Optical Parametric Oscillator
Why OPOs?
In the past, several detection methods were used for trace gas measurements in biomedical applications, ranging from mass spectrometry, gas chromatography, chemiluminescence, electronic nose, to different laser-based spectroscopic methods. As many molecular gases have strong absorption in the mid-IR wavelength region, it is advantageous to use laser absorption spectroscopy; it is an excellent method for highly sensitive and selective detection of VOCs. For trace gas sensing, OPOs have to be developed in the mid-infrared wavelength region between 2.5 and 25 µm.
History
In 1965, five years after the demonstration of the first laser, Joseph Giordmaine and Bob Miller developed the Optical Parametric Oscillator (OPO) at Bell lab's. While lasers found almost directly widespread use in the field of gas sensing, this only happened for OPOs thirty years later. Its initial slow development was caused by the fact that OPOs were far from being a workhorse. Lack of proper non-linear crystals with a sufficiently high damage threshold, the unavailability of good pump lasers, and the need for skilled operators restricted their usefulness in applications. The advent of new crystals like Periodically Poled Lithium Niobate (PPLN) and a new generation of (fiber) pump lasers have changed this situation. Nowadays, OPOs are commercially available from many suppliers covering the visible, near- and mid-infrared and (recently) part of the ultraviolet wavelength region, using the pulsed and continuous wave (cw) regime.
Physical basics of OPOs
In an optical parametric process the pump photon (ωp) is split in two parts, forming two new photons with different energies (Fig. 2). The generated photon with the highest energy is termed signal (ωs), and the other photon is termed idler (ωi). The energy conservation has to be considered for the generated pair of photons, but is not sufficient to determine which frequencies will be amplified. In addition to that, destructive interference between the photons needs prevention. This gives rise to a second restriction, the phase matching condition. In bulk crystals birefringence is used to offset the dispersion and it satisfies the phase matching condition for specific wavelength combinations. Altering the propagation direction in such crystals (angle tuning) changes the selected frequency combination. Changing the crystal temperature may also change the refractive indices nidler, nsignal and npump resulting in temperature tuning.
Figure 2: Singly resonant Optical Parametric Oscillator (SRO) configuration generating mid-infrared light. For this singly resonance case, all mirrors are highly reflective for the signal wavelength, while transparent for the pump and idler.
Instead of trying to match the phase velocity of the pump laser and harmonics beams throughout the entire bulk crystal, an alternative way is Quasi phase matching (QPM), which was devised independently by Armstrong and Franken in 1962 and 1963, respectively. In materials with a second order susceptibility c(2), the pump wave (E) generates a nonlinear polarization wave (P) which oscillates with the different frequency. These waves E and P, while travelling through the crystal, will get a phase difference of p after a certain distance which is called coherence length (Lc). The invention of QPM reduces the effect of the destructive interference over long path length in bulk crystals, by resetting the phase difference between P wave and E wave after an odd number of coherence lengths (Fig. 3). This can be achieved by periodically inverting the generated polarization wave, which is done by domain inversion in the crystal, thereby changing the sign of the nonlinear coefficient. The distance of two Lc is called period of the crystal (L), and can be made within a wide range of lengths from a few micrometers to hundreds of micrometers. However, QPM only partially compensates for the phase mismatch. This does not mean that overall conversion efficiency will be lower for QPM. With birefringent phase-matching, the idler, signal and pump waves will often have different paths inside the crystal, limiting the range in which the beams overlap. Within QPM the waves propagate collinear, resulting in a maximum overlap between the waves. This means that very long crystals can be used with an effective interaction zone over their full length. Another advantage of QPM is that very high nonlinear coefficients can be used. With birefringent phase-matching, the propagation direction through the crystal is determined by the angle required for phase-matching. Generally this means that the crystal direction with the highest nonlinear coefficient cannot be used. With QPM, this problem does not exist, resulting in the use of much higher nonlinear coefficients. Furthermore, with QPM it is also possible to use nonlinear crystals that are not birefringent but have a very high nonlinear coefficient, such as GaAs and ZnSe.
Figure 3: Quasi phase matching in a periodically poled nonlinear crystal. The induced light field (P) undergoes a phase shift of p at every change in poling, thereby staying in phase with the pump light field (E). With quasi phase matching a special crystal is used, which has a characteristic poling period in the crystal axis of L which is twice longer than the coherence length Lc in the crystal axis. Coherence length is determined as a distance when P and E waves will be out of phase. This gives rise to an extra term in the phase-matching condition. By poling crystals with the right value of Lc the phase-matching condition can always be met for any combination of signal and idler photons.