We report a continuous-wave Nd:GdVO4 self-Raman laser generating wavelength-selectable output in the green-yellow spectral region. The laser combines stimulated Raman scattering (SRS) with intracavity second harmonic and sum-frequency generation (SHG/SFG) to enable “on-demand” operation at any one of three wavelengths, 532 nm, 559 nm and 586 nm, each with output power greater than 3.5 W. Using experimental and numerical modelling results, we show how the highly-interactive, simultaneous non-linear processes of SRS and SHG/SFG may be balanced to achieve stable, selectable multi-wavelength visible operation.
©2010 Optical Society of America
Stimulated Raman scattering, arising from the third-order nonlinear susceptibility of an optical medium, was amongst the first nonlinear optical processes observed following the discovery of the laser . Over the past 40 years, SRS in gas, liquid and solid media has been studied extensively, but despite important advances [2,3], its practical application has been limited, notably to frequency-conversion of high-power pulsed lasers in high-pressure H2 or D2 gas . In contrast, the second-order nonlinear processes of SHG and SFG in crystalline materials (particularly KTP, BBO and LBO) have become ubiquitous in frequency-conversion of solid-state lasers.
While the potential for efficient frequency conversion of Q-switched solid-state lasers using SRS in crystalline materials was first demonstrated by Amman in1975 , another 25 years of Raman material and laser development was necessary before SRS in crystals emerged as a practical means of wavelength extension [6–8]. Achieving efficient continuous-wave (CW) operation of crystalline Raman lasers is particularly challenging due to the very high (kW) circulating powers required at the fundamental wavelength. The first report of a CW crystalline Raman laser was made only in 2004 , but in the past two years we have demonstrated all-solid-state CW visible laser sources based on simultaneous intracavity SRS and SHG in crystalline materials with output powers over 2 W at single wavelengths in the yellow, with diode-visible conversion efficiencies over 15% .
SRS offers great potential for wavelength flexibility and switching. While it is possible to achieve different fundamental wavelengths in the infrared by using different laser host materials and dopants, Raman-active materials can also be used to enable frequency shifting, typically ranging from 200 to 1330 cm−1. In 2005, Mildren et al.  combined the processes of SRS and SFG in a pulsed (Q-switched), Nd:YAG/KGW Raman laser to demonstrate wavelength-selectable emission across four visible lines, with average output power up to 1.8 W.
In this letter we report a compact and efficient CW laser source which uses simultaneous intracavity SRS and SHG/SFG in an otherwise standard diode-pumped Nd3+ laser to deliver multi-Watt output, selectable at three wavelengths (532, 559 and 586 nm as denoted by the solid arrows in Fig. 1 ) across the green-to-yellow spectrum. The Raman-activity of the Nd:GdVO4 laser crystal itself is exploited to generate an intense first-Stokes Raman intracavity laser field, and efficient frequency conversion to visible output is achieved by phase-matching the intracavity LBO crystal for SHG or SFG of the fundamental and/or first-Stokes wavelengths.
2. Design concept for a Wavelength selectable CW visible laser
The concept of simultaneous intracavity SRS and SHG/SFG to generate multiple, selectable wavelengths is illustrated in Fig. 1. For an optical cavity incorporating laser and Raman media and high-reflectivity cavity end-mirrors, the fundamental, first-Stokes, and higher-order Stokes wavelengths can resonate simultaneously, with energy cascading to the highest-order Stokes line. Insertion into the cavity of a nonlinear crystal for which phase-matching conditions may be set for either SHG or SFG of the various circulating fields, allows selection of an ascending range of visible outputs: the second harmonic of the fundamental; the sum-frequency of the fundamental and first-Stokes; the second-harmonic of the first-Stokes; and so on up to the second harmonic of the highest Stokes order resonating in the cavity. In this paper, we focus on the generation of three selectable, visible emission lines (532, 559 and 586 nm), through SHG and SFG of the fundamental and first Stokes wavelengths from a single self-Raman resonator, using mirrors designed to oscillate the fundamental and first-Stokes wavelengths. Using these mirrors, it was not possible to resonate the second-Stokes optical field and achieve red selectable output (denoted as dashed lines in Fig. 1). We have however, recently demonstrated second-Stokes emission at 1308 nm and visible emission at 620 nm from an Nd:GdVO4, intracavity Raman laser  by using mirrors designed to resonate the second-Stokes beam.
