Investigation of Structural and Optical Homogeneity of LiNbO₃:ZnO Crystals of Different Genesis

Abstract

Nominally pure congruent LiNbO₃ crystal and LiNbO₃:ZnO ([ZnO] ~ 5.4–6.4 mol % in the melt) crystals of different genesis are studied using photoinduced light scattering, laser conoscopy, and X-ray diffraction (method of moments). The photorefractive properties and optical and structural uniformity are analyzed. It is found that LiNbO₃:ZnO crystals obtained by direct solid phase doping have the best optical and structural uniformity, the lowest photorefractive sensitivity, and a high compositional uniformity along the polar axis. Crystals obtained by homogeneous doping are compositionally uniform, although their optical and structural uniformity is worse than those of LiNbO₃:ZnO crystals obtained by direct solid phase doping. Anomalies of conoscopic patterns are caused by the presence of charged structural defects in LiNbO₃:ZnO crystals and the distortion of the optical indicatrix induced by mechanical stresses and compositional inhomogeneity of the crystals. High temperature annealing in short-circuited state of LiNbO₃:ZnO crystals with high dopant concentration leads to healing of charged defects and improvement of optical characteristics in general. The distribution coefficient for LiNbO₃:ZnO crystals obtained by the homogeneous doping method is significantly higher than that for the crystals obtained by the direct solid phase doping method.

INTRODUCTION

Lithium niobate crystal (LiNbO₃) is a long-term center of attraction for specialists in integral and nonlinear optics, acoustoelectronics, quantum electronics, and solid state physics [1–5]. Interest in the study of highly doped LiNbO₃:ZnO crystals (~4.0–9.0 mol % ZnO in the melt) is due to their high resistance to optical damage [5–7]. In addition, highly doped LiNbO₃ crystals, such as LiNbO₃:ZnO, form polar clusters owing to localization of impurity and intrinsic defects along the polar axis [5]. Such clusters according to mathematical modeling can reach the size of ~5–10 nm [8]. In contrast to crystallography, methods of fractal geometry allow the study of disordered structures under the condition of repeatability of elements. A common characteristic of fractal structures is that they are formed in the absence of thermodynamic equilibrium, as is the case when growing crystals. Periodic nanoscale structures of fractal type with a step from 5–10 to 100 nm were detected by atomic force microscopy in doped LiNbO₃ crystals. Moreover, the periodic splitting occurs both in the direction parallel and in the direction perpendicular to the polar axis of the crystal. Thus, periodic structural formations with characteristic sizes of 5–10 nm appear in the crystal, which agrees with the results of mathematical modeling [8]. In addition, it was shown that, during high-temperature annealing of the original polydomain and monodomain LiNbO₃:ZnO crystalls in the short-circuited state, there occurs a significant increase in unipolarity [9] accompanied by the collapse of polar clusters that stabilize the domain boundaries, the disappearance of domain boundaries, and overall significant restructuring of the crystal microstructure.

The study [10] reports on the influence of the doping method on the structure and properties of LiNbO₃:Me crystals. At the same time, LiNbO₃:Me crystals (Me: Mg, Zn, B, etc.) with similar impurity concentrations and grown using different doping methods have noticeable differences in structure and optical and physicochemical characteristics [10].

The purpose of this work is a comparative study of the photorefractive properties and optical and structural homogeneity of LiNbO₃:ZnO crystals ([ZnO] ~ 5.4–6.4 mol % in the melt) grown by the Czochralsky method from a charge synthesized using precursors Nb₂O₅:ZnO obtained by homogeneous doping with zinc at the stage of Nb₂O₅ extraction and a LiNbO₃:ZnO charge obtained by the direct solid-phase doping method.

