barrier, the current rises rapidly. The exact onset and magnitude depend on the contacts, mobilities, band offsets and recombination parameters chosen in the material models.

Unlike a resistor, the LED does not produce light in direct proportion to the applied voltage. Forward bias first controls how readily carriers enter the active region; the radiative recombination rate determines how many photons are generated. High forward current does not automatically imply high optical efficiency.

4. Understanding the energy bands and carrier injection

From the Output tab, double-click snapshots to open the Simulation snapshots viewer. Use the + button to add four lines, choosing Ec.csv (conduction band), Ev.csv (valence band), Fn.csv (electron quasi-Fermi level) and Fp.csv (hole quasi-Fermi level). Move the slider under the graph to inspect the device at different applied voltages.

At 0 V in the dark, the electron and hole Fermi levels coincide and are flat: the junction is in thermal equilibrium. The steps in the band edges at the GaAs/AlGaAs interfaces are the heterojunction band offsets. They form a lower-energy region for electrons and a corresponding confinement region for holes in the GaAs layer.

Under forward bias, the contact electrochemical potentials separate and carrier injection increases. The electron and hole quasi-Fermi levels become distinct; their positions relative to the conduction and valence bands indicate the injected carrier populations. As both electron and hole populations increase in the GaAs layer, radiative recombination becomes stronger.

A flat quasi-Fermi level indicates no driving force for the corresponding carrier current; a spatial gradient indicates transport. Large quasi-Fermi level changes close to an interface may also reveal an unwanted extraction or injection barrier. The band diagram therefore helps explain the shape of the electrical J–V curve.

5. Where are the infrared photons generated?

The purpose of the GaAs active region is to bring injected electrons and holes together so that they recombine radiatively. In a simple direct-gap bulk model, the net radiative recombination rate is described by

\[ R_{\mathrm{rad}}=B\left(np-n_i^2\right), \]

where B is the radiative recombination coefficient and n and p are the electron and hole densities. The quantity R_rad has units of m⁻³ s⁻¹. The active GaAs layer should contain a substantial fraction of the radiative recombination, with the AlGaAs barriers helping to confine the injected carriers.

In the Simulation snapshots viewer, add a curve for the radiative recombination rate and another for the electron or hole density using the File to plot selectors. Plot the rates at a low forward voltage and at a higher forward voltage. The available recombination-file names depend on the enabled electrical model and output settings; select the radiative rather than the total recombination quantity.

Not every recombination event creates a photon. Defects and traps can provide nonradiative recombination pathways that dissipate energy as heat. Raising the current can increase photon generation while also increasing nonradiative loss. The balance between these processes controls the LED's internal quantum efficiency.

6. Examining the infrared emission spectrum

In the main Device structure tab, click Emission parameters to inspect the material emission settings. The preset uses an emission model appropriate for the GaAs active layer, centred near the GaAs band edge. The emitted spectrum is distinct from a light-source spectrum: it is calculated for the electrically driven device rather than supplied as incident illumination.

After the forward-bias run, open the emission or optical output files from the Output tab and inspect the calculated spectrum. The emitted light should lie in the near-infrared, close to 870 nm for room-temperature GaAs. Its exact shape depends on the emission model, carrier populations and any reabsorption or optical-outcoupling calculation included in the preset. A radiative-recombination-rate calculation alone does not automatically determine the emitted spectrum or extracted optical power.

GaAs has a high refractive index, so a large fraction of the internally generated light can remain trapped by total internal reflection. Some photons may also be reabsorbed in the GaAs layers or lost at contacts. The spectrum and intensity that escape from the device therefore need not match the photons generated internally.

7. Internal efficiency and light extraction

An LED is assessed by how effectively it converts injected charge into photons. The internal quantum efficiency (IQE) is the fraction of injected carriers that produce radiative recombination. If the device is in steady state and the recombination currents are accounted for consistently, a useful estimate is

\[ \eta_{\mathrm{int}}\simeq\frac{q\int_{\mathrm{device}}R_{\mathrm{rad}}(y)\,dy}{|J|}. \]

Here J is the injected current density in A m⁻². The estimate assumes radiative events create approximately one photon per recombination event and requires consistent treatment of leakage and contact currents. The external quantum efficiency of an LED measures photons escaping the device per injected electron and is generally lower:

\[ \eta_{\mathrm{ext}}=\eta_{\mathrm{int}}\,\eta_{\mathrm{out}}, \]

where \(\eta_{\mathrm{out}}\) is the photon extraction efficiency. To estimate extracted optical power or angular emission, use the available optical outcoupling tools with appropriate wavelength-dependent optical constants, device geometry and emission distribution. Do not equate the integral of radiative recombination with the optical power leaving the LED.

For monochromatic light near 870 nm, the internal optical power generated per unit area can also be estimated from the photon-generation rate:

\[ P_{\mathrm{generated}}\simeq\frac{hc}{\lambda} \int_{\mathrm{device}}R_{\mathrm{rad}}(y)\,dy \qquad [\mathrm{W\,m^{-2}}]. \]

This quantity is an internal photon-energy estimate. The external emitted power is smaller when photons are trapped, reabsorbed or absorbed at contacts. For a more detailed optical treatment, see the optical transfer-matrix documentation.

8. Further investigations

Once the reference LED is running, change one parameter at a time to understand which parts of the device govern light generation and extraction.

9. Summary and what you have learned

This tutorial has demonstrated how OghmaNano models an 870 nm GaAs/AlGaAs infrared LED. We opened a working semiconductor structure, inspected the materials and doping, and followed carrier injection through the J–V curve and energy-band diagrams. The wider-bandgap AlGaAs layers help confine electrons and holes in the GaAs active region, where radiative recombination generates near-infrared photons.

We also distinguished between internal photon generation and light extracted from the LED. Both the electrical recombination physics and the optical layer stack matter when predicting practical device performance. You can now explore how forward bias, alloy composition, active-layer thickness and nonradiative losses affect the LED.

For further information, see the electrical material-model library, drift–diffusion theory and optical modelling documentation.