1. Introduction

High quality light sources with specific wavelengths within the visible range are needed for many applications, e.g. dermatology, excitation of biological makers, autofluorescence and for RGB imaging. Due to a limited selection of wavelengths within the visible range, often the application has to be fitted to existing light sources rather than the opposite. The result may be an inferior design or increased development time. The traditional route to the visible spectrum has been through second harmonic generation (SHG) of efficient solid-state lasers, tuning has been obtained using optical parametric oscillators (OPOs), or by use of inefficient dye or gas lasers. These approaches are in general complicated to implement, costly and inefficient, particularly in the CW regime.

In this paper a generic approach for efficient generation of CW light at a predetermined wavelength within the visible spectrum is presented. The suggested approach is based on singly-resonant sum-frequency generation (SFG) of the circulating intra-cavity power of a high finesse diode pumped CW solid-state laser (DPSSL) and the output from a tapered, single-frequency external-cavity diode laser (ECDL). The periodically poled KTP crystal (PPKTP) is situated inside the high finesse DPSSL cavity leading to enhanced conversion efficiency of the nonlinear mixing process [1].

This approach combines desirable properties of different solid-state technologies: the wide range of available diode lasers in the near infra-red (NIR) spectral region, the tunability of tapered ECDLs, the high intra-cavity power of DPSSLs and the flexibility of using quasi phase-matching (QPM).

The generated wavelength is found according to the energy conservation law, Eq. (1). By appropriate choice of solid-state laser wavelength, λ_DPSSL, and the ECDL centre wavelength, λ_ECDL, any specific wavelength in visible spectrum, λ_SUM, can be reached.

One possible realisation of the suggested generic approach is demonstrated in the following. The setup is based on intra-cavity SFG of a high finesse folded cavity 1342 nm Nd:YVO₄ laser and a single-frequency tapered ECDL with a central wavelength of 765 nm. Using PPKTP more than 300 mW of blue light at 488 nm is generated.

Generation of blue light utilizing sum-frequency mixing of DPSSLs and laser diodes in singly-resonant configurations has previously been demonstrated [2–4], however, the output power presented here, is to the authors knowledge one order of magnitude higher than what has previously been reported. This is due to the implementation of a tapered semiconductor device in the presented design.

2. Setup and laser characteristics

The setup used in the experiment is shown in Fig. 1. It consists of a single-frequency 765 nm external-cavity tapered diode laser, a folded cavity 1342 nm solid-state laser, and a 10 mm long intra-cavity PPKTP crystal.

 

Fig. 1. Schematic of the experimental setup. The 765 nm beam from a tapered laser diode (TD) in Littrow configuration is single-passed through a PPKTP crystal placed in the beam-waist of a high-finesse 1342 nm laser for efficient SFG into the blue spectral region.

Download Full Size | PDF

The 1342 nm solid-state laser comprises an 8 mm long a-cut Nd:YVO₄ crystal with a Nd doping of 0.5 atm%. The high finesse cavity is formed by three mirrors: M1 (plane end surface of the laser crystal), M2 (r=-100 mm) and M3 (r=-150 mm). M1 is coated for high reflection at 1342 nm and high transmission at 808 nm. Mirrors M2 and M3 are coated for high reflection at 1342 nm and high transmission at 765 nm. Mirror M2 serves as the input coupling mirror for the 765 nm beam. The distance between M1 and M2 is 200 mm, and the separation of M2 and M3 is 208 mm. The 1342 nm cavity forms a beam-waist of 63 µm and 54 µm in horizontal and vertical axis, respectively, between mirrors M2 and M3 inside the Brewster’s cut PPKTP. The position of the beam-waist is 54 mm from mirror M2. The beam intensity profile was measured to be nearly Gaussian when pumped by up to 2.3 W of 808 nm diode power.

The intra-cavity losses of the 1342 nm laser are determined from slope efficiency measurements, using two partly reflective (PR) mirrors at the position of M3 (T=1.35% and T=3.5%). From these measurements the passive round-trip loss of the 1342 nm cavity is found to be as low as ^P αP₁₃₄₂=0.34 %.

The intra-cavity power of the 1342 nm laser is measured using a partly reflecting mirror (T=1.35%) placed instead of the HR mirror M3. The small fraction of the 1342 nm beam transmitted through the HR coated mirror M2 is measured. The ratio between the transmitted power of M2 and M3 is used for determining the actual transmission coefficient of M2. Thus the power leaked through mirror M2 can be used to estimate the intra-cavity power of the high finesse cavity. Special care is taken to avoid detection of any other signal but the 1342 nm beam. Intra-cavity powers of up to 200 W were measured when the 1342 nm laser was pumped with 2.3 W of 808 nm light and with the 765 nm beam blocked. When generating 230 mW of blue light the intra-cavity power drops to around 180 W because of the loss introduced by nonlinear conversion.

