1. Introduction

UV solid-state lasers emitting at 266 nm are used in various applications such as microelectronics, precise cutting of some materials (polymer, glasses) which only absorb below 300 nm, mass spectrometry etc. Such UV solid-state devices can be achieved from an infrared fundamental laser light through several stages of frequency conversion using nonlinear optical (NLO) crystals. Borate crystals are often chosen because of their wide transparency in visible and UV range, their good chemical stability and their high damage threshold [1]. Nevertheless, materials which are transparent at 266 nm, which have a phase matching angle for SHG of the 532 nm fundamental wavelenght, and which possess a high conversion efficiency are less numerous. Two main nonlinear crystals are currently commercially available for the fourth harmonic generation (FHG) at 266 nm: β-BaB₂O₄ (BBO) and CsLiB₆O₁₀ (CLBO). Although these crystals demonstrate good performances for UV generation at 266 nm [2–4], they present some practical drawbacks which limit their application as high hygroscopy for CLBO and large walk-off angle as well as high two-photon absorption for normal BBO [5]. Until recently, YAl₃(BO₃)₄ (YAB) single crystal, which is a negative uniaxial crystal and belongs to space group R32, was only studied doped by rare earth ions as Nd³⁺ and Yb³⁺ for self-doubling application at 532 nm. The growth of YAB by Top Seeded Solution Growth (TSSG) using flux based on K₂Mo₃O₁₀ did not allow obtaining transparent crystals below 300 nm. Since 2008, undoped YAB arouse a renewed interest as a nonlinear crystal for applications below 300 nm thanks to the use of new flux (Li₂WO₄-B₂O₃ [6], Li₂O-B₂O₃ [7] and Li₂O-Al₂O₃-B₂O₃ [8]). YAB is a promising nonlinear material because of similar nonlinear properties to CLBO [9], a relative high hardness and chemical stability (non-hygroscopic) which make it easy to cut and polish [10]. So far, only two experimental reports on the use of YAB as NLO crystal for 266 nm generation were published [7,9]. Q. Liu et al [9] obtained 5.05 W average power of a 266 nm UV laser at 65 kHz pulse repetition frequency and 19.5 ns pulse width by performing type I SHG in a YAB single crystal of dimensions 3 x 3 x 6 mm³. They obtained a pulse peak power of 4 kW and a conversion efficiency of 12.3% from 532 to 266 nm.

In this paper, we report the fourth harmonic generation of microchip laser Nd:YAG/Cr⁴⁺: YAG by performing type I SHG in a YAB single crystal. The fundamental wavelength at 1064 nm was first doubled in frequency at 532 nm with a LBO crystal of dimensions 5 x 5 x 10 mm³. The optical transmittance of the YAB crystal over the range 180-1200 nm as well as the optical conversion efficiency from 532 to 266 nm were measured. The influences of optical homogeneity and absorption of YAB on the conversion efficiency are discussed.

2. Crystal characterization

A YAB single crystal was grown using the top seeded solution growth (TSSG) method with La₂O₃-B₂O₃ as a flux [11]. A YAB slab of dimensions 15 x 15 x 2.94 mm³ (Fig. 1) was cut in ZX plane with θ = 66.9 ° and φ = 0° for type I SHG leading to 266 nm generation according to the Sellmeier equations determined by J. Yu et al [8]. The value of the effective nonlinear coefficient d_eff is 0.69 pm/V with d₁₁ = 1.7 pm/V [10]. The crystal was optically polished but uncoated. A large section of 15 x 15 mm² has been chosen for probing in order to evaluate the homogeneity of the crystal. The moderate thickness of the crystal (2.94 mm) is limited by the absorption at 266 nm (see below).

 

Fig. 1 Photograph of the polished slab of YAB single crystal cut in the XZ plane (θ = 66.9°, φ = 0°).

Download Full Size | PDF

The crystal transmittance through a 2.94 mm thick slab cut and polished for FHG was measured over the range 180-1200 nm at room temperature using unpolarized light with a Cary 6000i UV-Vis-NIR spectrophotometer (spectral bandwidth (SBW) is 2 nm). We can observe on the transmittance spectrum (Fig. 2) two weak absorption bands (1, 2) in the visible region at λ₁ = 585 nm, λ₂ = 415 nm and two strong absorption bands (3, 4) in the UV region at λ₃ = 285 nm and λ₄ = 245 nm. Because of the strong absorption in UV, the transmittance drops sharply from T = 81% at 338 nm to T = 38% at 266 nm. Considering the optical loss by reflection calculated using Sellmeier equations [8], the absorption of the crystal at 266 nm is 55% corresponding to an absorption coefficient of 2.7 cm⁻¹. Normally, the cut-off wavelength is about 170 nm [7] but it is 200 nm in this study because of the absorption and the thickness of the sample.

