1. Introduction

There is a severe need for nonlinear crystals able to cover the near-IR and mid-IR spectral regions and applicable in practical frequency conversion systems possessing generation of coherent light beams with wide wavelength tuning. Such key factors as stability of optical properties upon exposure of the crystal to high power laser beam, high optical homogeneity and sufficient mechanical properties are limiting for material selection. Layered p-type ε-GaSe is an excellent crystal for frequency conversion over a wide spectral region from near-IR to mid-IR and even to terahertz region [1], but an application of GaSe in commercial systems is hampered by very poor mechanical properties. GaSe is characterized by nearly zero hardness by Mohs scale, high cleavage and noticeable irreproducibility of nonlinear properties from sample to sample [2]. On the other hand, it is well known that comparatively flexible GaSe crystal lattice well incorporates different doping elements and even compounds, such as In [3–8], Er [9], S [10,11] and AgGaSe₂ [6] with solid solution formation and significant modification of physical properties. The GaSe crystals doped with indium (In) show a strengthening of the mechanical properties and can be cut and polished in crystallographic direction selected according to phase-matching conditions for interacting light waves. Moreover, an increase of optical nonlinear coefficient of GaSe:In in comparison to that of pure GaSe has been obtained [6]. Information on influence of In doping on optical transition range of GaSe-based solid solution is inconsistent. No effect of In doping on GaSe transmission curve has been reported in Ref. 6. However, a shift of long-wavelength transmission end towards longer wavelengths has been detected in Ref. 10. It was experimentally shown that phonon absorption spectrum in GaSe:In is generally shifted towards long-wavelengths in comparison with that in pure GaSe [12]. Earlier, non-monotonic dependence of absorption coefficient on In content in GaSe:In was found with minimum absorption at In content as 0.5 mass% [7]. Contrary to that, monotonic increasing of absorption coefficient with In content increase in solid solution Ga_1-xIn_xSe up to solubility limits ≤14.6 mass% and ≥47.2 mass% (x≤0.2 and x≥0.75, respectively) was also reported [12]. So, even general tendency in influence of In doping on transmission range and optical quality of GaSe is not clear.

In high power nonlinear devices a crystal may be heated due to residual absorption of the pump laser radiation. This effect changes the phase matching (PM) conditions and reduces the conversion efficiency. On the other hand, temperature tuning can allow the PM which is unreachable by angular tuning [13] and fine adjustment to non-critical PM is possible [14]. Again, very often, the exact requirement of the doping concentration can not be realized in the growth process [15]. Thus an appreciable deviation in PM occurs that reduces the conversion efficiency of the frequency conversion process. Then, negative effect of composition variation from element ratio calculated for most benefit PM conditions can be compensated by crystal temperature tuning. Thus crystal temperature stability and tunability are important parameters in view of desired operation of nonlinear crystal.

Although there is extensive literature on the dispersion properties of pure GaSe [16–20] and for its temperature dispersions [14,21,22], until now the PM conditions in indium doped crystals have not been studied in details. In particular, no effect has been earlier found for indium doping on dispersion properties of GaSe [3,6]. Contrary to that, Das et al. [10] reported ~0.4° difference in internal PM angle measured for second harmonic generation (SHG) of 10.64 µm CO₂ laser emission line in reference to theoretical value obtained using the Sellmeier equation of pure GaSe crystal [19] and 0.01° [8] in reference to theoretical value obtained using historically first data for Sellmeier equations [17]. No experimental data have been reported on temperature tuning of PM in In doped GaSe.

In present study optical transparency and PM conditions are studied for SHG of Er³⁺:YSGG and CO₂ laser illumination in indium doped GaSe:In(0.1, 1.23, 2.32 mass%) crystals throughout temperature range from -165 to 230 °C in comparison with pure and sulfur doped GaSe:S(0.09, 0.5, 2.2, 3 mass%) crystals.

2. GaSe growth technology

Pure GaSe and doped GaSe:In, GaSe:S crystals were grown by Bridgman technique in SPhTI of TSU, Tomsk and FGUP GOKB Ametist, Krasnodar, Russia. Fabrication of GaSe single crystals involves two principal steps: (1) GaSe chemical synthesis starting from high purity (six-9,s) elemental reagents, and (2) single crystal growth starting from synthesized GaSe polycrystal. Synthesis of 100 to 150 g of GaSe compound was carried out in quartz ampoules 15–25 mm in diameter evacuated to 10^-5 mm Hg. The gallium and selenium were placed in different boats, which were inserted into the opposite sides of the ampoule. The selenium was then transported into the boat with Ga in the vapor phase and GaSe polycrystalline ingot synthesized by interaction of the vapor phase with Ga melt surface was contained entirely in the Ga-boat after cooling. Doping was accomplished by adding an appropriate amount of indium to the boat with gallium and sulfur admixture to the boat with selenium charge. The synthesis has been performed by three sequential stages. At first step, the effective sublimation of chalcogen atoms to a vapor phase and transport to Ga/In occur. Second step ensures homogenization of GaSe melt due to diffusion in the melt at constant temperature. At final step slow cooling of the melt has been produced for 36 h with formation of homogeneous crystalline ingots with dimensions 8-10 cm. Best results were achieved when the temperature of the melt was T=1015±5 °C, temperature gradient at the crystallization front was 10 deg/cm and crystal pulling rate was in the range 0.5-1 mm/h.