The tapered diode laser is used in a standard Littrow configuration [5] (see Fig. 1). The rear end of the AR coated tapered diode receives feedback from a reflective grating with 1800 lines/mm. The result is single-frequency operation (sub MHz) and the option for wavelength tunability of the output beam of ±6 nm. The maximum output power is 1.4 W at a drive current of 3.1 A. The output beam is collimated by two lenses, CL, and passed through a Faraday isolator (FI). Mirrors CM1 and CM2 are used for alignment and the lens, L, focuses the beam into the nonlinear PPKTP crystal. The focal length of the lens is f=100 mm.

The beam-quality parameters of the tapered ECDL along the horizontal and vertical axes at the position of the nonlinear PPKTP crystal are measured to be M²_H=1.48 and M²_V=2.1, respectively, at an output power of 1.4 W. However, only 68% of 765 nm power from the laser reaches the nonlinear crystal due to losses in the Faraday isolator and uncoated lenses. The 765 nm laser is linearly polarized in the horizontal plane and parallel to the polarization axis of the 1342 nm laser.

The spot size (radius) of the 765 nm laser at the beam waist of the 1342 nm cavity is 49 µm and 15 µm in the horizontal and vertical plane, respectively. Furthermore, the spot size in the PPKTP crystal is magnified to 88 µm in the horizontal plane, as a result of the oblique angle of incidence at the Brewster cut PPKTP. Considering these spot sizes and the measured M² values, focusing parameters of ξ_H=0.23 and ξ_V=11 are found in the horizontal and vertical plane, respectively. Optimal focusing of the 765 nm would correspond to a focusing parameter close to unity [6]. A significant increase in the nonlinear coupling efficiency is therefore to be expected from an improved mode-matching of the fundamental 765 nm beam.

Figure 2 shows the spectra emitted from the two lasers. Note that the linewidth of the ECDL is smaller than the resolution of the instrument (5 pm). The wavelength stability of the 765 nm ECDL was measured over a period of one hour. The maximum deviation from the nominal wavelength during this period was measured to be 5 pm, again limited by the resolution of the spectrometer. The spectrum of the 1342 nm laser contains multiple peaks which are believed to be etalon effects in the 8 mm long Nd:YVO₄ crystal. The spectrum of the 1342 nm laser is not stable over time, but changes in a slowly manner.

 

Fig. 2. (left) Spectrum of the 765 nm tapered diode laser and (right) spectrum of the highfinesse 1342 nm laser.

Download Full Size | PDF

The power stability of the 765 nm ECDL was measured over a period of one hour. After reaching thermal equilibrium, power fluctuations of less than 1 % were measured.

3. Sum-frequency generation

The PPKTP crystal is temperature controlled using a peltier element. Optimizing the temperature of the PPKTP crystal for maximum conversion efficiency a phase-match temperature of approx. 42.5 °C was found. This resulted in a maximum of 308 mW of blue light at 488 nm when pumping the Nd:YVO₄ laser crystal with 2.3 Watts of 808 nm pump power corresponding to a drive current of 2.5 A. The tapered ECDL power incident on the PPKTP crystal was 950 mW, corresponding to a maximum drive current of 3.1 A. Thus the conversion efficiency of the incident 765 nm power to 488 nm blue light was 32% at this pump level for the 1342 nm mixing module.

When the two incident beams can be considered non-depleted, a linear relation exists between the fundamental powers and the generated power.

where P ₇₆₅ is the power of the 765 nm beam, P _1324,circ is the circulating steady state intra-cavity power at 1342 nm, P ₄₈₈ is the power of the generated beam, and η_SFG is the SFG constant. The SFG constant depends on beam alignment, phase-matching, the interacting wavelengths, spot sizes and the nonlinear crystal characteristics. A theoretical estimate of this constant is found to be 0.008 W^-1[7], when using a beam waist of 60 and 50 µm for the 1342 and 765 nm beams respectively, and a d ₃₃=9.8 pm/V. This estimate is 4.5 times higher than the measured value of 0.0018 W^-1 (using a circulating 1342 nm power of 180 W), indicating that the system can be further optimized.

Figure 3 (left) shows the power of the generated blue light as a function of the injected 765 nm power, when changing the current of the tapered ECDL from 0 to 3.1 A. An almost linear dependency is seen as expected from Eq. (2). The deviation from a straight line is believed to be due to small changes in the intensity profile of the focused 765 nm beam as the diode laser current changes. The change in the intensity profile in turn alters the efficiency of the SFG process. The conversion efficiency of the 765 nm beam to the 488 nm beam in the present configuration is found to be approx. 32%, independently of the power of the 765 nm beam.

The output power at 488 nm as a function of pump power of the 1342 nm laser is shown in Fig. 3 (right). The cavity was optimised at a pump power of 2.3 W. Varying the pump power changes the thermal-lens in the laser material, thus displacing the beam-waist as well as perturbing the lasing spectrum of the 1342 nm laser. These effects resulted in the nonlinear characteristics of the curve shown in Fig. 3 (right).

 

Fig. 3. (left) Power of the generated 488 nm light as a function of the injected 765 nm power when P ₈₀₈=2.3 W. (right) Power of the generated 488 nm light as a function of 808 nm pump power, P ₈₀₈, for the 1342 nm laser and with P ₇₆₅=950 mW incident on the nonlinear crystal.

