1. Introduction

High power green light sources have numerous applications such as laser displays, material processing, biological investigations, and eye surgery. Due to degrees of freedom in material choice, second harmonic generation (SHG) by the quasi-phase-matching (QPM) technique has been an attractive method to obtain an efficient continuous wave (CW) green laser. Among QPM materials, periodically poled Mg doped stoichiometric lithium tantalate (PPMgSLT) crystal has demonstrated to be the most promising for 10W-level CW green SHG because of a high effective nonlinear coefficient and a high thermal conductivity [1]. A single-pass SHG is an approach for high-power CW green generation, because it has not only the very compact architecture but also high freedom in combination with steadily improving high power infrared lasers such as an Yb-doped fiber laser. Single pass scheme is also the simplest and direct method to determine the performance of the material or device for practical application. Several groups have achieved efficient CW single pass SHG with PPMgSLT crystals [2–6].

The main obstacle in high power green SHG is thermal effects such as thermal lensing and thermal dephasing. Thermal lensing, transverse effect, imposes a limit on the local intensity and thermal dephasing, longitudinal effect, imposes a limit on the overall power [7]. Those problems in SHG crystals are related to several factors, such as linear and nonlinear absorption, thermal conductivity of crystal and contacted material, aperture (width and height) size and focusing parameter. Heat is loaded in SHG crystal by infrared- (IR), and visible absorption, and even green-induced-infrared absorption (GRIIRA). Thermal conductivity of material has been increased by developing ferroelectrics with near-stoichiometric composition. Although thermal limitation is basically related on the inherent thermal conductivity of SHG crystals, the thermal disposal performance in SHG can be improved by introducing macroscopic architecture. When the crystal is set in a metal module, the performance of heat removal can be improved from lateral faces by metal contact with high thermal conductivity, ~100 W/m∙K. In combination, reduction of crystal aperture close to beam diameter can lead to a significant reduction of peak temperature in the crystal [8]. In this paper, we investigated heat disposal performance of SHG modules depending on different crystal widths and concluded the relatively tight aperture crystal is suitable for CW high power green generation. Finally, we demonstrated CW 19 W single-pass green SHG with a tight aperture crystal without saturation.

2. Phase-matched calorimetry (PMC)

Temperature of SHG crystal is commonly controlled by an external temperature controller such as thermoelectric cooler (TEC) for satisfying phase-matching condition. In high power SHG, temperature increase in a crystal mainly comes from the thermal loading by absorption of first harmonic (FH) and SH waves having different absorption coefficients. When the initial phase-matching temperature is disturbed by heat from FH and SH power, TEC temperature changed for optimum SH power indicates the temperature change in crystal. The technique is named as Phase-Matched Calorimetry (PMC) measuring the temperature increase by supplying laser heat, where the accurate temperature measurement is allowed under the phase matching. If an SHG module requires smaller TEC temperature change for optimum SH power, the module has better performance in heat removal. In order to make quantitative analysis on thermal disposal performance, we propose a new fitting equation to express the TEC temperature change as a function of FH or SH power.

Heat capacity, C [J/°C∙m³∙s] is defined as the ratio between the quantity of heat energy Q [J/m³∙s] transferred to the object, and the resulting temperature increase of the object ΔT [°C].

If temperature distribution in the crystal is uniform along the crystal length, TEC temperature change is directly related to the increased temperature of the crystal. But, in practice, the temperature distribution in the crystal strongly depends on increasing SH power [9,10] along the crystal length in high power SHG, because of several times higher absorptivity at SH than that of FH. However, if the non-uniformity is highly suppressed by high thermal conductivity of the crystal and good thermal disposal performance of the SHG module, TEC temperature change can be approximated to the increased temperature of crystals. For an example, Fig. 1 shows typical SHG temperature tuning curves at different FH and SH power measured with a module. Although the used module has relatively lower heat removal performance (C_α = 1.7 W/°C) than that of further modules, no distortion of the symmetry of curves is observed up to multi-W SH power, which guarantees unique temperature under uniform temperature distribution along the crystal length.

 

Fig. 1 SHG temperature tuning curves for a 20 mm-long PPMgSLT device with a QPM period of 8.4 μm.

Download Full Size | PDF

Final fitting equation for TEC temperature based on SH power is as follows.

 

Fig. 2 (Left): SH Power versus FH power for a 10 mm long crystal with a QPM period of 8.4 μm, quadratically fitted to determine η_norm. (Right): TEC temperature versus SH power, fitted with Eq. (3).

Download Full Size | PDF

3. Experiment

We fabricated 1mol% Mg doped PPSLT crystals, while details of the fabrication processes were reported in [4]. The PPMgSLT crystals were set in a metal module to improve the heat disposal at four lateral faces. In order to investigate the heat removal performance depending on boundary conditions, we used three different practical crystals with dimensions of 0.3, 0.4 and 0.5 X 0.5 X 10 (or 20) mm (width, height, length). The fabricated modules have two different QPM periods. We employed modules with a QPM period of 8.4 μm, combining with a CW Yb-doped fiber laser (1083 nm) with a single transverse mode and a maximum output power of 50 W. Another module with QPM period of 8.0 μm was utilized for high power SHG of a CW single-mode and single-frequency Nd:YAG laser (1064 nm) with a maximum output power of 100 W, which was originally developed for a light source of an interferometric gravitational wave detector [12]. The linewidth of Yb-doped fiber laser with M² = 1.0 was 0.17 nm at a maximum output power of 47 W. The Nd:YAG laser with M² = 1.2 presented the linewidth ~3.8 X 10⁻⁹ nm (1 kHz) at a maximum output power of 100 W.

