Cart 0

The history of semiconductor lasers

This is a continuation from the previous tutorial - guided-wave photodetectors.


The advent of the laser dates back to 1958, the year in which the seminal paper of Schawlow and Townes appeared. It was followed by the successful operation of a solid-state ruby laser in May 1960 and of an He-Ne gas laser in December 1960.

The feasibility of stimulated emission in semiconductor lasers was considered during this period, and in 1962 several groups reported the lasing action in semiconductors.

The device consisted of a forward-biased GaAs p-n junction. Electron-hole recombination in the depletion region of the p-n junction provided the optical gain, and the polished facets perpendicular to the junction plane provided the optical feedback - the two necessary ingredients for any laser.

Soon p-n junctions of other direct-band-gap semiconductor materials such as InAs, InP, GaAsP, GaInAs, and InPAs were used to obtain semiconductor lasers at different wavelengths.

Practical utility of these earlier devices was, however, limited since a large value of the threshold current density (\(J_\text{th}\gtrapprox50\text{ kA/cm}^2\)) inhibited their continuous operation at room temperature.

As early as 1963 it was suggested that semiconductor lasers might be improved if a layer of one semiconductor laser material were sandwiched between two cladding layers of another semiconductor that has a relatively wider band gap. Such a device consisting of two dissimilar semiconductors is commonly referred to as a heterostructure laser, in contrast to the single-semiconductor devices, which are labeled as homostructure lasers.

Both of these structures are shown schematically in Figure 1-1, which also indicates their typical physical dimensions.


Figure 1-1. Schematic illustration of (a) homostructure and (b) double-heterostructure semiconductor lasers with their typical physical dimensions. The dotted area represents the depletion region in the vicinity of the homojunction. The hatched area shows the thin (~0.2 μm) active layer of a semiconductor material whose band gap is slightly lower than that of the surrounding cladding layers.


Heterostructure lasers are further classified as single-heterostructure or double-heterostructure devices depending on whether the active region, where lasing occurs, is surrounded on one or both sides by a cladding layer of higher band gap.

The use of a heterostructure, however, requires a careful matching of the lattice constants of the two semiconductors. It was only in 1969 that the successful room-temperature operation of a heterostructure laser was demonstrated using the liquid-phase epitaxial technique for the growth of GaAs and AlxGa1-xAs layers.

However, these lasers operated in the pulsed mode. Further work led in 1970 to heterostructure lasers operating continuously at room temperature.

Notation such as (Ga,Al)As or AlGaAs/GaAs is often used to emphasize the heterostructure nature of these GaAs lasers. However, since homostructure lasers are no longer used, we shall simplify the notation in our tutorials, whenever no confusion is like to arise, by denoting a heterostructure laser only by the composition of its active layer.

Already in 1969 double-heterostructure GaAs lasers with a room-temperature value of \(J_\text{th}\approx5\text{ kA/cm}^2\) were reported. This value was reduce to about \(1.6\text{ kA/cm}^2\) in 1970, and by 1975 AlGaAs layers with \(J_\text{th}\approx0.5\text{ kA/cm}^2\) were demonstrated using thin (~0.1μm thick) active layers.

This was an improvement by more than two orders of magnitude over the simple homostructure lasers first made in 1962. It converted the semiconductor laser from a laboratory curiosity to a practical, compact, coherent light source useful for numerous applications.

The physical reason for the reduction in the threshold current density with the use of a heterostructure device is twofold. The cladding layers surrounding the active layer hav e higher band gap and at the same tie a lower refractive index compared with those of the active layer. (see Figure 1-2).

The band-gap difference helps to confine electrons and holes to the active layer, where they recombine to produce the optical gain.

At the same time the refractive-index difference confine the optical mode close to the active layer, which acts as a dielectric waveguide. The optical-mode confinement significantly reduces the internal loss that would otherwise occur in the absence of index guiding due to the spreading of the optical mode in the lossy regions.


Figure 1-2. Schematic illustration of the simultaneous confinement of the charge carriers and the optical mode to the active region occurring in a double-heterostructure semiconductor laser. The active laser has a lower band gap and a higher refractive index than those of the cladding layers.