Download Full Size | PDF

The spectrum of the generated blue light at the optimal PPKTP temperature is shown in Fig. 4 (left). Comparing with Fig. 2 (right), a strong resemblance to the asymmetrical spectrum of the 1342 nm laser having two clusters of characteristic peaks is clearly seen. Narrowing the spectrum of the 1342 nm laser, e.g. by inserting an etalon in the cavity, could improve the output spectrum at 488 nm. The wavelength acceptance bandwidth of the PPKTP crystal is calculated to be 0.2 nm [8], which is considerably narrower than the spectrum of the 1342 nm laser. Thus, an increase in conversion efficiency is also expected from narrowing the 1342 nm spectrum.

The measured 488 nm output power as a function of PPKTP temperature is shown in Fig. 4 (right, black squares). The measured FWHM temperature acceptance bandwidth is approx. 3.5 °C. Theoretically, a FWHM temperature acceptance bandwidth of 3.0 °C is found, see Fig. 4 (right, green curve) [9–11]. The increased bandwidth seen in Fig. 4 (right) is primarily due to the two clusters in the spectrum of the 1342 nm laser seen as shown in Fig. 2 (right).

 

Fig. 4. (left) Spectrum of the generated 488 nm light and (right) normalized 488 nm output power as a function of the temperature of the PPKTP crystal (black squares). The green curve shows the theoretical acceptance temperature assuming spectral delta-functions. The corresponding change in the intra-cavity 1342 nm power (red circles) is also shown.

Download Full Size | PDF

Figure 4 (right, red circles) shows the circulating intra-cavity power of the 1342 nm laser as a function of the PPKTP temperature. Since the phase-match condition is a function of temperature, also the intra-cavity power changes, since a part of the 1342 nm photons are converted into 488 nm photons. The nonlinear loss at 1342 nm, α^SFG ₁₃₄₂, is proportional to the generated 488 nm power, P ₄₈₈. At optimum phase-match temperature the intra-cavity 1342 nm power is decreased by approx. 10 % (P ₁₃₄₂,PM=182 W) compared to the non-phase matched power (P ₁₃₄₂,NPM=200 W). The nonlinear loss at 1342 nm can be found, as:

where η_Q is the ratio between the generated wavelength 488 nm, and the laser wavelength 1342 nm. When the PPKTP temperature scan was performed, the phase-matched output power was P ₄₈₈=229 mW.

Knowing the nonlinear loss, it is possible to calculate the passive loss of the 1342 nm laser, using a different approach compared to the previously described method which relied on the slope efficiencies. Since the pump power of the 1342 nm laser is kept constant throughout the temperature scan, the following relation exists between the circulating 1342 nm power and the nonlinear loss:

This confirms that the passive intra-cavity loss α^P ₁₃₄₂ is indeed very small, thus resulting in the very high circulating intra-cavity power reported here. The actual difference between the 0.34% passive loss of the 1342 nm cavity calculated using the slopes efficiency approach and the 0.42% passive loss found using the change of intra-cavity 1342 nm power due to the nonlinear conversion we attribute primarily to measurement errors, but the difference may also be explained by an additional scattering loss induced by thermal lensing of the blue light.

In addition to the expected decrease in intra-cavity power proportional to the generated sum-frequency signal, there are some additional drops in circulating power at 44–46 °C. The reason for this decrease in power is not know, but it may be caused by thermal effects in the PPKTP crystal thereby changing the spectral or the spatial properties of the 1342 nm laser. This is to be investigated further.

The stability of the 488 nm beam was measured, showing a slow power variation of less than 3 %. The fluctuation is ascribed to an instability of the intra-cavity power of the DPSSL, primarily due to mechanical changes in the setup. Finally, in spite of very high intensities inside the PPKTP crystal, no grey-tracking was observed after several hours’ of operation at maximum currents.

4. Conclusion

More than 300 mW of coherent 488 nm light was generated by sum-frequency mixing in an intra-cavity singly-resonant configuration using a diode pumped solid-state 1342 nm Nd:YVO₄ laser and a single-frequency tapered ECDL at 765 nm. A power conversion efficiency of 32 % of the 765 nm beam to 488 nm was achieved in the mixing module.

Furthermore, it was shown that the power conversion efficiency at a constant 808 nm pump level of the mixing module is practically independent of the injected power of the 765 nm beam. Thus, the technique is equally efficient for high and low-power seed lasers, as long as the small signal regime is maintained. This is in strong contrast to SHG.

There are several ways to improve the experimental realization of the system. The mechanical, thermal and spectral properties of the system can be significantly improved, and it is believed that power scaling to the Watt level of the suggested system is straightforward, since no saturation effects have been observed.

In conclusion, we have demonstrated the possibility to synthesize visible light at a predetermined wavelength using the suggested generic approach. The particular system generates light at 488 nm, but by using a tapered diode laser with an appropriate wavelength or by using the different solid-state laser lines, combined with the flexibility of using PPKTP makes it possible to generate light efficiently at virtually any desired wavelength within the visible spectrum.