1. Introduction

Nonlinear ε-GaSe crystal has many of attractive physical properties [1] and, in particular, has been applied for extreme wide range generation from 2.7 to 38.4 µm and further to 58.2–3540 µm range [2]. Unfortunately, its application has been held back by very low mechanical properties: near zero hardness in Mohs and easy cleaving [1]. On the other hand, it is well known that GaSe lattice well incorporates different doping elements, such as In [3] and S [4] with noticeable modification of mechanical and optical properties. Doped crystals grown in accordance with chemical composition GaSe:GaS → GaSe_1-xS_x are the only mid-IR crystal with end GaSe crystal that can meet the requirements for mid-IR nonlinear optical devices pumped by high peak power picosecond Nd:YAG, or femtosecond Ti:Sapphire and Cr:Forsterite lasers due to a shift of transparency range towards shorter wavelength with doping leading to a reduction of linear and nonlinear absorption. Doping with light Al also results in the shift of the transparency range towards shorter wavelength but leading to drastic degradation in optical quality even at low (by 0.1–2 mass%) level of Al content [4].

The layered GaSe and GaS crystals have identical arrangement of atoms in a layer with four different types of regular stacking of the layers, as centrosymmetric β and acentrosymmetric δ, ε and γ polytypes, respectively. p-type GaSe crystals grown by Bridgeman technique are often related to ε-polytype but can contain a mixture of all acentrosymmetric polytypes. The stacking of the layers in GaS crystal is invariably of the β-polytype, but on the other hand even Maker fringe experiments on pure GaS revealed the presence of second harmonic generation (SHG) [5]. GaSe and GaS form a continuous series of mixed GaSe_1-xS_x compound with mixing ratio x=0-1, but again in single crystals the presence of several polytypes is possible depending on x. The formation of ε-type was found for 0≤x≤0.01, the γ-type for 0.05≤x≤0.4, the β-type for x≥0.4, and a mixture of the ε- and γ-types for 0.03≤x≤0.05 [6]. The ε- to β-polytype transition for x=0.2-0.3 is mentioned in Ref. [6,7], but only presence of the β-polytype was observed for x=0.216 (5% sulfur) [8] and even for x=0.4-0.5 [9].

There is extensive information on the refractive index dispersion in GaSe [10–17] and GaS [7]. However, the experimental results on frequency conversion in mixed crystal GaSe_1-xS_x, x=0.2, 0.4 and even 0.8 were reported only in one study (in contradiction to data on the crystal structure [5–9]) [7]. Phase matching (PM) in GaSe_1-xS_x crystals in optical devices have not been studied in detail and any experimental data for PM angles are not available yet. In present study we report the PM properties for SHG in pure p-type GaSe and mixed GaSe_1-xS_x, x=0.04, 0.023, 0.090, 0.133, 0.175, 0.216, 0.256, 0.362, 0.369, 0.412, crystals at room temperature (RT).

2. Model estimations

Some of the available dispersion equations for GaSe_1-xS_x crystals are reported in Table 1.

Table 1. Sellmeier Constants for GaSe_1-xS_x Crystals

View Table | View all tables in this article

Eq. (1): n ²=A/λ⁴+B/λ²+C+Dλ²+Eλ⁴;

Eq. (2): n ²=A+B/(λ²+C)+Dλ²;

Eq. (3): n ²=A+B/λ²+C/λ⁴+D/λ⁶+Eλ²/(λ²−F).

Here Δλ is the validity range, Eq. is the equation type;

“o” and “e” indices denote ordinary and extraordinary waves respectively.

Birefringence (B) dispersion curves estimated with available dispersion data for pure GaSe crystals are depicted in Fig. 1(a) and estimated SHG PM diagrams in Fig. 1(b).

 

Fig. 1. Dispersion of B (a) and SHG PM diagrams (b) for pure GaSe. Points show available experimental data; PM diagrams estimated with data of [13] and [14] are coincident. Source of the data is identified in the figure insets; F.2 and F.3 are formula numbers.