EXPERIMENTAL

Granulated lithium niobate charge for growing LiNbO₃ and LiNbO₃:ZnO crystals was obtained from mixtures of Nb₂O₅–Li₂CO₃, Nb₂O₅–ZnO–Li₂CO₃, and Nb₂O₅:ZnO–Li₂CO₃ by granulation synthesis. LiNbO₃ and LiNbO₃:ZnO crystals with a diameter of 40 mm and a length of the cylindrical part of ~35–40 mm were grown in the direction (001) by the Czochralsky method in an air atmosphere from Pt crucibles with a diameter of 75 mm. The method of direct solid-phase doping consists in the synthesis of a charge of lithium niobate from a mixture of Nb₂O₅–ZnO–Li₂CO₃. The method of homogeneous alloying is based on the fact that the alloying additive is introduced into niobium pentoxide at the stage of its isolation from high-purity niobium-containing solutions. The resulting doped niobium pentoxide (Nb₂O₅:ZnO) is used as a precursor in the synthesis of the charge from a mixture of Nb₂O₅:ZnO–Li₂CO₃ [10]. Crystals 1 and 2 were grown by direct solid-phase doping, and crystal 3 was grown by homogeneous doping (Table 1). LiNbO₃:ZnO crystals were monodomainized by high-temperature electrodiffusion annealing (HTEDA) by applying a constant electric field when the crystal was cooled in the temperature range of ~1513–1173 K. The degree of monodomainity of the crystals was controlled by analyzing the frequency dependence of the electric impedance and by determining the value of the static piezoelectric modulus (d_333st) of the crystal blank.

Table 1.   Concentration of ZnO in the melt (C_m) and in the upper part of the LiNbO₃:ZnO crystal (C_u), parameter ΔC = C_u – C_l, and distribution coefficient K_d

The concentration of zinc in the crystal was determined in the plates cut from the upper (cone-shaped) and lower (flat-end) parts of the blank. The zinc concentration in these plates, C_u and C_l, respectively, was determined by the method of atomic emission spectrometry using a Shimadzu ICPS-9000 spectrometer.

Samples for the study of mosaic structure, photoinduced light scattering (PILS), and conoscopic patterns were cut from LiNbO₃:ZnO crystals in the form of parallelepipeds (size ~ 8 × 7 × 6 mm³), the edges of which coincided in the direction of the crystal-physical axes X, Y, Z (Z is the polar axis of the crystal). In order to study PILS and conoscopic patterns, the faces of the parallelepipeds were carefully polished.

PILS was excited by a Nd:YAG (MLL-100) laser (wavelength λ₀ = 532 nm) with power P = 160 mW. In the PILS experiments, the laser beam was directed along the Y axis, and the intensity vector E of the electric field of the laser radiation was parallel to the polar axis Z of the crystal. The experimental setup and method for determining the PILS indicatrix are described in detail in [11].

Optical homogeneity and structural distortions in LiNbO₃:ZnO crystals were controlled by laser conoscopy, which allows observing large-scale and high-resolution conoscopic patterns [12, 13]. This is especially important for doped single crystals of LiNbO₃ because of the possibility of uneven distribution of the dopant into the structure. The conoscopic patterns were recorded when excited by Nd:YAG (MLL-100) laser radiation (λ₀ = 532 nm) with a power of P = 1 and 90 mW.

The characteristics of the mosaic structure of crystals of LiNbO₃:ZnO were calculated using the method of moments, which is quite universal and allows calculations for the case where it is possible to register only one order of reflection from the family of planes with indices (hkl). Diffraction patterns were recorded on an equatorial diffractometer in the θ–2θ geometry in CuK_α radiation both with and without rotation of the sample. Reflections were registered from three mutually perpendicular crystal planes (XY, YZ, ZX in the orthogonal coordinate system): (006), (0012), (110), (220), (330), (300), respectively. The angles of disorientation of the blocks were estimated from the analysis of the profile shape of the X-ray patterns obtained without rotation of the samples. The average size of blocks and the magnitude of micro-distortion were calculated from X-ray images obtained with the rotation of the sample using the method of moments [14–16] based on the analysis of the profile, not the width of the diffraction lines. The profile of the diffraction line h(x) is a convolution the “physical” profile f(x) and the instrumental profile g(x), which are associated with the deviations of the sample microstructure from the ideal and the geometry of the scheme, respectively. In turn, the function of the “physical” profile is a convolution of the functions f_D(x) and f_m(x), whose shape and width are determined by the final dimensions of the mosaic blocks and microdeformations, respectively. The moment of a function is an integral of the form

$${{M}_{n}} = \int\limits_{{{\sigma }_{1}}}^{{{\sigma }_{2}}} {{{x}^{n}}f(x)dx} ,$$

(1)

where n is the order of the moment and f(x) is the current value of the function.