In experiments, we measured and compared TEC temperature as a function of SH power for various crystal width of 0.3 mm and 0.4 mm and width configuration of 0.3 mm and 0.5 mm, respectively. The beam diameter on crystal surfaces was calculated ~100 μm for 10 mm long- and 20 mm long crystals resulting in different focusing parameters. We adjusted the fundamental power by a combination of a half-wave plate and a polarizer, and used high power dichroic mirror to separate fundamental and SH power. The temperature of modules was controlled by a TEC with an accuracy of 0.1 °C.

4. Results and discussion

Measured TEC temperatures versus SH green power for modules with different width crystals are shown in Figs. 3(a) -3(c). Case (a) and (b) are configurations of 0.3 mm and 0.4 mm width. Case (c) is a configuration of 0.3 mm and 0.5 mm width. We measured the SH data for fitting in lower power range to reduce risk of damage. Low heat capacity induces a rapid decrease in PMC curve in the low power range, and gradually shifts to a slow decrease. The initial phase-matching temperature seems to be lower for low heat capacity, when we extrapolate the linear region to the Y axis. We used the devices with the same initial phase-matching temperature, except slightly misaligned case. Each heat capacity is determined by fitting with Eq. (3). In the case of Fig. 3(a), heat capacity of 6.1 W/ °C and 5.6 W/ °C is calculated for 0.3 mm and 0.4 mm, respectively, when focusing parameter, ξ = 2.73. To investigate the dependence of the heat capacity on focusing parameter, we used the same module in Fig. 3(b) with different focusing parameter, ξ = 1.90. Heat capacity value of 8.7 W/ °C and 6.5 W/ °C are measured for 0.3 mm and 0.4 mm, respectively. The higher heat capacity values of case (b) for the relatively looser focusing with bigger beam waist agrees with the newly defined C_α proportional to the size of cross section A. In the case of (c), we obtained the heat capacity of modules with 2 times longer crystals (L = 20 mm) with width of 0.3 mm and 0.5 mm. Heat capacity of 8.9 W/ °C and 5.8 W/ °C are determined for 0.3 mm and 0.5 mm, respectively, when ξ = 0.75. Relative heat capacity values for different crystal widths of 0.3, 0.4 and 0.5 mm are plotted in Fig. 3(d). It is clear that the reduction of crystal width improve the heat capacity of the module. The reduced width leads to the reduction of peak temperature in the cross-section of the crystal due to the more closed metal with high thermal conductivity, ~100 W/m∙K and then reduces the thermal gradient along the crystal length. The relatively weak focusing parameter effect can be neglected in the comparison. Maximum 50% increase of heat capacity was achieved by using 0.3 mm width, compared than 0.5 mm width. On the other hand, optimum size of crystal aperture for the highest heat removal is equal to the fundamental beam size if we assume heat dissipation from four lateral faces. However, the optimum aperture size is limited when we consider SH conversion efficiency. The optimum aperture size for higher heat removal and SH efficiency could be determined by the beam size on the aperture with optimum focusing condition, ξ = 2.84 and crystal length.

 

Fig. 3 TEC temperature versus SH power for (a): 0.3 mm and 0.4 mm width configuration for QPM period, Λ = 8.4 μm, ξ = 2.73, L = 10 mm, (b): 0.3 mm and 0.4 mm width configuration for Λ = 8.4 μm, ξ = 1.90, L = 10 mm, (c): 0.3 mm and 0.5 mm width configuration for Λ = 8.4 μm, ξ = 0.75, L = 20 mm. (d): normalized relative heat capacities versus crystal widths of 0.3, 0.4 and 0.5 mm.

Download Full Size | PDF

Finally, in order to achieve high SH power we employed a module with 10 mm-long crystal with width of 0.3 mm having high heat capacity value of 18.8 W/ °C and a QPM period of 8.0 μm using the Nd:YAG laser at focusing parameter ξ = 1.98. The high C_α results from not only combination of metal module and tight aperture but also use of a wafer with relatively low absorption coefficient. The dependences of the SH power (left-hand scale) and efficiency (right-hand scale) on fundamental power are presented in Fig. 4 . A maximum 19 W at green 532 nm with 26.5% conversion efficiency were achieved at fundamental power of 71.6 W. No signs of saturation in SH power were observed up to the maximum 19 W SH power.

 

Fig. 4 CW 532 nm green SH power and conversion efficiency for a 10 mm long crystal versus fundamental power. Solid line: Tanh² dependence of SH power. Inset: TEC temperature of the used module versus SH power fitted by Eq. (3), result in the heat capacity value as 18.8 W/°C.

Download Full Size | PDF

5. Conclusion

We proposed a new PMC method to determine heat removal performance of SHG modules. The thermal performance of SHG modules was improved by combination of housing PPMgSLT crystal in a metal module and use of a tight aperture with 0.3 mm width. Maximum 50% increase in heat capacity was achieved by using a crystal width of 0.3 mm compared with 0.5 mm. The improved heat removal demonstrated CW single pass 532 nm power of 19 W at an SHG conversion efficiency of 26.5% in a 10 mm-long PPMgSLT crystal without saturation.