Invoking a concept that dates from the early days of Nd:YAG lasers, when engineers aligned multiple laser rods in the same resonator, a collaboration between researchers in the US and Germany has demonstrated a technique of coherently combining the outputs of two vertical external cavity surface-emitting laser (VECSEL) chips to generate almost twice as much power as either chip alone can produce.Optically pumped VECSELs produce high output powers in low-divergence, round output beams, but their power is ultimately limited by overheating. Overheating can be alleviated by expanding the pump-beam size to deposit the same heat over a broader area of the chip, but this maneuver works only up to a point. Eventually, amplified spontaneous emission and other effects, which increase as the size of the pump beam increases, limit the power output.When an application requires more power than can be obtained from a single VECSEL, one might consider combining the outputs from two or more into a single beam. But because VECSELs are lasers — which produce coherent light — this must be done carefully lest interference effects play havoc with the power in the combined beam. Polarization beam combining, where two beams of orthogonal polarizations are combined, is one possible approach because the orthogonally polarized beams cannot interfere with each other. However, this technique requires polarizing each laser, and it cannot accommodate more than two lasers.Another possibility is spectral beam combining, where the beams are offset spectrally from each other and cannot interfere. The drawback of this technique is that the large bandwidth of the combined lasers makes the beam incoherent.The most sophisticated approach is coherent beam combining, where the lasers are artificially phase-locked with each other, and the electromagnetic fields of their beams combine coherently. While this technique sounds good in theory, in practice it requires complex phase-locking loops, which can be extremely difficult to implement and maintain. Instead of using any of these methods, the team — which included scientists from the University of Arizona and Areté Associates, both in Tucson, from Wright-Patterson Air Force Base in Ohio and from Philipps Universität in Marburg, Germany — put two VECSEL chips together in the same external resonator (Figure 1). The chips served as the resonator’s folding mirrors, reflecting more than 99.9 percent of the incident radiation from the 25 pairs of AlGaAs/AlAs that formed a distributed Bragg reflector behind the quantum wells. The InGaAs compressively strained quantum wells (14 of them on one chip and 10 on the other) were spaced so that each well aligned with an antinode of the resonator’s standing wave. Each chip had a quarter-wave antireflection coating on its surface, for a reflectivity of less than 1 percent at the 970-nm laser wavelength and less than 3 percent at the 808-nm pump wavelength.Figure 1. The laser with two intracavity gain chips produced nearly twice as much power as could be obtained from either chip alone. Images reprinted with permission of Optics Letters.Both chips were optically pumped at 808 nm from diode lasers, but because the researchers were limited to equipment on hand, pumping of the two VECSEL chips was not symmetric at higher powers. The researchers applied the same power density to both chips until they reached the maximum power (~20 W) for one of the pump lasers. But because the other pump laser could go to higher power, they continued increasing its power up to 40 W.For comparison purposes, they first characterized each VECSEL chip in its own linear resonator. They then aligned the two chips in a common resonator, as indicated in Figure 1. The comparison showed that the two-chip resonator produced almost twice as much power as either individual chip did in its own resonator (Figure 2). The two-chip resonator showed a definite roll-off at the upper end, an indication of overheating of one (or both) of the chips. The researchers believe that only one chip — the one pumped at 40 W — was overheated and that the roll-off could be avoided by pumping both chips at the same power.Figure 2. The roll-off at the upper end of the two-chip trace could probably be avoided by pumping both chips at the same power.The beam quality of the two-chip resonator was excellent. The measured M2 value was a perfect 1.0 at low pump powers and increased to about 2.2 at the highest powers. At the highest powers, the beam profile started to show the two peaks of the TEM01 mode. Again, the researchers suspect that beam quality also would be improved by pumping both chips at the same power, reducing the thermal load of the overheated chip.Optics Letters, Dec. 15, 2006, pp. 3612-3614.