Download Full Size | PDF

From Fig. 1(a) it is seen that estimated B dispersions are highly spread at wavelength λ<1.5-2 µm independent of the time of publication. Over the maximal transparency spectral range earlier published dispersion equations are resulting in higher B in spite of the fact that during this period lower quality crystals were grown and these crystals are occasionally possessing few lower B. Possible reason for that is formulating of disperse equations for n _e from nonlinear measurements on frequency conversion of laser radiation, which were often at the wavelength range λ<1.5-2 µm but not from direct measurements due difficulties in prism fabrication. Transparency spectrum, refractive index dispersions and resulted PM conditions are significantly varied within this spectral range due to variations in crystal stoichiometry and optical quality. Later published dispersion equations describe dispersion properties of the crystals with moderate α~0.1-0.05 cm^-1 over maximal transparency range with lower variations in optical properties at short-wavelength cut-off. It can be proposed that better crystal quality results in almost identical B at maximal transparency range [13–15] and Ref. 22 in [14]. Most of available experimental data on CO₂ laser SHG (Fig. 1(b)) are well related with data estimated on the basis of dispersion equations reported in Ref. 13, 14. Most recent published dispersion equations were developed with refractive indices directly measured for high quality GaSe crystal and show minimal B [16] that is close to results given in Ref. 13–15.

Estimated SHG PM diagrams for mixed GaSe_1-xS_x crystals are displayed in Fig. 2 in comparison with PM diagrams for moderate [16] and high [16] quality GaSe, and doped GaSe:Er(0.5%) crystals with α≈5 cm^-1 (Fig. 1 in [13]).

 

Fig. 2. SHG PM diagrams for pure GaSe after [13], GaSe_1-xS_x after [7] and imitation diagrams (dashed lines). Points are this study experimental data for crystals studied (labeled in the figure inset by number reported in Table 1/mixing ratio.

Download Full Size | PDF

In Fig. 2 it is seen that SHG PM diagrams estimated for centrosymmetric GaS (x=1) and GaSe_1-xS_x crystals with use the dispersion data of [7] are in unrealistic position to that of GaSe (x=0) crystal. Short-wavelength transparency end is at 0.48 µm for GaS (x=1) and at 0.62 µm for GaSe (x=0) [19] that is why shorter-wavelength position of GaS and GaSe_1-xS_x PM diagrams may be supposed. But in Fig. 2(a) shift of the GaSe_1-xS_x PM diagrams versus x is noticeable to longer-wavelength range, so as its irregularity at wavelength range λ<7.62 µm. In principle, it can be due to GaSe_1-xS_x composition and/or crystal structure change with x. PM diagrams for GaSe:Er(0.5 mass%), α=5 cm^-1 [17], has reasonable trend to that for lower quality GaSe crystals.

3. Experimental details

p-type pure GaSe and GaSe_1-xS_x single crystals used in this study were grown by the Bridgeman-Stockbarger method [15]. Sliced off samples (Table 2) have been used without any additional treatment and polishing and some of them had visual defects such as broken layers and local layer pieces on the lone of high optical quality faces.

Table 2. Parameters of GaSe and GaSe_1-xS_x crystals and SHG PM angles

View Table | View all tables in this article

Electron probe microanalysis was used to determine mixing ratio x=0.04, 0.023, 0.090, 0.133, 0.175, 0.216, 0.256, 0.362, 0.369, and 0.412. The GaSe_1-xS_x crystals with different x form red to yellow color with respect to the increase of the sulfur composition. Transparency spectra were recorded with spectrophotometer TU-1901, Puing Corp, China: Δλ=0.2-0.9 µm range, spectral resolution 0.05 nm and ATAVAR 360 FT-IR spectrophotometer, ThermoNicolet, USA: Δλ=2.5-25 µm, Δν=4 cm^-1. These transparency spectra are depicted in Fig. 3(a).

 

Fig. 3. Short-wavelength transparency spectra (a) and absorption coefficient (b) of GaSe, GaSe_1-xS_x and GaSe:In crystals.

Download Full Size | PDF

Calculated spectra of absorption coefficients are shown in Fig. 3(b). Point measurements with CO2 laser exclude an influence of surface defects on the absorption coefficient estimations and show that for all crystals α≤0.1-0.2 cm^-1 that couldn’t be determined with higher accuracy due to small thickness of the crystals. The only pure GaSe crystal #18 has α≈0.25 cm^-1.

