Laser Arrays

This is a continuation from the previous tutorial - strongly index-guided lasers.

InGaAsP semiconductor laser structures described in the previous tutorials have been developed for low-power (5-10 mW/facet) applications, such as providing a source in a lightwave communication system. With a proper design, these lasers can generally emit continuously up to powers in the range of 30-60 mW/facet near room temperature.

This limitation in the output power mainly arises from the leakage current, which increases with an increase in the applied current.

Semiconductor laser diodes for high-power operation have been extensively studied using the AlGaAs material system.

High-power lasers developed so far can be classified as one of two types, the single-emitter laser and the laser array.

A laser array has closely spaced multiple active regions, each of which emits light. The radiations from neighboring emitters in a laser array overlap, establishing a definite phase relation between them. As a result the array emits in a narrow, coherent beam.

Single-emitter laser structures are similar to those described in the previous tutorials. The leakage current limits the highest operating power from these devices. Nonetheless, 140 mW of CW power has been obtained from a 1.3-μm DCPBH InGaAsP laser with design optimization and facet coatings. The leakage current can be reduced in a laser structure with a p-InP substrate. The use of unstable-resonator geometry has also provided output powers of ~ 400 mW in a single lateral mode.

Laser-array structures using the AlGaAs material systems have been extensively studied. More recently, InGaAsP laser arrays have attracted attention.

Laser-array structures can be classified into two groups: gain-guided arrays and weakly index-guided arrays. The difference between the two types is that the mode profile along the junction plane is determined by gain guiding for the former and by weak index guiding for the latter.

Figure 5-36 shows three laser-array structures.

The gain-guided array is a dielectric-stripe device. The openings in the dielectric (\(\text{SiO}_2\)) and their center-to-center spacings are chosen so that there is adequate overlap between oscillating modes of neighboring emitters.

The weakly index-guided laser array can employ a ridge waveguide or a rib waveguide structure. The dimensions of and spacing between emitters are again chosen in such a way as to provide adequate overlap of the oscillating modes. This overlap establishes the coupling necessary for phase-locked operation. 

Figure 5-37 shows the \(\text{L}-\text{I}\) curves for a 1.3-μm gain-guided InGaAsP laser array with ten emitters. InGaAsP laser arrays emitting at 1.3 μm have been operated to pulsed output powers of 800 mW/facet.

An often-encountered problem in the performance of a laser array is that the far field along the junction plane consists of two lobes, a few degrees apart, whereas a single-lobe far field is desirable for practical applications.

The nature of the far field depends critically on the phase relationships between neighboring emitters, and single-lobe and twin-lobes far fields are generally related to in-phase and out-of-phase operations, respectively.

The relative phase between adjacent emitters, however, is generally determined by the lateral variation of the optical gain and the refractive index. With proper design, single-lobe far-field operation has been observed in some semiconductor laser arrays.

Figure 5-38 show such far fields obtained under the in-phase operation of a 1.3-μm InGaAsP gain-guide laser array. The subsidiary maxima in the far field are due to the presence of other transverse modes. Such higher-order modes have slightly different wavelengths and can be identified in the optical spectrum.

Two schemes have been proposed that lead to single-lobe far fields. These are (i) a chirped laser array, i.e., a laser array with variable emitter dimensions or a variable spacing between emitters, and (ii) a laser structure with higher optical gain in between the emitters.

The normal modes of a phase-locked laser array have been studied using coupled-mode analysis. These normal modes, generally referred to as the supermodel or array modes, oscillate at different wavelengths with different near-field and far-field patterns. Coupled-mode analysis is very useful for a qualitative analysis of device behavior.

In a more recent approach the modes of the composite-array cavity are determined through a self-consistent solution of the paraxial wave equation that incorporates lateral variations of both the optical gain and the refractive index.

Even though the concept of coupling between emitters is not invoked, the near filed has a multipeak profile with each peak confined to a single-emitter region. This approach is useful for understanding the role of the lateral guiding mechanism and its influence on in-phase and out-of-phase coupling between emitters.

An important aspect of laser-array research is the determination of suitable structures capable of producing high output power in a diffraction-limited beam.

Schemes that have been demonstrated for this purpose include a diffraction-type spatial filter, Talbot cavities, antiguided arrays and Y-junction arrays.

Of all these schemes, antiguided laser arrays have produced very high power levels in a nearly diffraction-limited beam. Linear arrays of antiguided lasers have strong interelement cooling due to the presence of leaky waves.

The structure of an antiguided laser array is shown in Figure 5-39.

The fabrication process consists of two MOVPE growth steps. The first step involves the growth of the active region and the n- and p-type cladding layers. The second step involves the growth of the p-type \(\text{Al}_{0.18}\text{Ga}_{0.82}\text{As}\) layer and the layers that follow as shown in Figure 5-39.

The lateral antiguiding property results from placing the \(\text{Al}_{0.18}\text{Ga}_{0.82}\text{As}\) guide layer in close proximity (\(\le0.2\) μm) to the active layer. 

The lateral far-field pattern of a 20-element array is shown in Figure 5-40. Close to threshold the output beam is diffraction-limited and at 3 times threshold the beam width is \(1.5^\circ\) (\(1.8\times\) diffraction limit) with 330 mW output power.

Surface-Normal Emitting Lasers

A conventional semiconductor laser has cleaved facets that form the optical cavity [see Figure 5-1 in the broad-area lasers tutorial]. The mirror facets are perpendicular to the surface of the wafer and light is emitted parallel to the surface of the wafer.

Many applications require semiconductor lasers emitting light normal to the wafer surface. Two categories of laser exist where light is emitted normal to the surface of the wafers.

In the first category are lasers in which the optical cavity is normal to the surface of the wafer. These lasers are also known as vertical-cavity surface-emitting lasers.

The second category consists of lasers where the active region has the conventional waveguide form, but the light is deflected normal to the surface using a mirror or a diffraction grating fabricated adjacent to the laser waveguide. Such lasers are described in this tutorial.

Liau and Walpole have reported a buried-heterostructure laser whose light output is deflected perpendicular to the wafer surface, so that the laser, in effect, is a surface-emitting laser.

The laser structure is shown in Figure 5-41. The mirrors of the laser are formed by the technique of vapor-phase transport. The beam is deflected normal to the surface using a monolithically integrated \(45^\circ\) parabolic mirror. Such lasers have room-temperature threshold currents in the range of 12-18 mA and exhibit performance characteristics comparable to those of conventional cleaved-facet lasers.

A grating can be used to deflect the laser output perpendicular to the wafer surface, such lasers are called grating-coupled surface-emitting lasers. The schematic of the laser is shown in Figure 5-42.

A diffraction grating is etched on the p-type cladding layer, outside the waveguide region, to couple the light out vertically. Bragg gratings also serve as end mirrors for the laser.

Threshold current densities as low as \(400\text{ A/cm}^2\) have been demonstrated using these devices. Diffraction gratings operate in the second-order diffraction mode. Lasers of this type have been fabricated using both the GaAs-AlGaAs (\(\lambda\sim0.85\) μm) and the InGaAs-GaAs (\(\lambda\sim0.9-1.05\) μm) material systems.

The grating-coupled surface-emitting structure has been used to fabricate high-power laser arrays. A schematic of a two-dimensional 10 x 10 array is shown in Figure 5-43(a). The active region in this device is continuous throughout the wafer.

In Figure 5-43(b) the arrays are coupled by the evanescent light, and in Figure 5-43(c) a Y-branch waveguide is used for coupling.

Two-dimensional arrays of this type are useful for applications requiring high power.

The next tutorial discusses about the rate equations for semiconductor lasers.