1. Introduction

Recent availability of large Nd:YVO₄ and Nd:GdVO₄ crystals with low Nd³⁺ doping concentration (<0.5at.%) has renewed interest in power scaling for efficient IR and visible laser operation [1–3]. The high heat loading of standard Nd:YVO₄ crystals at high pump powers (>5W) leads to gross thermal distortions and crystal fracture. Operation on the longer wavelength 1342nm Nd³⁺ transition further increases thermal loading of the crystal due to the increased quantum defect between the pump and laser wavelength. Nevertheless Nd:YVO₄ and Nd:GdVO₄ are favoured for operation at 1342nm since the emission cross-section is roughly equal to that of the 1064nm transition in Nd:YAG [4].

Approaches to reducing thermal loading for Nd:vanadate lasers include several key techniques (or combinations thereof) such as side-pumping schemes [5], reduced quantum defect pumping using 885nm diodes [6], use of GdVO₄ as the laser host material and use of longer crystals with lower Nd³⁺ doping. The high absorption coefficient of Nd:vanadate has enabled very efficient operation in side-pumping regimes, however astigmatic thermal lensing and poor beam quality reduce the attractiveness of these techniques. High-efficiency pumping with 885nm diodes is very promising, but the current cost and limited availability of high power 885nm diodes greatly limits this potential. Nd:GdVO₄ is very similar to Nd:YVO₄, except its thermal conductivity is approximately twice as large [6], enabling a higher pump fracture limit and reduced thermal distortion. Use of longer (10–15mm long) crystals with Nd³⁺ doping in the range 0.25–0.5at.% gives substantially reduced heat loading, though mode-matching is generally more difficult than for highly doped crystals.

Q-switching of high power Nd:YVO₄ lasers has been extensively reported for the dominant 1064nm transition [1], however q-switching of the 1342nm transition and related intracavity SHG at high power have received less attention [7–10]. Recently type II non-critical phase matched (NCPM) SHG in LBO has been used intracavity to demonstrate 1.3W of average output power at 671nm [10]. Type I critical phase matching for SHG in LBO has higher non-linear coefficient than the type II process [7], though it is subject to beam walk-off.

Here we report on q-switched operation of a high power, diode end-pumped Nd:YVO₄ laser at 1342nm and intracavity frequency doubling to 671nm. Over 1.7W average power at 671nm was obtained at 100kHz. The maximum repetition rate demonstrated was 350kHz, where the average red output power was 1.34W. Operation at higher repetition rates is expected with increased pump power.

2. Experimental details

The laser configuration is illustrated in Fig. 1. The pump source consisted of two 15W 808nm Optopower diodes coupled into 1mm fibre bundles with a measured NA=0.11. Following the approach of Taira et al [11], the fibre bundle outputs were collimated and focussed through the dichroic mirrors (T~×85% at 808nm) M1 and M2 into either end of the crystal with x0.4 magnification, giving a pump mode radius of ~200µm. To obtain this magnification the confocal length of the pump beam was shorter than optimal (~5.6mm). The a-cut Nd:YVO₄ crystal (supplied by Coretech) had 0.3at.% Nd³⁺ concentration and dimensions of 3×3×12mm. Both crystal faces were AR coated for 808, 1064 and 1342nm. The crystal was wrapped in indium foil and placed inside a water-cooled copper mount, which was held at ~15°. This system was previously optimised for 1342nm generation, producing >8W cw in a TEM₀₀ beam (>40% slope efficiency and >30% optical conversion efficiency) [12].

For the present experiments an acousto-optic q-switch (NEOS N33027-25-2-I) was added to the M2–M3 arm together with the LBO frequency-doubler, which was located close to M3. The LBO crystal, cut for type I CPM at 1342nm (30°C), was wrapped in indium foil and placed inside a copper mount which was held at 30°C using a thermoelectric cooler with active feedback.

 

Fig. 1. Laser pump and resonator configuration.