3. Crystal characterization

Chemical composition of doped GaSe single crystals has been measured with electron probe micro-analysis (EPMA) with using LEO 1430 device. Phase purity has been tested with X-ray diffraction (XRD) analysis and ε-GaSe polytype was found. As-grown highly In doped GaSe crystals were significantly deficient in indium content in reference to initial charge composition, some of them had spherical precipitates and/or local surface deformations. Highly doped GaSe:S crystals were ~1-2% deficient in chalcogen content in reference to initial charge composition and weakly doped GaSe:S crystals show chalcogen deficiency as ~10–15%. Single crystal GaSe ingots grown by use of non-stoichiometric element proportions in the charge were often almost stoichiometric, free of cracks and containing indium and sulfur in the amounts of <3 mass%. For present study, optical quality samples were cleaved up to thickness l_c=0.5–1.5 mm or 3–4 mm.

Microhardness measurements were performed with hardnessmeter PMT-3, Russia. The hardness of sulfur-doped crystals was dependent proportionally on doping level. The GaSe:S crystals show weaken tendency to cleavage and were suitable for handling during optical characterization. In particular, for the crystal #4 (2 mass% S content) the hardness is 15 kg/mm² that is twice as large as that of pure GaSe hardness of 7-8 kg/mm². The hardness of In-doped crystals with doping level <3 mass% varies noticeably from sample to sample and often well correlates with the crystal optical quality as it is also seen in Fig. 6 and Fig. 5(b) of Ref. [7]. In our experiment the hardness of In-doped crystals is only slightly higher than that of pure GaSe. In particular for the GaSe:In crystal #20 with 2.32 mass% of indium content the hardness is measured as 9.5 kg/mm². For lower quality crystals, with doping level over 1 mass%, the hardness is similar to or little higher than that of GaSe:S solid solutions with the close doping level. For example, as high hardness as 14 kg/mm² has been determined for the crystal #17 with 1.23 mass% of In. We presume that higher hardness is induced by significant In intercalation into interlayer space in GaSe:In crystals [23]. To retain the planarity of cleaved samples, we mounted them on metal plates with appropriate ellipsoidal apertures with dimensions 5×15 mm² for optical access. Low accuracy in hardness measurements reported in [7] can be partially explained by significant difference between crystal and initial charge compositions that was not accounted.

Parameters of selected crystals are presented in Table 1.

Table 1. Parameters of pure and doped GaSe crystals and external PM angles for Er³⁺:YSGG and CO₂ laser SHG.

View Table

Here x is a mixing ratio in solid solutions Ga_1-xIn_xSe and GaSe_1-xS_x, l_c is a crystal length, 2.79 µm and 9.58 µm are the pump wavelengths of Er³⁺:YSGG and CO₂ laser, respectively, θ_ext is an external PM angle.

Cleaved samples of different optical quality have been used in this study without any additional treatment or polishing. Several samples have such visible surface defects as broken layers and local layer pieces on the lone of high optical quality faces. The optical characterization includes bulk transparency measurement and determination of PM conditions for Er³⁺:YSGG and CO₂ laser SHG. Transmission spectra shown in Fig. 1 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. Direct point measurements at CO₂ laser wavelengths show that all samples are characterized by absorption coefficient α≤0.1–0.2 cm^-1, pure GaSe crystal #1 and In doped crystal #17 by α≈0.25 cm^-1.

 

Fig. 1. Transmission spectra of pure and doped GaSe crystals. Curve identification is given in the insert.

Download Full Size | PDF

4. Experimental setup

Traditional optical set-up shown in Fig. 2 is applied for I type SHG.

 

Fig. 2. Scheme of the experimental setup for CO₂ and Er³⁺:YSGG laser SHG. BS1 and BS2 are removable beam splitters, M1 and M2 is two-mirror optical system for a laser beam alignment, ID1, ID2 and ID3 are iris diaphragms, L1, L2 and L3 are lenses, NLC is a non-linear crystal, F is a block filter and D1 and D2 are detectors, PC is personal computer, OS is oscilloscope.