A double-heterostructure semiconductor laser such as shown in Figure 1-1 is sometimes called a broad-area laser since it does not incorporate any mechanism for the lateral (parallel to the junction plane) confinement of the injected current or the optical mode. [refer to the lateral structures of semiconductor lasers tutorial]

As early as 1967, stripe-geometry homostructure lasers were proposed to limit the lateral spread of the injected carriers inside the active layer. In these lasers the current is injected over a narrow (~10 μm) central region using a stripe contact. The stripe geometry was adopted for heterostructure lasers in 1971. [refer to the lateral structures of semiconductor lasers tutorial]

Such lasers are also referred to as gain-guided since it is the lateral variation of the optical gain that confines the optical mode to the stripe vicinity. By contrast, heterostructure lasers where the optical mode confinement occurs mainly through lateral variations of the refractive index are termed index-guided. [refer to the lateral structures of semiconductor lasers tutorial

GaAs lasers are of continued interest. Using the growth techniques of vapor-phase epitaxy and molecular-beam epitaxy, multiquantum-well laser structures have been developed.

In these devices, the active region is not a single GaAs layer but rather consists of several ultra thin (~0.01 μm) layers composed alternatively of GaAs and AlGaAs materials.

Considerable effort has been directed toward developing high-power GaAs lasers and phased-array semiconductor lasers.

In one approach, multiple stripes are used to generate distinct regions of optical gain in the junction plane and the near field consists of several spots. However, since the stripes are not widely separated, the optical field in the gain region of each emitter overlaps with that of the neighboring emitter. Such a coupling leads to a phase-locked array of emitters providing well-collimated high power output.

Whereas the output power from a conventional GaAs laser is usually below 50 mW, more than 100 W of power has been obtained from monolithic laser arrays.

So far we have followed the development of GaAs lasers operating usually in the wavelength range of 0.8-0.9 μm. Long-wavelength semiconductor lasers int eh range of 1.1-1.6 μm are of considerable interest for optical fiber communications.

Although several material systems were considered, the combination of InGaAsP-InP turned out to be the most suitable in view of its nearly perfect lattice match.

The active layer is composed of the In1-xGaxAsyP1-y quaternary alloy. By varying the mole fractions \(x\) and \(y\), almost any wavelength in the 1.1-1.6 μm range can be selected.

The cladding layers in this heterostructure laser (see Figure 1-1) consist of either InP or InGaAsP itself with different mole fractions \(x\) and \(y\).

Room-temperature operation of a 1.1-μm InGaAsP laser in the pulsed mode was reported in 1975. The adoption of stripe geometry led to continuous operation of such lasers in 1976.

In 1977 the wavelength was extended to 1.3 μm. Since low-loss dispersion-free fibers at 1.3 μm were already available, considerable attention was focused on developing a practical InGaAsP laser at this wavelength.

Motivated by the realization of an ultra-low-loss (~0.2 dB/km) fiber at the 1.55-μm wavelength, several groups in 1979 reported on InGaAsP lasers operating in the vicinity of 1.55 μm.

After that, the development effort for InGaAsP lasers operating in the wavelength range of 1.3-1.6 μm proceeded at an enormous pace. The primary motivation was due to their application in optical fiber communications, and by 1984 the use of InGaAsP lasers in long-haul optical communication systems had reached the commercial stage.



Semiconductor Materials

In the above discussion we have followed the development of heterostructure lasers based on two semiconductor materials, AlGaAs and InGaAsP.

However, in view of their potential application in such diverse fields as optical fiber communication, optical data recording, high-speed printing, and molecular spectroscopy, the list of semiconductor materials that have exhibited lasing action has continued to grow.

Figure 1-3 shows the range of emission wavelengths for various semiconductor lasers. Taken together, these materials cover the optical spectrum from near ultraviolet to far infrared.

Recently CdZnSe semiconductor lasers operating near 0.46 μm have been developed.


Figure 1-3. Wavelength range of semiconductor lasers covered by different material systems. Semiconductor lasers emitting at \(\lambda\gt3\) μm usually require low-temperature operation.


The most important criterion in selecting the semiconductor material for a specific heterostructure laser is related to the quality of the heterojunction interface between the two semiconductors of different band gaps.

