Volume Bragg Grating Narrows Parametric Oscillator’s Output
Bandwidth is cut in half with almost no reduction in output energy.
Optical parametric oscillators (OPOs) provide an important mechanism to extend and tune the wavelength range of conventional lasers. Conceptually, each photon of the laser beam entering the OPO is split into two photons that share the energy of the original photon, so two beams of lower-energy photons (i.e., longer wavelength) emerge from the OPO. The parameters of the OPO can be adjusted to vary the way in which the incoming photon’s energy is split between the two new photons so that the wavelengths of the emerging beams can be tuned.
An OPO consists of a nonlinear crystal inside an optical resonator. In practice, singly resonant OPOs — those in which only one of the two new wavelengths oscillates — have proved effective and reliable devices for tuning nanosecond laser pulses to longer wavelengths. But, as is the case for any oscillator, the bandwidth of an OPO matches the bandwidth of either its gain medium or its resonator feedback, whichever is narrower. For an OPO, the gain mechanism is the phase matching in the nonlinear crystal, and the feedback mechanism is the reflection from the resonator mirrors, both of which are relatively broad. The resulting broad output spectrum is unsuitable for many applications.
Figure 1. The OPO resonator was folded, with a corner cube at each end and a flat folding mirror in the middle (top). Scattered green pump light illuminates the OPO components (bottom). (PPKTP = periodically poled KTiOPO4.) Images reprinted with permission from Optics Letters.
Recently, Björn Jacobsson and his colleagues at KTH-Royal Institute of Technology in Stockholm substituted a volume Bragg for one of the reflectors in a singly resonant OPO, thereby narrowing the resonator bandwidth and reducing the OPO’s bandwidth by a factor of two without significantly reducing the output energy. A volume Bragg grating is a block of transparent material whose refractive index is modulated with a series of closely spaced planes, usually created by exposure to interference fringes of ultraviolet light. As with any grating, it reflects only a narrow band of wavelengths that resonate with its periodicity and transmits all other wavelengths.
The scientists in Sweden designed a ring resonator for their OPO, with a corner-cube reflector at each end and a dichroic folding mirror in the middle (Figure 1). They fabricated one corner cube by polishing a right-angle prism on the end of the nonlinear crystal, a periodically poled KTiOPO4 (KTP) crystal. For the other corner cube, they assembled a volume Bragg grating on a mirror so that the reflected beam always returned parallel to the input beam, regardless of the assembly’s orientation.
Volume Bragg gratings have been used to reduce OPO bandwidths, but because they were not in the corner-cube arrangement shown in Figure 1, tuning the OPO wavelength resulted in an undesirable change in the direction of the output beam. The corner-cube arrangement has been demonstrated, but only with laser diodes and solid-state lasers. The scientists believe that theirs is the first OPO to embody the corner-cube technique.
They pumped the nonlinear crystal through the folding mirror, which was highly transmissive at the 532-nm pump wavelength and at the wavelength of one of the two new photons (the “idler”) but highly reflective at the wavelength of the other new photon (the “signal”). Thus, most of the idler, along with any remaining pump, was dumped out of the resonator after emerging from the KTP crystal and hitting the folding mirror.
Figure 2. By rotating the retroreflector in Figure 1, the scientists tuned the OPO output by 0.2 nm (100 GHz).
However, the signal was reflected off the folding mirror toward the corner-cube assembly. Regardless of the assembly’s angular orientation, the signal always was reflected on the same path back to the folding mirror. But the reflectivity peak of the grating depended on the incident angle of the light, so the corner-cube assembly not only reduced the OPO’s bandwidth but also allowed the output wavelength to be tuned (Figure 2).
The signal intensity diminished considerably at the ends of the tuning range shown in Figure 2 because the OPO’s gain bandwidth — the phase-matching bandwidth in the KTP crystal — was only several tenths of a nanometer.
To tune over a wider range, the scientists adjusted the phase-matching wavelength by changing the KTP temperature while simultaneously adjusting the orientation of the grating-corner-cube assembly (Figure 3).
Figure 3. By simultaneously adjusting the phase-match wavelength and the grating reflectivity, the scientists tuned the OPO output from 757 to 762 nm (2.6 THz). The broken line shows the output of a similar OPO using mirrors instead of the grating-corner-cube assembly. The 0.50-nm bandwidth in this case was twice the 0.25-nm bandwidth of the OPO with the Bragg grating.
Figure 3 also shows that the bandwidth of the OPO with the volume Bragg grating is half that of a similar OPO with mirrors instead of a grating. The scientists observed similar output energies from both OPOs, both producing ∼0.2 mJ of signal from 0.6 mJ of pump. But they pushed the grating OPO to higher pump energies, observing a little more than 0.4 mJ of signal when they boosted the pump to 1.3 mJ.
Optics Letters, Nov. 15, 2007, pp. 3278-3280.
MORE FROM PHOTONICS MEDIA