Download Full Size | PDF

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 lens with focal distance f=50 mm is applied for focusing of ∅3.5 mm pump beam onto the crystal attached to the temperature-controlled holder installed into the evacuated Duar vessel and placed at 1 m distance from the pump laser. The crystal temperature can be tuned within the range from -165 to 230 °C with accuracy ±2.5 °C and temperature gradient about 40 °C/hour. Second ZnSe lens with f=50 mm is applied for collimating of SHG beam after nonlinear crystal. Step-motor-drive rotational stage RCA100, Zolix Instruments Co., Ltd, with positioning accuracy 4.5″ is used for determination of the PM angles. One direction (clockwise) rotation is used to exclude an influence of backlash on PM angle measurements. For careful determination of “0”-angle position of a crystal, reference He-Ne laser beam, reflected from the crystal surface, was aligned at the 100 µm slit located at 2 m distance from the crystal. Third ZnSe lens with f=50 mm was installed before the slit. 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 is applied to measure wavelength values and record SHG pulses, respectively. Pulse time shape-form was recorded with digital storage oscilloscope TDS3052, Tektronix Inc., Δf=500 MHz. The residual pump radiation was blocked by two 3 mm LiF plate located close to the nonlinear crystal and detector. Home-made Q-switched 250 ns Er³⁺:YSGG pulse laser operating at λ=2.79 µm with ∅3 mm TEM₀₀ beam was used as an alternative pump source. High pulse output energy of Er³⁺:YSGG laser, up to 24.5 mJ, let us to carry out the SHG experiments without using of focusing and collimation lenses. A set of organic plates is used as a block filter for 2.79 µm radiation.

5. Results and discussions

No variation is found in transmission curves shown in Fig. 1 for GaSe:In crystals with different doping levels that confirms earlier results [6]. Transmission range of GaSe:S crystals is significantly shifted toward shorter wavelengths with increasing of S content. The reason for immeasurable variation of transmission spectrum for the GaSe:In crystals becomes clear if to calculate the x value in chemical formula of the solid solution. For GaSe:In or Ga_1-xIn_xSe and GaSe:S or GaSe_1-xS_x crystals it can be determined from relations:

x·A(In)/A(Ga_1-xIn_xSe)=x·A(In)/[(1-x)·A(Ga)+x·A(In)+A(Se)]=P(In, mass%)/100%,

x·A(S)/A(GaSe_1-xS_x)=x·A(S)/[A(Ga)+x·A(S)+(1-x)·A(Se)]=P(S, mass%)/100%,

where x is a mixing ratio or mole content of a dopant, A(In)=114.82, A(Ga)=69.723, A(Se)=78.96, A(S)=32.066, A(Ga_1-xIn_xSe) and A(GaSe_1-xS_x) are atomic weights of and In, Ga, Se, S, Ga_1-xIn_xSe and GaSe_1-xS_x, relatively; P(In, S, Ga_1-xIn_xSe, GaSe_1-xS_x mass%) are mass contents of In, S, Ga_1-xIn_xSe and GaSe_1-xS_x, relatively; P(In, mass%) and P(S, mass%) are In and S mass content in mass%. Due to drastic difference of atomic mass of In and S, the value 2 mass% of S or In is equivalent, respectively, to mole content of x=0.09 for sulfur and x=0.03 for indium. This means that doping level written in mass% relates to more than three times lower number of Ga atoms substituted by In atoms in GaSe:In (without intercalation) solution than that for Se atoms substituted by S atoms in GaSe:S solution. Besides this, significant intercalation of In into interlayer space was found in GaSe:In crystals [23]. Thus, the combination of these factors promises for much smaller shift of transparency curve for GaSe:In crystals in comparison with that in GaSe:S solid solutions.

No optimal levels of In or S doping are found to have minimum optical losses. GaSe:In crystals show rapid degradation of optical quality with In doping increase over 1–2 mass% that well correlate with the crystal hardness variation and possible intercalation. This can be an explanation why no studies of optical properties were reported for GaSe:In crystals with In content over 7184 ppm while at least 10000 ppm In-doped crystals were grown [3–8]. The cleaved crystals characterized by α≤0.1–0.2 cm^-1 are almost uniform in optical quality and no domain structure is found. GaSe:S crystals, however, show no noticeable changes in optical quality that well correlates with supposed small S intercalation.

In Fig. 3 the dependencies of I type PM conditions for Er³⁺:YSGG and CO₂ laser SHG on sample temperature are shown.

 

Fig. 3. External PM angles for SHG under pumping by (a) Er³⁺:YSGG and (b) 9.58 µm CO₂ laser emission line versus temperature. Experimental data of this study are given by circles. Triangles are the PM angles estimated with available literature data for refractive indices. Crystal parameters are shown in the inserts.