 

Fig. 2 Transmittance at room temperature with unpolarized light of a 2.94 mm thick type I YAB single crystal (θ = 66.9 °, φ = 0°).

Download Full Size | PDF

The absorption bands result from the presence of transition metal impurities in the crystal and more precisely from chromium Cr³⁺ and iron Fe³⁺, which can come from the raw materials and even from the platinum crucible. Fe³⁺ and Cr³⁺ ions can easily substitute Al³⁺ which is 6-fold coordinated in YAB. The two absorption bands in the visible (1, 2) are characteristic of Cr³⁺ ion (Fig. 2) and correspond respectively to d-d transitions: ${}^{4}A{}_{2g}\to {}^{4}T{}_{2g}$ and ${}^{4}A{}_{2g}\to {}^{4}T{}_{1g}$ [12]. The two strong absorption bands (3, 4) at 285 nm and 245 nm are assigned to charge transfer transitions between O^2- and Fe³⁺. They are similar to the ones observed in YAB crystals grown in air using other flux (Li₂O – B₂O₃ [7], Li₂O – B₂O₃ – Al₂O₃ [8]). A similar phenomenon was also observed in others crystals as BaAlBO₃F₂ (BABF) [13] and especially in K₂Al₂B₂O₇ (KABO) where the UV absorption is consistent with a charge transfer between the surrounding ligands O^2- and Fe³⁺ [14].

3. Experimental set-up and laser beams characteristics

The fundamental wavelength at 1064 nm is produced by an amplified diode pumped passively [110]-cut Cr⁴⁺:YAG Q-switched microchip Nd:YAG laser (100Hz rep. rate). The pulse duration is 580 ps with a maximum energy of 1.7 mJ which corresponds to a maximum peak power of 3 MW [15].

The experimental setup used for the fourth harmonic generation with YAB crystal is presented in Fig. 3. A type I LBO crystal, cut in the XY plan (θ = 90°, φ = 11.3°) with dimensions of 5 x 5 x 10 mm³ was used as frequency doubling material to generate the 532 nm laser beam. The choice of LBO for this first stage of conversion is explained by its high crystalline quality and both high angular acceptance (7.4 mrad.cm) [16] and good nonlinear effective coefficient (0.83 pm/V) [17]. Two half-wave plates (λ/2) and a polarizing beam splitter (PBS) are used to control the power and the polarization of the infrared beam which is focused in the LBO crystal through a lens of focal length f1 = 175 mm. The waist of IR laser beam in LBO crystal is 350 μm. An energy of 0.9 mJ at 532 nm with 530 ps pulse width corresponding to a mean conversion efficiency of 53% (energy ratio) was obtained.

 

Fig. 3 Experimental setup for fourth harmonic generation with YAB crystal.

Download Full Size | PDF

Then, two dichroic mirrors M1 coated for HR at 532 nm and HT at 1064 nm are used to filter the residual 1064 nm beam. A silica window (S) was used to take off a part of the beam in order to measure precisely the input energy at 532 nm which value is varied thanks to the use of half wave plate at 532 nm and PBS. The green beam is then focused in the YAB crystal through a lens of focal f2 = 150 mm. The beam quality M² and the waist of the green beam were measured by the knife-edge method and were calculated to be 1.2 and 115 μm respectively (Fig. 4). The UV beam and the residual 532 nm beam are respectively transmitted and reflected to power meters thanks to the dichroic mirror M2 coated for HR at 266 nm and HT at 532 nm. The confocal length (2 x Rayleigh range) is 13.3 cm, so we can consider that for the YAB crystal, the input beam is almost parallel. Knowing that the walk-off of YAB for second harmonic generation of 532 to 266 is 33 mrad [17], the interaction length $L_C\text{,}$which is the maximum length before the beams of the interacting fields separate, can be calculated as follows:

 

Fig. 4 M² and waist measurements for the laser beam at 532 nm.