To reduce the formation of lattice defects, the lattice constants of the two materials should typically match to better than 0.1%.

Figure 1-4 shows the inter-relationship between the band gap \(E_\text{g}\) and the lattice constant \(a\) for several ternary and quaternary compounds.

Solid dots represent the binary compounds and solid lines correspond to the ternary compounds.

The clear region bounded by the polygon (whose edges represent ternary compounds) denote the possible values of \(E_\text{g}\) and \(a\) for the quaternary solid solution of In1-xGaxAs1-yPy obtained by varying the mole fractions \(x\) and \(y\).

The dotted line shows the range of band-gap values that can be achieved by varying the compositions \(x\) and \(y\) to obtain a quaternary material that is lattice-matched to the binary InP. 

Figure 1-4. Band gap and lattice constant for In1-xGaxAs1-yPy (clear region) and (AlxGa1-x)yIn1-yP (shaded region) obtained by varying compositions \(x\) and \(y\). Dashed lines separate indirect-band-gap regions (shown hatched). Dotted lines show the wavelength range (top scale) for a semiconductor laser whose quaternary active layer is lattice-matched to the binary compound.


Figure 1-5 shows the constant band-gap contours (solid curves) in the \(x-y\) compositional plane. The dashed lines correspond to the fixed values of the lattice constant.

For given values of \(E_\text{g}\) and \(a\), the intersection of solid and dashed curves provides the compositional values \(x\) and \(y\) used to obtain the active-layer quaternary material.

Since the photon energy \(E=h\nu\) is approximately equal o the band-gap energy, the lasing wavelength \(\lambda\) is obtained using \(E_\text{g}=hc/\lambda\), where \(h\) is the Planck constant and \(c\) is the speed of light in vacuum.

If \(E_\text{g}\) is expressed in electron volts, the lasing wavelength \(\lambda\) in micrometers is given by


Figure 1-5. Contours of constant band gap (solid lines) and constant lattice spacing (dashed lines) in the \(x-y\) compositional plane for In1-xGaxAsyP1-y. The composition values \(x\) and \(y\) can be chosen to obtain a particular band gap (or laser wavelength) for a given lattice constant. Band gap is indirect in the shaded area. 


For In1-xGaxAsyP1-y lasers a wavelength range of 1.1-1.65 μm can be covered by choosing \(x\) and \(y\) according to Figure 1-5 such that the active layer is lattice-matched to InP (\(a=0.587\text{ nm}\)).

Semiconductor lasers emitting at 1.3-μm and 1.55-μm wavelengths are of particular interest because of their application in optical fiber communications.

From Figure 1-5 the active-layer composition at 1.3 μm (\(E_\text{g}\approx0.95\text{ eV}\)) corresponds to \(x=0.28\) and \(y=0.6\).

The cladding-layer composition can also be chosen from Figure 1-5; the only requirement is that its band gap should be somewhat larger than that of the active layer.

Diagrams similar to Figures 1-4 and 1-5 have been constructed for other heterostructure materials in order to provide guidance for the lattice-matched growth of the active and cladding layers.

Recent advances in the epitaxial growth techniques permit a lattice mismatch of up to a few percent without degrading the interface quality significantly. Such semiconductor lasers are referred to as strained-layer lasers and have attracted considerable attention because of their superior performance.

To cover the longer wavelength region (\(\lambda\gt1.6\) μm), other material systems shown in Figure 1-3 have been successfully exploited.

Lattice-matched active layers of the quaternary InxGa1-xAsySb1-y material have been grown on GaSb substrate and can cover the wavelength range of 1.7-4.4 μm.

Another important class of materials is lead salts. Lead salts have been used to make semiconductor lasers emitting in the far-infrared region ranging from 3-34 μm, although low-temperature operation is required.

Two material systems of particular interest are PbTe-Pb1-xSnxTe and PbS-PbS1-xSex. The longest wavelength of about 100 μm is obtained using the semiconductor material Bi1-xSbx.

By contrast, short-wavelength semiconductor lasers require the use of so-called II-VI semiconductors such as ZnS or ZnSe, often doped with Cd.



The next tutorial covers the topic of operating principles of semiconductor lasers

Share this post



Sold Out