The efficiency with which a fundamental optical field can be converted to the first-Stokes or higher optical field order depends critically on cumulative losses within the resonator . Thus, it is essential to minimise resonator losses, by careful selection of resonator mirrors and optical crystals. The spatial mode characteristics of the resonating fundamental and Stokes optical fields are also critical in determining the SRS conversion efficiencies from fundamental to first- and higher-Stokes lines, and SHG/SFG efficiencies. Specifically, small mode diameters within the Raman-active and nonlinear crystal elements maximise the optical field intensities, and thereby the non-linear coupling for both SRS and SHG/SFG. As will be elucidated in this paper, these couplings effectively determine the output power at each visible wavelength; correspondingly, the optimal resonator design is different for each visible wavelength. It follows also that the effects of thermal loading (arising from pumping inefficiencies and the inelastic nature of the SRS process) on the optical properties of the laser and Raman crystals are important and must be considered in the resonator design [13,14].
3. Experimental investigation
3.1 Experimental setup
The laser resonators used in this work had a linear layout as outlined in Fig. 1, and were formed using a flat input mirror and various curved output couplers (radius-of-curvatures of 300 mm, 150 mm and 50 mm were used). These mirrors were all coated for high transmission (T> 98%) at 879 nm and high reflectivity (R >99.994%) at 1063 and 1173 nm. The resonator also incorporated a flat intracavity mirror, coated for high-reflectivity (S2: R> 98%) at 586.5 nm and low reflectivity (S1/S2: R< 0.05%) at 1063 and 1173 nm. A 4 × 4 × 20 mm long, 0.3% at. doped a-cut Nd:GdVO4 crystal, coated anti-reflecting (R<0.2% @ 1063/1173 nm) was used. The lithium borate crystal (LBO) with dimensions 4 × 4 × 10 mm, was cut for non-critical phase matching (NCPM) (θ = 90°, φ = 0) and was anti-reflection coated (R< 0.3%@ 1063 nm/1173 nm). The all crystals were wrapped in indium foil and housed in copper blocks. The Nd:GdVO4 was water cooled to 18°C, while a resistive heater contacted the copper block housing the LBO to enable heating of the LBO from ambient to 160°C. The output from a fibre-coupled 879 nm pump diode (200 µm fibre core diameter, 0.22 NA) was focussed onto the laser crystal with a 430 µm diameter spot size (63 mrad half-angle divergence) using a pair of aspheric lenses with effective focal lengths of 8 mm and 18.3 mm.
The resonator used to demonstrate wavelength selectable output consisted of the flat input mirror, 150 mm ROC output coupler and intracavity mirror. Thin-film optical filters were used to separate the different output wavelengths and the power was monitored using a silicon power meter (Ophir: Nova/10A-SH). The spectral properties of the output were measured using Ocean Optics fibre-coupled spectrometers (USB2000 and HR4000) and the beam quality was measured using a scanning-slit-beam-profiler (Gentec: Beamscope P8).
3.2 Experimental Results
We first consider the maximum powers achievable when the laser resonator was optimised separately for each wavelength (see Table 1 ). Maximum power at 586 nm was 4.3 W corresponding to 17.1% diode-to-visible conversion efficiency. This is almost 2 W higher than previously reported , the improvement being due to improved optical coatings and an improved design for coupling the diode pump light to the resonator. Up to 5.3 W was observed at 559 nm (20.4% optical conversion efficiency), equal to the highest visible power yet reported from a crystalline, CW Raman laser system , and using the same resonator design. Maximum power at 532 nm was 4.6 W and was limited by the onset of “parasitic” SRS which in this case acted as a loss. We suppressed the SRS by inserting into the resonator, a thin BK7 substrate coated to transmit the fundamental and present a substantial loss to the first-Stokes wavelength (~16% per round-trip). This resulted in a significant improvement ( + 2.3 W) in maximum output power, thereby generating 6.9 W at 532 nm.