Download Full Size | PDF

4. Fourth harmonic generation at 266 nm

Measurements of conversion efficiency from 532 nm to 266 nm have revealed significates inhomogeneities in YAB crystals. Some areas in the crystal scatter the laser beam. As we can see from Fig. 5, the quality of the UV beam is variable depending on the position in the crystal where we do the measurement. Moreover, for each position in the crystal, the orientation of the crystal has to be modified in order to maximize the UV energy. Twins which have been already observed in YAB crystals [18–20] could be responsible for these effects. Fourth harmonic generation experiments were, therefore, performed in positions where losses by scattering were negligible.

 

Fig. 5 Photographs of two UV beams at 266 nm taken in two different positions in the crystal.

Download Full Size | PDF

The conversion efficiency from 532 nm to 266 nm and output energy at 266 nm versus input energy and intensity at 532 nm were measured. The results are shown in Fig. 6. Considering the losses by reflection of the crystal at 532 nm and 266 nm, an energy of 113 μJ at 266 nm with 445 ps (Fig. 7) pulse width corresponding to a mean conversion efficiency of 12.2% has been obtained. The pulse width was measured by a 12.5 GHz GaAs detector (ET-4000, Electro-Optics Technology, Inc.) and a 12 GHz oscilloscope (DSO 81204B, Agilent Technology). Assuming a Gaussian pulse shape, the pulse peak power $P$and the intensity $I$at 266 nm can be expressed as follows:

 

Fig. 6 Conversion efficiency from 532 nm to 266 nm and output energy at 266 nm versus input energy and intensity at 532 nm.

Download Full Size | PDF

 

Fig. 7 Pulse width of UV beam at 266 nm.

Download Full Size | PDF

The pulse peak power is 240 kW. If we assume that UV and green waists are similar, UV intensity is 1.15 GW.cm⁻². No damage in YAB crystal was observed for this test.

5. Discussion

The results shown above highlighted a substantial loss of energy$E_{LOSS}$which can be calculated as follows:

 

Fig. 8 $E_{LOSS}$, $E_{266}$ and $E_{532}^R$ versus $E_{532}^0.$

Download Full Size | PDF

For example, the maximum obtained energy of 113 μJ at 266 nm corresponds with a loss of energy of 317 μJ. This loss of energy can have multiple origins:

Losses by scattering can be neglected because the experiment was performed at a position where the scattering was very low. Linear absorption at 532 nm is also negligible because of the low linear absorption coefficient (${\alpha }_{532}$ = 0.069 cm⁻¹). These two latter assumptions were checked by rotating the polarization of 532 nm laser beam by 90°. In this situation, there is no phase matching and no 266 nm generated beam. Then, we checked that $E_{LOSS}=E_{532}^0-E_{532}^R\approx 0.$Two-photon absorption (TPA) is a nonlinear phenomenon which is characterized by the coefficient $\beta$(cm.GW⁻¹) in the following equation:

It is highly likely, therefore, that linear absorption at 266 nm and absorption of two photons at 266 are mostly responsible for the energy loss. Indeed, the linear absorption at 266 nm is about 55% (Fig. 2) so that we can assume that most of the energy loss comes from the linear absorption at 266 nm. Regarding the TPA at 266 nm, it could not be measured directly. Nevertheless, the plot of the transmittance $T=E_{266}/E_{266}^{TOTAL}$ where $E_{266}^{TOTAL}=E_{266}+E_{LOSS}$ versus the total intensity at 266 nm is significant of the occurrence of TPA. In our case, the total intensity is the sum of the loss and of the measured intensity at 266 nm. As shown in Fig. 9, the transmittance decreases nonlinearly with increasing total intensity while it should have been constant if no TPA. Moreover, the intensity at 266 nm is in the order of GW.cm⁻² where TPA is generally assumed to occur. We can conclude that the energy loss mostly comes from linear absorption at 266 nm but also from TPA at 266 nm. If we assume no linear absorption and no TPA at 266 nm, the conversion efficiency would be 46%.

 

Fig. 9 Intensity-dependent transmittance at 266 nm.

Download Full Size | PDF

6. Conclusion

A YAB single crystal was grown using La₂O₃-B₂O₃ as a flux. A 2.94 mm thick type I YAB crystal was cut in ZX plane (θ = 66.9°, φ = 0°) for 266 nm generation. The transmittance spectrum reveals two strong absorption bands at 216 nm and 264 nm with an absorption coefficient of 2.7 cm⁻¹ at 266 nm. In spite of this absorption, 240 kW peak power at 266 nm with 445 ps pulse width corresponding to a mean conversion efficiency of 12.2% has been obtained. These results reveal that YAB is a promising NLO crystal for 266 nm generation.