The combined illumination provided by two automobile headlights is twice that provided by a single headlight. However, it doesn’t work that way with lasers. Even if the lasers are aligned with closely spaced parallel beams — so that the beams essentially overlap in the distance, as automobile headlights do — interference effects among the coherent beams still can cause a drastic reduction in beam quality and also in the power delivered to a target.There are two, or maybe three, ways around this problem. The first, called spectral combining, is to offset the lasers’ wavelengths from each other slightly, so the beams cannot interfere. And while it would work to place the lasers close together so their beams will essentially overlap in the distance, there’s a better way to do it. Because each of the lasers has a different wavelength, their beams will be diffracted from a grating at different angles. If they are incident on the grating from appropriate angles, they can all be diffracted into a single beam. Other dichroic elements, such as a prism or a set of dichroic mirrors, can serve the same function as the grating. But a drawback of spectral combining is the increased spectral width of the combined beam.If the beams are of the same wavelength, combining them becomes more complicated. One approach, as noted above, is to align them in a closely spaced parallel array so the beams overlap in the distance. If the lasers are locked together in phase, then they interfere constructively, and the combined beam behaves pretty much as if it had originated from a single laser. (How much “pretty much” is depends on the beam quality and the spacing of the original lasers.) The drawback of this approach, called coherent combining, is that it can be very difficult to phase lock an array of lasers.Another approach to coherent combining, to be discussed below, does not need to pack the lasers together so they overlap in the distance. Instead, it utilizes a special optical element to combine the beams inside their respective resonators.The third approach is polarization combining, where two orthogonally polarized beams with identicalwavelengths are combined with a polarizing prism and cannot interfere because their electric fields are normal to each other. The “maybe” qualifier accompanies this technique because it cannot be applied to more than two lasers.Recently, Moti Fridman and his colleagues at the Weizmann Institute of Science in Rehovot, Israel, have used both spectral combining and coherent combining to add together the outputs of four fiber lasers with better than 80 percent overall combining efficiency. They speculate that the technique eventually could be effective in combining the beams of several hundred fiber lasers.In their proof-of-principle demonstration, the scientists carefully aligned four fiber lasers in a 2 × 2 matrix so that their longitudinal axes were parallel (Figure 1). They combined one pair of intracavity beams with the other pair using an interferometric combiner, resulting in a single pair of intracavity beams in the right-hand half of the multibeam resonator in Figure 1. These two beams are focused onto a grating that diffracts them onto the output coupler. Each laser self-selects the wavelength that will allow it to lase, depending on its angle of incidence on the grating. Each of the beams strikes the grating at a different angle, so two different wavelengths emerge from the output coupler. In other words, the two intracavity beams in the right half of the resonator are spectrally combined into a single output beam.Figure 1. Two pairs of intracavity beams are coherently combined with an interferometric combiner, and the resulting two beams are then spectrally combined with a grating. The insets show the four beams to the left of the interferometric combiner (bottom left), the two beams to the right of it (bottom right) and the output beam emerging from the resonator (top). Reprinted with permission of Optics Letters.The plane-parallel interference combiner simply translates one intracavity beam sideways so that it overlaps the other beam (Figure 2). At first glance, it would appear that half of each beam is lost at the 50-percent reflector — that is, the dotted arrows in Figure 2 represent half the power in each beam. Why this doesn’t happen is perhaps more easily understood philosophically than mathematically.Figure 2. The interferometric combiner translates one beam sideways so that it overlaps the other beam. The transmitting surfaces are antireflection-coated. In the perspective of Figure 1, this view is looking down from above. R= reflectivity.Philosophically, just as Mother Nature abhors a vacuum, she likewise abhors a population inversion. Both are distasteful departures from equilibrium. Stimulated emission is a good way to get rid of a population inversion, and the faster that happens, the better Mother Nature likes it. The more circulating power there is in a laser, the faster atoms are removed from the population inversion. Thus, Mother Nature will always seek to adjust all the parameters at her disposal to maximize a laser’s circulating power. Grasping this simpleconcept leads to an intuitive understanding of all sorts of resonator dynamics, from mode-locking to intracavity nonlinear optics to relaxation oscillations. It also explains why each laser self-selects its optimal wavelength, as discussed above.So, in Figure 2, each laser adjusts its phase to minimize its intracavity loss and thereby maximize its circulating power. Destructive interference occurs in the beams represented by the dotted arrows, and virtually all the power disappears from those beams. Figure 3. The spectrum of the beam emerging from the output coupler (c) is the straightforward addition of the two intracavity beams (a) and (b). Reprinted with permission of Optics Letters.The combined beam emerging from the output coupler in Figure 1 contained about 131 mW, or 82 percent of the total power from the four 40-mW lasers. The beam quality of the combined beam, M2 = 1.15, was essentially identical to the beam quality of each laser oscillating by itself. When the scientists looked at the spectrum of the combined beam, they saw two distinct peaks corresponding to the two spectrally combined intracavity beams. With approximately 1.3 nm between the peaks, and given the fiber lasers’ 70-nm spectral width, it theoretically should be possible to spectrally combine as many as 50 fiber lasers. And because the scientists have previously demonstrated that at least five lasers can be combined with the interferometric combiner, it seems conceivable that eventually several hundred fiber lasers could be combined with this technique.Optics Letters, April 1, 2008, pp. 648-650.