Next we consider wavelength-selectable operation of the laser using a resonator with output coupler ROC 150 mm and resonator length 39 mm. Only the temperature of the LBO crystal was varied in order to select between the three visible wavelengths. This particular set of resonator parameters was established empirically by way of parametric studies of output power versus resonator length and output coupler ROC. Data showing input/output power scaling characteristics for the wavelength-selectable laser are shown in Fig. 2 . In excess of 3.5 W was achieved for all 3 visible wavelengths, with diode-to-visible conversion efficiencies >11% for each individual line. The rollover in green emission for incident pump powers beyond ~25 W, was again due to “parasitic” SRS (the BK7 intracavity element was not included in this resonator for demonstrating wavelength-selectable operation). The output beam quality parameter (M2) of the outputs were 5.1, 5.16 and 6.87 at 532 nm, 559 nm and 586 nm respectively, at maximum incident pump power.
4. Numerical Modelling
A rate equation model was used to simulate the behaviour of the laser at each wavelength . The model simulates the power flow from the pump diode to the population inversion, on to the fundamental field though stimulated emission, and on again to the Stokes field through SRS. The rate of power flow from the fundamental field to the Stokes field is , where the Raman coupling coefficient is defined as , in which is the Raman gain coefficient, c is the speed of light, is the length of the Raman crystal, is the average mode area in the self-Raman crystal, and L is the optical cavity length.
The desired frequency mixing process is modelled by introducing appropriate loss terms for the fundamental and/or Stokes fields. The powers generated at wavelengths 586 nm, 559 nm and 532 nm are , , and respectively, where the additional factor of four is appropriate for the 559 nm generation that mixes both intracavity fields rather than frequency-doubling a single field. The SFG coupling coefficient is proportional to in which is the effective non-linear coefficient, is the length of the SFG crystal, is the average mode area in the SFG crystal, and γ is a correction factor between 0 and 1 conveying the quality of the phase matching along the crystal length relative to perfect phase matching. This factor is reduced by effects such as temperature gradients and beam divergence within the SFG crystal. Using the known crystal properties and an estimated round trip loss for the fundamental and Stokes fields of 0.8% and 1.25% respectively, we can use the model to predict the output of the laser at each wavelength as a function of the two non-linear coupling strengths and .
We generated contour plots showing the laser output power as a function of the χ(3) and χ(2) coupling strengths (κSRS and κSFG respectively), for maximum incident pump power. These contour plots are shown in Fig. 3(a) –3(c), with our estimate of the operating values of and at full pump power in each laser, derived from measurements of the cavity mode sizes. Note that to enable easy comparison we normalised and to the operating values for the 586 nm laser. If the lasers were all to operate with the same mode sizes for the fundamental and Stokes field, and if the quality of the phase matching was uniform for the three selected processes, then the operating values for the three lasers would be equal. However, two factors lead to the lasers having very different operating values: differences in the thermal loading and the associated lensing lead to different mode sizes in the cavity; and non-uniform temperature control for the SFG phase matching as we move from generating 586 nm to 559 nm to 532 nm with a corresponding increase in SFG crystal temperature from 45.5 to 95 to 155°C.
The 586 nm laser operates best when both non-linear couplings are high. The 559 nm laser, in contrast, requires a lower SFG coupling: too much SFG coupling can entirely suppress the Stokes field as Stokes photons are depleted by SFG faster than the field can build up by SRS. The 532 nm laser ideally needs low SRS coupling, so that the fundamental field is not converted to Stokes field, which in this case acts as a loss mechanism.
