Active region

The active region is the part of a semiconductor light emitter in which electrons and holes meet and recombine: the gain layer of a laser diode or semiconductor optical amplifier, and the emitting layer of an LED. It is a narrower-bandgap material sandwiched between wider-gap barriers or cladding layers, so that the heterojunctions on either side trap the injected carriers and concentrate them to a carrier density of order 10¹⁸ cm⁻³. Its composition sets the emission wavelength, its thickness and number of layers set the current needed to reach gain, and its overlap with the optical mode sets how much of that gain the light sees.

Bulk, wells and dots

Three forms are in use. A bulk double-heterostructure active layer is typically of order 0.1–0.2 µm thick, thick enough that the carriers behave as in a three-dimensional crystal. A quantum well is 5–10 nm thick, thin enough that the motion perpendicular to the layer is quantized; most modern lasers use a multi-quantum-well stack of 3–10 wells separated by barriers, a few tens of nanometers of active material in total. Quantum dots confine carriers in all three directions; a quantum-dot laser stacks several layers of self-assembled dots. In all three, the active layer is too thin to guide light efficiently by itself, so it sits inside a separate-confinement heterostructure: a thicker waveguide core of intermediate index, a few hundred nanometers across, that holds the optical mode while the wells hold the carriers.

Confinement factor and threshold gain

Only the fraction $\Gamma$ of the mode that overlaps the active material is amplified, the confinement factor. A single 8 nm well in a separate-confinement structure captures roughly 0.01–0.03 of the mode; five wells together capture about 0.1, or 0.02 per well. The threshold condition is

$$\Gamma\, g_\text{th} = \alpha_i + \alpha_m .$$

For a 300 µm Fabry–Pérot cavity with uncoated facets ($R = 0.32$), the mirror loss is $\ln(1/R)/L = 38.0$ cm⁻¹. With an internal loss of 10 cm⁻¹ and $\Gamma = 0.1$, the wells must supply a material gain of

$$g_\text{th} = \frac{10 + 38.0}{0.1} \approx 480\ \text{cm}^{-1},$$

a value within reach of strained InGaAsP or InGaAlAs wells. Adding wells raises $\Gamma$ and lowers the gain each well must provide, but each added well also has to be filled to transparency; the threshold current density passes through a minimum at an intermediate number of wells, which for 1.55 µm edge emitters is usually 5–8.

Active volume and threshold current

The threshold current scales with the active volume $V = W L d$, stripe width times cavity length times total active thickness:

$$I_\text{th} = \frac{q\,V\,N_\text{th}}{\tau}.$$

A 2 µm wide, 300 µm long ridge with 40 nm of wells has $V = 24$ µm³ (2.4 × 10⁻¹¹ cm³). With $N_\text{th} = 2.5 \times 10^{18}$ cm⁻³ and a carrier lifetime of 2 ns, $I_\text{th} \approx 4.8$ mA, or 800 A/cm² over the stripe area. Measured thresholds of such ridges are often higher, around 10 mA, because current spreads laterally beyond the ridge and some carriers leak over the barriers. The same scaling explains why VCSELs, with active volumes below 1 µm³, run at sub-milliampere thresholds, and why broad-area laser diodes with stripes of 100 µm need hundreds of milliamperes.

Junction heating

Electrical power that does not leave as light heats the active region. A ridge laser driven at 60 mA and 1.3 V that emits 12 mW dissipates 66 mW; with a thermal resistance of 40 K/W from active region to heat sink, typical of a ridge mounted junction-side down,, the active region runs 2.6 K above the heat sink, enough to shift a DFB wavelength by roughly a quarter of a nanometer at a typical 0.1 nm/K. Heating raises the threshold through the characteristic temperature $T_0$, lowers the efficiency, and at high current produces the thermal rollover of the light-current curve. The article on measuring laser diode junction temperature covers how the actual temperature is found.

Pitfalls

Values of transparency and threshold density are quoted per well in some papers and per total active thickness in others, and the two differ by the number of wells. "Active region" is also used loosely for the whole separate-confinement core; the gain and $\Gamma$ refer to the wells alone. In an LED, light generated in the active region must still escape the high-index chip, so the internal and external efficiencies differ widely.

Common questions

How thick is the active region of a laser diode?

A single quantum well is 5–10 nm thick; a multi-well stack totals a few tens of nanometers; an older bulk active layer is of order 0.1–0.2 µm. The surrounding optical waveguide is thicker, a few hundred nanometers.

Why use quantum wells instead of a bulk active layer?

A thin well holds far fewer carriers, so it reaches transparency at a much lower current, and its step-like density of states gives more gain per carrier. Strain in the well can further reduce the threshold. The price is a small confinement factor, recovered by stacking several wells.

References: L. A. Coldren, S. W. Corzine and M. L. Mašanović, Diode Lasers and Photonic Integrated Circuits, 2nd ed. (Wiley, 2012); G. P. Agrawal and N. K. Dutta, Semiconductor Lasers, 2nd ed. (Van Nostrand Reinhold, 1993); E. F. Schubert, Light-Emitting Diodes, 2nd ed. (Cambridge University Press, 2006).