Download Full Size | PDF

In Fig. 3 it is seen that external PM angles for Er³⁺:YSGG and CO₂ laser SHG in the In and S doped crystals have opposite trends with the doping from PM diagram for pure GaSe. In indium doped crystals PM angles for Er³⁺:YSGG laser SHG are increasing with doping but decreasing for CO₂ laser SHG with about 3-fold lower gradient. Only repeated measurements and averaging of the results allow detection of as low as about 0.06° difference in PM angles for good quality doped and pure GaSe crystals. These features of the PM in GaSe:In crystals reveal a shift of the PM diagrams to longer wavelengths in difference to the shift to shorter wavelengths for GaSe:S crystals. Very small changes in PM angles with In doping explain why they were not found in SHG experiments [3,6]. The PM angles can be estimated with available dispersion data [16–20]. At T=27 °C our results obtained for pure GaSe crystals are in good relation with PM angle θ=39.43° estimated for CO₂ laser SHG with dispersion formula given in [19] but are 5.85° higher than the PM angle estimated with dispersion data of [17]. Thus, the 0.05° difference between experimental data for CO₂ laser SHG (Fig. 4 [8]) and estimated with Sellmeier equations of [17] reported in [8] and also 0.4° difference between experimental data and estimated with Sellmeier equations of [10] seems to be an impact of low accuracy rotational stages used for crystal positioning.

In both GaSe:In and GaSe:S crystals θ_ext for SHG at λ=2.79 µm is increasing with heating. Contrary to that, small decreasing in θ_ext for SHG at λ=9.58 µm is observed on T increasing in pure GaSe:In and GaSe:S crystals. From Fig. 3(a) it is seen that at T>-23 °C the variation of PM angle with temperature can be approximated by linear function with the slope dθ/dT=22″/1 °C for pure and In doped crystals. At T ~-165 °C the slope is half as that at positive temperatures. This seems to be a consequence of birefringence decrease as it was observed for ZnGeP₂ [24]. Measured PM temperature bandwidth is of 22 °C·cm FWHM. The PM angles for CO₂ laser SHG are linearly decreasing with temperature with the slope dθ/dT=-4.9″/1 °C. PM temperature bandwidth at 9.58 µm pump is of 219 °C·cm FWHM, which is close to earlier reported value of 172 °C·cm FWHM for pure GaSe crystal at 10.59 µm pump [14].

6. Conclusion

We have reported the transmission range and phase matching conditions for second harmonic generation of Er³⁺:YSGG and CO₂ laser in doped GaSe:In(0.1, 1.23, 2.32 mass%) in comparison with pure and sulfur doped GaSe:S(0.09, 0.5, 2.2, 3 mass%) crystals. Transmission range of GaSe crystals is independent on indium doping up to 2.32 mass%. Optical quality of studied GaSe crystals is rapidly degraded on In doping and GaSe:In crystals become not useful in nonlinear devices at In content of >3 mass%. If an improvement of mechanical and nonlinear properties of GaSe:In crystals with doping level increase is followed by rapid decreasing in optical quality and x=0.2 as a principle indium solubility limit [12], then it is possible to estimate optimal doping level that gives maximum efficiency of frequency conversion processes. Optimal doping level for the studied crystals is estimated to be between 0.5 and 1 mass% for GaSe:In crystals. Small changes ~0.18° in PM angle for 2.79 µm Er³⁺:YSGG laser SHG and ~0.06° for 9.58 µm CO₂ laser emission line SHG have been observed. Variations of In doping over the crystal volume do not disturb phase matching conditions. Dispersion properties of the crystals are studied as a function of temperature over the range from -165 to 230 °C. GaSe:In crystals show low sensitivity to phase matching angle tuning with temperature with slopes dθ/dT=22″/1 °C at Er³⁺:YSGG laser pump and dθ/dT=-4.9″/1 °C at CO₂ laser pump. Linear changing of PM with temperature is an indicator of absence of crystals structure transformation within temperature range from -165 to 230 °C. Thus, application of GaSe:In solid solutions in high average power nonlinear optical systems instead of pure GaSe crystals seems to be reasonable. Improved growth technology of GaSe:In crystals with high, up to x=0.2, level of In doping without significant In intercalation into interlayer spaces should be designed. Due to noticeable shift of transparency spectrum to shorter-wavelength range sulfur doped GaSe:S (≤3 mass%) crystals are preferable to pure and indium doped GaSe crystals for frequency conversion of short-wavelength near-IR lasers into mid-IR range. Due to both short-wavelength transparency shift and higher nonlinearity GaSe:S crystals are also preferable for the same application to another known crystals: AgGaS₂, LiInS₂, LiGaS₂, AgGaGeS₄, LiInSe₂, etc. that are transparent from visible to near-IR and further to mid-IR region.