Download Full Size | PDF

M1, M2 and M3 were all high reflecting mirrors (R>99.5%) at the laser wavelength (1342nm); in combination they also exhibited sufficient transmission (up to 70% per mirror) to suppress lasing on the higher gain 1064nm transition. M3 and M2 were both highly transmissive for the second harmonic (@ 671nm T~88%), so an additional mirror M4 was used to redirect the output from M2 back through M2 to obtain a single red output. All of the mirrors were flat apart from M2, which had a 25cm concave radius of curvature. The total resonator length was ~145mm, the shortest possible length given the physical size of the laser crystal, q-switch and LBO mounts. Under these conditions and in the absence of significant thermal lensing in the laser crystal, the resonator is stable with a small mode size in the LBO frequency doubler.

3. Results and discussion

Figure 2 shows second harmonic (671nm) average output power for q-switch pulse repetition rates between 40 and 350kHz. Average output power at the second harmonic was found to vary by only ~20% over the full range of repetition rates investigated. Maximum 671nm power of 1.7W was obtained at 100kHz and diode pump power incident on the laser crystal of 14W, above which value the thermal lens induced in the laser crystal rendered the laser resonator unstable.

 

Fig. 2. Second harmonic average power variation with q-switch repetition rate.

Download Full Size | PDF

Pulse energy and duration are shown as a function of pulse repetition rates in Fig. 3. Second Harmonic output pulse widths varied between 20–30nsec FWHM at lower repetition rates up to 150kHz, but increase to >90nsec FWHM above 300kHz. As expected, maximum pulse energies were obtained at low prf’s where the interpulse period approaches the fluorescence lifetime (~100µsec). The minimum average pulse energy was 4µJ at 350kHz and the maximum was 37µJ at 40kHz. Damage to the AR coatings on the laser crystal has previously been identified as an issue at repetition rates <40kHz, so such low pulse rates were not investigated in the present experiments. Note that for q-switch rates above 350kHz the laser gain was sufficient to reach threshold only for every second q-switch opening, and the laser repetition rate reverted to half the q-switch rate.

 

Fig. 3. 671nm pulse energy variation with repetition rate.

Download Full Size | PDF

Figure 4 shows second harmonic (671nm) average power as a function of incident pump power for a range of Laser pulse repetition rates. The laser output in continuous (cw) mode reaches maximum at around 12.5W incident pump power, whereas maximum q-switched output power was obtained at around 14W pump power (the thermal loading is approximately the same in both cases since an additional 1.5W is coupled out of the resonator in q-switched operation). Maximum slope efficiency achieved was 17% at 100kHz.

 

Fig. 4. Comparison of laser characteristic with varying pulse repetition rate.

Download Full Size | PDF

As for our previous report of cw 671nm generation [12], cavity mirror transmission and curvatures used were not optimal for second harmonic generation. Given both end mirrors (M1 and M3) were flats and M2 curved, the resonator mode was highly elliptical in both the laser crystal and the LBO: calculated dimensions of the resonator mode radius in the laser crystal were 640×240µm, which was larger than the ~200µm circular pump mode radius.

The induced Nd:YVO₄ thermal lens focal length with approximately 12W absorbed pump power was calculated to be around 36cm (from Chen et al. [1]). This lensing was sufficiently strong enough to render the laser resonator unstable for incident pump powers above 14W. Re-engineering of the q-switch and LBO mounts will allow the resonator to be shortened to approximately 120mm, thus stability could be maintained with a laser crystal focal length as short as 15cm, equivalent to 24W deposited power, for which 671nm output powers exceeding 2.5W are anticipated. Replacement of the 25cm concave turning mirror with a flat mirror will also allow a much stronger laser crystal thermal lens to be tolerated, permitting an even greater pump power. Resonator modeling suggests the original 147mm cavity length and a flat turning mirror will permit a laser crystal thermal lens as short as ~11cm, which would correspond to more than 30W of absorbed pump power, for which we could expect 671nm output approaching the 5W level.

4. Conclusion

We have demonstrated 1.7W of red (671nm) laser output from an intracavity-doubled, q-switched, diode-pumped Nd:YVO₄ laser based on a low-dopant-concentration laser crystal. To our knowledge this is the highest output for an all-solid-state Nd:YVO₄ laser source operating at 671nm. Operation at prf’s from 40 to 350kHz has also been obtained at average powers above 1.3W. Power scaling to the multiwatt level should be achievable with improved mirror selection and resonator layout, and increased diode pump power.