1. Introduction

Lasers in the visible spectrum have over the past decade found an increasing number of applications, e.g. within the health care industry, for spectroscopy and imaging use. Many wavelengths in the visible spectrum can be reached through second harmonic generation of efficient solid-state lasers both in continuous wave and pulsed operation. In the CW regime efficient light sources based on sum-frequency mixing have been demonstrated [1–3] using various power enhancement schemes. In the pulsed regime sum-frequency mixing of synchronized Q-switched pulses has been demonstrated [4, 5]. Raman shifted pulsed lasers based on solid-state [6] or fibre lasers [7] have been realised. Alternatively cascaded nonlinear frequency conversion of Q-switched lasers has been used for the generation of pulsed light in the visible spectrum [8, 9].

In the present paper, a novel generic approach for generation of pulsed light in the visible spectrum based on sum-frequency mixing of the circulating intracavity power of a high finesse CW laser and a single pass Q-switched laser is proposed. The fundamental idea is that the high circulating intracavity power of an efficient IR laser can be cavity dumped through nonlinear frequency conversion into the visible spectrum as a high peak power pulse from a second laser is propagating through the nonlinear medium. The nonlinear loss due to sum-frequency mixing is proportional to the injected peak power of the single pass pulses. Therefore, the nonlinear loss is zero between the pulses, but in the presence of an injected pulse the nonlinear conversion can be highly efficient, thus essentially cavity dumping the entire circulating intracavity field.

The advantage of the proposed system is that no synchronization is needed between the two lasers since one of them is operated as a continuous wave laser. The pulsed laser can be realised either as a passively or actively Q-switched system. The duration of the sum-frequency generated pulses is controlled by the pulse duration of the Q-switched laser or ultimately by the cavity roundtrip time of the CW laser, provided that the nonlinear coupling is sufficiently strong to dump the entire circulating intracavity field within a single pass of the cavity. It should be noted, however, that the generated peak power in the visible is limited by the circulating power of the CW operated laser multiplied by the photon energy ratio.

2. Setup

We present in this paper one possible realization of the generic nonlinear cavity dumping approach. It is based on sum frequency generation, SFG, between a high finesse folded cavity 1342 nm Nd:YVO₄ laser and a passively Q-switched 1064 nm Nd:YAG laser using periodically poled KTP, PPKTP, as shown in Fig. 1.

 

Fig. 1. Schematic setup of the suggested generic approach for generation of visible laser pulses.

Download Full Size | PDF

The 1342 nm laser comprises a 8×3×3 mm3 Nd:YVO₄ crystal with a Nd doping of 0.5 atm%, a coupling mirror M2 (planar), a mirror M3 (r=-100 mm) and an end-mirror M4 (r=-50 mm). Mirrors M2, M3 and M4 and the left end facet of the laser crystal are coated for high reflection at 1342 nm, thereby forming a high finesse cavity.

The Q-switched 1064 nm laser consists of a 5 mm long Ø 3 mm Nd:YAG crystal (1.0 atm%), a Brewster plate, BW, for polarisation control, a Cr:YAG saturable absorber, QS, acting as passive Q-switch and an output coupling mirror, M1 (r=500 mm, R=84 % @ 1064 nm). The left end facet of the Nd:YAG crystal and the output mirror form an unstable cavity, supporting a stable Gaussian mode when sufficient thermal lensing is induced by the pump diode. Two different saturable absorbing crystals having a small signal transmission of 80 % and 90 %, respectively are used.

The mirrors M2 and M4 are coated for high transmission at 1064 nm, whereas the steering mirror M5 and the folding mirror M3 are coated for high reflection at 1064 nm. Lens, L, is used for mode matching of the 1064 nm laser into the nonlinear crystal. Both beam waist radii are approximately 40 µm. The nonlinear material used in this setup is a 20 mm long PP:KTP, with a QPM period of 12.65 µm. The PP:KTP was mounted in a peltier controlled oven in order to adjust temperature for optimum phase matching. Both lasers are pumped by 4W@808 nm, RPMC-3415-808, diode lasers.

3. Measurements

First the performance of the Q-switched 1064 nm laser was evaluated. Experimental data of the repetition rate as a function of pump power are shown in Fig. 2. As expected, the repetition rate increases as a function of the pump power. The saturation effect seen in the measurements is due to changing beam parameters caused by increased thermal lensing. The transverse intensity distribution is, however, Gaussian at all pump power levels. The full width half maximum (FWHM) pulse duration is almost constant at 7 ns for all pump power levels. The peak power remained constant at approximately 10 kW independent of the pump power, when using the 80 % saturable absorber.

 

Fig. 2. Measured operation of the Q-switched 1064 nm laser. Left graph shows the pulse repetition rate as a function of diode pump current. Right graph shows a typical trace of a single pulse and multiple pulses at a diode pump current of 3.5 A, using the 80 % Cr:YAG saturable absorber.

Download Full Size | PDF

An intracavity circulating power of approx. 50 W was obtained for the 1342 nm laser in the absence of 1064 nm pulses. It is known from CW experiments [3], that a conversion efficiency of more than 1 %/W/cm can be obtained. The nonlinear conversion of the 1342 nm intracavity field into the generated 593 nm, scales linearly with the peak power of the single pass laser. Thus even modest peak power of the single pass Q-switched laser will be sufficient to dump the circulating field of the 1342 nm laser. The result is a visible light pulse with a duration corresponding to a few roundtrip times of the high finesse cavity.

