Vertical-cavity surface-emitting lasers (VCSELs) producing several hundred watts are an attractive alternative as a high-power semiconductor laser source, not least because they can be easily processed in 2D arrays to scale up the output power.
Fabrication methods
Semiconductor lasers consist of layers of semiconductor material grown on top of each other on a substrate. This is known as the epitaxial structure.
For VCSELs and edge-emitters, this growth is typically done in molecular-beam-epitaxy or metal-organic-chemical-vapour-deposition growth reactors. The grown wafer is then processed accordingly to produce individual devices. Figure 1 summarizes the differences between VCSEL and edge-emitter structures.
In a VCSEL, the active layer is sandwiched between two distributed Bragg reflectors made up of several quarter-wavelength-thick layers of semiconductors of alternating high and low refractive index. The reflectivity of these mirrors is typically between 99.5 and 99.9%. As a result, the light oscillates perpendicular to the layers and escapes through the top (or bottom) of the device.
Current and/or optical confinement is typically achieved through selective-oxidation of an aluminium-rich layer, or ion-implantation (or even both for certain applications). The VCSELs can be designed for top emission at the epitaxial/air interface or bottom emission through the transparent substrate, in cases where "junction-down" soldering is required for more efficient heat-sinking for example.
Since VCSELs are grown, processed and tested while still in the wafer form, there is significant economy of scale resulting from the ability to conduct parallel device processing. Here, equipment utilization and yields are maximized and set-up times and labour content are minimized. In fact, VCSEL processing is identical to the well-established, low-cost silicon integrated-circuit planar processing.
In the case of a VCSEL, the mirrors and active region are stacked sequentially during epitaxial growth. The VCSEL wafer then goes through etching and metallization steps to form electrical contacts. At this point, the wafer goes to test and individual laser devices are characterized on a pass–fail basis. Finally, the wafer is diced and the good lasers (a yield of greater than 95% is typical) move on to higher-level assembly.
The wafer can be diced into single devices or arrays of single devices effectively connected in parallel. The arrays can be linear (1D), rectangular or square (2D). Furthermore, since the position of the individual elements in a VCSEL array is defined by photolithography, arbitrary design layouts of the elements with a placement accuracy on the micron level are possible. Depending on the application, 2D VCSEL arrays can contain anything from a few hundred to several thousands of single devices. Figure 2 shows a cross-section schematic of a processed VCSEL array.
The unique architecture of VCSELs also leads to other distinct advantages over edge-emitters. For example, VCSELs are not subject to catastrophic optical damage (COD) and therefore have a much higher reliability than edge-emitters. VCSELs can also operate at very high temperatures.
High-power arrays
Designing efficient single devices is the key to producing high-power 2D VCSEL arrays (Proc. SPIE 6908 690808 [2008]). At Princeton Optronics, we have developed high-efficiency single devices (10–20 µm selectively oxidized aperture), emitting around 976 nm, and with power conversion efficiencies (PCE) greater than 51%. When a large number of these devices are combined in a 2D array, the high PCE can be preserved and the output power is increased. Figure 3 shows a fully packaged 976 nm 2D VCSEL array on a micro-channel cooler.
This array has an emission area of approximately 0.22 cm2, a peak power conversion efficiency of greater than 44% and was operated under a constant heat-sink temperature of 15 °C. A record 231 W continuous-wave (CW) output power was reached at a drive current of 320 A, limited by thermal roll-over. This corresponds to a power density of 1 kW/cm2, which is similar to that achieved by high-power edge-emitter stacks. We have also fabricated lower-power arrays (90 W maximum) with PCE of 51%, identical to that of our single devices.
We have also examined the spectral and beam properties at a CW output power of 100 W for the array shown in figure 3. Results are shown in figure 4. The spectral full-width half-maximum was only 0.8 nm, which is about five times less than the spectral width of edge-emitter bars or stacks (typically in the 3–5 nm range).
We measured the wavelength shift as a function of the heat-sink temperature to be 0.065 nm/K, which is identical to the value for our single devices. This value is five times less than that of edge-emitters (typically 0.33 nm/K). It is clear that our high-power VCSEL arrays benefit from the same intrinsic narrow spectrum and stable emission wavelength as our single devices. These properties are useful for many pumping applications where the medium has a narrow absorption band.
The far-field beam profile of the array is circular, with a quasi-top-hat profile, and a 1/e2 full-width divergence angle of 17°. Since such beam characteristics can be achieved without any optics, VCSEL arrays present a cost-effective solution for end-pumping applications.
Smaller arrays with an area of 0.028 cm2 were soldered onto submounts and tested in quasi-CW (QCW) mode. The chip-on-submounts were tested on a thermo-electrically cooled stage maintained at 20 °C. Pulse width and duty factor were 100 µm and 0.3%, respectively. The maximum power reached was 100 W, which was limited only by the QCW current driver (125 A). This corresponds to a power density of 3.5 kW/cm2, which again is similar to what is achieved with QCW edge-emitter stacks. Higher power levels can be achieved by connecting several chips in series.
High-energy arrays
Beyond CW and QCW operation, 2D VCSEL arrays can also be pulsed in the 10–100 ns regime for a variety of applications. For example, there is increased interest from the military for a compact rangefinder to equip soldiers with a lightweight tool for high-precision targeting engagements.
One of the main components of this micro laser rangefinder (LRF) is a passive saturable absorber Cr4+:YAG Q-switch (Applied Optics 39 2428). A pulse-coded signal is desired for rangefinder applications, which means that the Q-switch must be actively turned on and off to generate a series of non-evenly spaced pulses.
To overcome this issue, Princeton Optronics has developed a high-energy 5 × 5 mm VCSEL array capable of producing 100 ns pulses that pump the chromium crystal. Our arrays have an output energy of 220 µJ at a 2 kA current, corresponding to a record peak power of 2.2 kW, as shown in figure 5(a). The fully assembled active Q-switch is shown in figure 5(b).
One of the main advantages of VCSELs is that, since they are not subject to COD, they do not fail when operated at many times their roll-over CW current at such high output energies. Also, their unique geometry and properties allow for direct assembly against the crystal facets, without the need for coupling optics or isolators.
In the active Q-switch module, each of the four VCSEL array chips is pressed directly against one of the crystal facets (which is anti-reflection coated at 980 nm). In fact, the crystal is held in place by the pressure of all four VCSEL chips against its facets. The four VCSEL arrays provide a combined energy of close to 1 mJ in the crystal and in turn provide a means to actively control the Q-switch by turning it on (pumped) and off (no pumping).
Conclusions
VCSEL arrays have many advantages such as low manufacturing costs, reliability, high-temperature operation, intrinsic spectral stability and good beam quality. For end-pumping applications, these 2D arrays can potentially be assembled directly against the facets of the doped medium, without the need for optics or isolators.
VCSEL arrays can also emit short pulses (<200 ns) at many times their roll-over CW current without any COD-induced failure, making them a reliable laser source for high-energy applications such as a designator, beacon and active Q-switch.
Because of their significant and unique advantages in terms of costs, reliability and performance, VCSELs could become the technology of choice for compact and efficient high-power or high-energy semiconductor laser sources in many applications.
• This article originally appeared in the April 2008 issue of Optics & Laser Europe magazine.