Traditional SHG optical set-up is used. Leading pulse of line-tuneable CO₂ laser with TEM₀₀ mode selection, 600 Hz pulse-repetition frequency and 500 W peak power is of 120 ns FWHM followed by 1 μs tail. ZnSe 50 mm focal length lens is applied for focusing of Ø3.5 mm pump beam into the room temperature crystal. Step-motor-drive rotational stage RCA100, Zolix Instruments Co., Ltd, with positioning accuracy 4.5″ is used for determination of the PM angles. UV-FIR monochromator SBP300, Zolix Instruments Co., Ltd.: 66 gr/mm grating and RT pyroelectric detector MG-30, Russia: Δλ=2-20 µm, NEP=1.5·10^-9 W/cm·Hz^1/2 are applied to measure wavelengths and record SHG pulses, respectively. Digital storage oscilloscope TDS3052, Tektronix Inc., Δf=500 MHz, is used to display pulses time shape-form. The residual pump radiation was blocked by two 3 mm LiF plate located close to the nonlinear crystal and detector. Homemade Q-switched 250 ns Er³⁺:YSGG operating at λ=2.79 µm with Ø3 mm TEM₀₀ beam also was used as a pump source. Its high pulse output energy, up to 24.5 mJ, let us to carry out SHG experiments without use of focusing lenses.

4. Result and discussions

In Fig. 3(a) it is seen that transparency spectra for GaSe_1-xS_x, x=0-0.405, crystals are linearly shifting toward shorter wavelength with x, as it was supposed accounting shorter-wavelength transparency cut-off of GaS [19]. PM angles are gradually decreasing versus x for Er³⁺:YSGG SHG and gradually increasing for CO₂ laser SHG (Fig. 2). Step changes in SHG efficiencies were also not observed. Linear shift of the short-wavelength transparency end, gradual shift of Er³⁺:YSGG and CO₂ laser SHG PM angles (Table 2), and SHG efficiency with x confirms that p-type GaSe_1-xS_x studied is a continuous series of mixed crystals without change of polytype. This result is in agreement with data of Ref. [19]. From Fig. 2 it is also seen that the spectral derivative of PM diagram at Er³⁺:YSGG laser wavelength is a few times to spectral derivative at CO2 laser wavelengths, but PM angle changes with x for Er³⁺:YSGG SHG is about a half to changes for CO₂ laser SHG. It can be explained by small (~0.07 µm) shift of short-wavelength transparency end in comparison with ~4.0 µm shift of long-wavelength transparency end of GaSe_1-xS_x crystals towards shorter wavelength and by possible shift of PM diagrams to upper positions [7]. For easy understanding of this point imitation diagrams are depicted in Fig. 2 by dashed lines. No evident domain structure of the crystals was found in the measurements. The error margin of ±15-20″ of the measurements was due to local deformations of the crystal surfaces and long time instabilities of the facility parts.

Stoichiometric variations in GaSe crystals of different origin are another reason for the scatter in phase matching experimental data. PM angles for Er³⁺:YSGG SHG in moderate optical quality pure GaSe crystals #1 and #15 (α≤0.1-0.2 cm^-1) with about 1 mass% excess of Se content are of 49.25° and 49.29°, they are of 39.48° and 39.38° for CO₂ laser 9.58 µm emission line SHG. These PM angles are in good coincidence with PM angles estimated with dispersion equations of [13] designed for close quality GaSe (α=0.05 cm^-1) to the crystals studied. SHG PM angles for lower quality, α≈0.25 cm^-1, pure GaSe #18 with about 1.5 mass% excess in Ga content are of 50.03° and 40.15°, respectively for these lasers, showing trend to PM diagram designed for lower quality (α≈0.25 cm^-1) doped GaSe:Er(0.5%) crystal [17].

5. Conclusion

High optical quality, α≤0.1-0.2 cm^-1, mixed single crystals of p-type GaSe_1-xS_x, x=0.04, 0.023, 0.090, 0.133, 0.175, 0.216, 0.256, 0.362, 0.369, and 0.412, are grown and studied. Through transparency spectra, PM conditions and efficiency for Er³⁺:YSGG and CO₂ laser SHG at room temperature it was determined that polytype structure of the p-type GaSe_1-xS_x is not changing with x in spite of predominantly different polytype structure of end GaSe (x=0) and GaS (x=1) crystals at RT. These crystals are useful for application in nonlinear devices. SHG PM diagrams for GaSe_1-xS_x crystals are shifting with x to shorter-wavelength range in full coincidence with shorter wavelength transparency cut-off of GaS and possibly to upper position, as it goes from available data. SHG efficiency in GaSe_0.91S_0.09 is 2.4 times of pure GaSe. It was shown that up to 1° difference in SHG PM angles can be caused by the difference in the GaSe astoichiometry.