The non-linear coupling strengths are only partially determined by fixed laser design factors such as the non-linear crystal properties and lengths, resonator length and mirror curvatures. They also depend on dynamic factors such as mode sizes in the two crystals, the thermal lens experienced by the resonating beams, and phase matching in the SFG crystal. When operating at 559 nm and 532 nm, the laser has significantly reduced SFG coupling, owing to larger mode sizes in the doubling crystal (as a result of a smaller thermal lens), as well as poorer temperature uniformity in the doubling crystal resulting from the elevated temperatures required for phase matching at these wavelengths.
For operation at 532 nm, reduced beam quality of the fundamental lowers the Raman coupling , but it would be advantageous if this could be lowered further to bring the operating point into the higher-power regime. Alternately, we can increase the Stokes losses, suppressing the Stokes field despite the strong Raman coupling. This was verified experimentally with the insertion of the coated intracavity element which was lossy at the first-Stokes wavelength, as discussed in section 3.
We have demonstrated that cascaded intracavity SRS and SHG/SFG of crystalline solid-state lasers can be used to generate selectable, multiple wavelength CW outputs at multi-Watt powers in the visible from a single device. To the best of our knowledge, such operation of a CW solid-state laser has not previously been reported. Compact, efficient, all-solid-state lasers generating multi-Watt powers in the visible are in high demand for a multiplicity of applications, and we anticipate our practical and efficient approach to wavelength versatility could offer unprecedented flexibility in wavelength control.
References and links
1. E. J. Woodbury and W. K. Ng, “Ruby operation in the Near IR*,” Proc. Inst. Radio Eng. 50, 2367 (1962).
4. A. Glass, “Design considerations for Raman lasers,” IEEE J. Quantum Electron. 3(11), 516–520 (1967). [CrossRef]
5. E. O. Ammann and J. Falk, “Stimulated Raman scattering at kHz pulse repetition rates,” Appl. Phys. Lett. 27(12), 662–664 (1975). [CrossRef]
6. T. T. Basiev, and R. C. Powell, “Solid-state Raman lasers,” in Handbook of Laser Technology and Applications Editor: Webb, C. E., ch. B1.7 (The Institute of Physics, London, Bristol, UK, 2003).
7. P. Cerný, H. Jelinkova, P. G. Zverev, and T. T. Basiev, “Solid state lasers with Raman frequency conversion,” Prog. Quantum Electron. 28(2), 113–143 (2004). [CrossRef]
8. J. A. Piper and H. M. Pask, “Crystalline Raman lasers,” IEEE Sel. Top. Quantum Electron. 13(3), 692–704 (2007). [CrossRef]
9. A. S. Grabtchikov, V. A. Lisinetskii, V. A. Orlovich, M. Schmitt, R. Maksimenka, and W. Kiefer, “Multimode pumped continuous-wave solid-state Raman laser,” Opt. Lett. 29(21), 2524–2526 (2004). [CrossRef] [PubMed]
10. A. J. Lee, H. M. Pask, P. Dekker, and J. A. Piper, “High efficiency, multi-Watt CW yellow emission from an intracavity-doubled self-Raman laser using Nd:GdVO4.,” Opt. Express 16(26), 21958–21963 (2008). [CrossRef] [PubMed]
13. D. J. Spence, P. Dekker, and H. M. Pask, “Modeling of continuous wave intracavity Raman lasers,” J. Sel. Top. Quantum Electron. 13(3), 756–763 (2007). [CrossRef]
14. P. Dekker, H. M. Pask, D. J. Spence, and J. A. Piper, “Continuous-wave, intracavity doubled, self-Raman laser operation in Nd:GdVO(4) at 586.5 nm,” Opt. Express 15(11), 7038–7046 (2007). [CrossRef] [PubMed]
15. A. J. Lee, H. M. Pask, D. J. Spence, and J. A. Piper, “Efficient 5.3 W cw laser at 559 nm by intracavity frequency summation of fundamental and first-Stokes wavelengths in a self-Raman Nd:GdVO4 laser,” Opt. Lett. 35(5), 682–684 (2010). [CrossRef] [PubMed]