The left part of Fig. 3 shows a measurement of the circulating power in the 1342 nm laser when a Q-switched 1064 nm pulse is single passed through the intracavity nonlinear material. It is clearly seen that the circulating 1342 nm power is effectively dumped by the 1064 nm pulse. It should be noted that the cavity roundtrip length of the 1342 nm laser is 885 mm corresponding to a roundtrip time of approx. 3 ns, while the 1064 nm pulse has a duration of 7 ns. This means that the intracavity power is effectively dumped within 2 roundtrips. The graph also shows the characteristic relaxation oscillations of the 1342 nm laser as the circulating field returns to its steady state value [10]. It is clear from this graph that the average power of the generated visible light can be optimized by proper adjustment of the repetition rate of the 1064 nm laser. This can be accomplished by having the next Q-switched pulse arrive just after the 1342 nm has settled at steady state (appr. 50 µs) or even at the point where the 1342 nm laser reaches its first maximum of the relaxation oscillation. See Fig. 3.

The right hand side of Fig. 3 shows the average yellow power as a function of diode pump current of the 1064 nm laser. Note, however, the average power has by no means been optimized in the present setup. Comparing the right part of Fig. 3 to the left part of Fig. 2, it is clearly observed that the average visible power scales linearly with the repetition rate of the Q-switched pulses, see the insert in Fig. 3. The reason for the lower average power as a function of repetition rate using the 90 % Cr:YAG, is that the peak power is insufficient to completely dump the circulating 1342 nm power.

 

Fig. 3. Measured power from the 1342 nm laser when a high peak power 1064 nm pulse is single passed through the intracavity nonlinear material. Right graph shows the average yellow power as a function of diode pump current for the 1064 nm laser.

Download Full Size | PDF

Looking closer at the time evolution of the three interacting fields, Fig. 4 shows normalized power levels for all three wavelengths. In the left graph the 1064 nm pulses have been attenuated to a peak power of approx. 6 W, resulting in a modest decrease in the circulating 1342 nm power, as indicated by the black curve. When the peak power of the Q-switched 1064 nm laser is increased, the cavity dumping of the circulating 1342 nm power is seen to be increasingly efficient. In the right part of Fig. 4 it is even seen that all the 1342 nm power is dumped in the form of a visible pulse before the trailing edge of the 1064 nm pulse has passed the nonlinear medium. Thus the peak power at 593 nm is reached before the peak power of the 1064 nm pulse is reached.

 

Fig. 4. Measured power of the three interacting fields at 1342 (Black curve), 1064 (Red curve) and 593 nm (Green curve). In the left graph the 1064 nm peak power has been attenuated to 6 W of peak power. In the right graph the 1064 nm peak power at the position of the nonlinear crystal was attenuated to 2.4 kW

Download Full Size | PDF

Increasing the peak power of the 1064 nm laser is seen to enhance this phenomenon further. However, this increase in the 1064 nm peak power does not result in increased 593 nm peak power as seen in the right part of Fig. 5. In fact, both peak power as well as average power of the generated visible light saturates completely after a certain peak power level of the 1064 nm laser is reached. This level corresponds to the point where all the circulating 1342 nm power has been converted into a 593 nm pulse, i.e. coupled out of the cavity. The only ways to increase the visible average power further is by increasing the circulating power of the CW laser or increasing the repetition rate of the single pass laser. The peak power necessary to reach the saturation point depends on the beam size, the mode overlap in the nonlinear crystal and the effective nonlinearity of the nonlinear material. The right part of Fig. 5 shows that the saturation point in the present configuration is reached at a peak power level of less than 1 kW.

 

Fig. 5. Measured power of the three interacting fields at 1342 (Black curve), 1064 (Red curve) and 593 nm (Green curve) at a 1064 nm peak power of 6 kW (left). The secondary peak on the green trace is residual 1064 nm pulse reaching the detector for the yellow light. The graph to the right shows saturation of the 593 nm peak power as a function of 1064 nm peak power.

Download Full Size | PDF

The measurements shown in the right part of Fig. 5 were obtained using reflection type ND filters to attenuate the 1064 nm peak power, without changing the mode overlap in the nonlinear crystal.

4. Conclusion

A novel generic approach for generation of pulsed radiation in the visible part of the spectrum is described. The approach has been exemplified in a specific setup generating pulses in the yellow part of the spectrum with peak powers of more than 100 W. The system has not yet been optimized in terms of average power, but this can easily be accomplished by increasing the repetition rate of the pulsed single pass laser system. An increased repetition rate would result in a proportional hereto average power increase without affecting the peak power of the visible pulses. Secondly, reducing the passive loss of the 1342 nm laser cavity would result in increased intracavity power and thus increased peak power of the visible pulses. Relatively low peak power is needed in order to saturate the nonlinear conversion, meaning that a Q-switched laser operating at a higher repetition rate and reduced peak power would improve the overall efficiency of the system.

A system optimized for overall efficiency is believed to be as efficient as the corresponding CW system corresponding to a 10% conversion efficiency form diode pump to visible output power [3]. However the suggested approach benefits from its simplicity, since e.g. the spectral overlap between two CW lasers can be difficult to control in the presence of a strong nonlinear coupling. In the proposed system the effective nonlinear coefficient of the nonlinear medium is of less importance, since the peak power of the single pass laser necessary to reach the depletion point is easily attained. In the suggested configuration low losses and high damage threshold of the nonlinear material is of higher importance than the high conversion efficiency, a nonlinear material such as a type I non-critically phase matched LBO, might be an appropriate choice of material for this configuration.