(1) The zero-order moment, n = 0, is the area under the curve, i.e., the integral intensity of the reflection on the radiogram.

(2) The first-order moment, n = 1, gives the position of the center of gravity of the reflection on the radiogram, i.e., the 2θ gravity center (gc).

(3) The moments of the second and fourth orders, n = 2, n = 4, are used in radiography to calculate the magnitude of micro-distortion and block sizes.

An important property of moments is their additivity: since each of the functions f(x) and h(x) is a convolution of two functions, their central moments are equal to the sum of the moments of the convolved functions. The term central moment means that the coordinate (±x) is counted from the coordinate x_gc of the line gravity center to the values of ±σ, i.e., from σ₁ to σ₂. Therefore, for the above functions, the central moment of the line profile of the sample is simply the sum of the contributions of the moments of the geometric profile and profiles due to the block structure and microdeformations, respectively, $M_{n}^{h}$ = $M_{n}^{g} + M_{n}^{D} + M_{n}^{M}.$ In the case where reflections of several orders can be registered from the given system of planes, only the second order moment is calculated for each of them. If there is only one reflection from the given system of planes, then the fourth order moments are used. The calculation method is described in detail in [14–16].

RESULTS AND DISCUSSION

The time dependences of the PILS patterns of crystals 1–3 before and after annealing in the short-circuited state are shown in Fig. 1. The patterns of PILS of LiNbO₃:ZnO crystals both before and after annealing hardly change in time or change very slightly (Fig. 1). Even with the exciting radiation power of 160 mW, there is no photorefractive response, the PILS indicatrix is not revealed, and only circular scattering on static structural defects is observed. In this case, the scattering pattern retains a shape close to a circle throughout the experiment (Fig. 1). However, the PILS patterns of the studied LiNbO₃:ZnO crystals differ from each other. After annealing, crystals 2 and 3 are characterized by a significantly larger size of the central layer corresponding to the cross section of the laser beam (Fig. 1b). The scattering patterns of these crystals both before and after annealing indicate a greater scattering capacity (i.e., optical and structural heterogeneity) compared to crystal 1 (Fig. 1). This is especially pronounced for crystal 3, which is a homogeneously doped LiNbO₃:Zn crystal (Fig. 1). The scattering pattern for this crystal is significantly less contrasting, there is a significant blurring of the image, and the asymmetry of the central PILS layer in the direction of the Z axis after annealing in the short-circuited state is replaced by an asymmetry in the direction of the X axis (Fig. 1). Analysis of the scattering patterns of the studied LiNbO₃:ZnO crystals indicates a significantly higher defectiveness of crystal 3 in comparison with crystals 1 and 2 (Fig. 1, Table 1).

Fig. 1.

PILS patterns of LiNbO₃:ZnO crystals 1–3 (a) before and (b) after annealing in the short-circuited state. λ = 532 nm, P = 160 mW.

When a uniaxial crystal placed in the optical system between the polarizer and the analyzer is illuminated by a divergent beam of linearly polarized radiation, the observed conoscopic pattern has the form of concentric isochromatic rings with the center at the output of the optical axis with characteristic intensity distribution in the form of a black Maltese cross superimposed on it. The branches of the Maltese cross consist of two isogyres which intersect at the center of the field of view and diverge to the edges of the cross. Anomalous optical biaxiality (deformation of the optical indicatrix of the crystal) is manifested in the form of distortion and rupture of the Maltese cross into two parts with lucidity at the center of the view field of the conoscopic pattern. The rupture and shift of the parts of the cross is possible in any azimuthal direction and is uniquely related to the direction of deformation of the optical indicatrix. Figure 2 presents conoscopic patterns of crystals 1–3 before and after annealing in the short-circuited state obtained at a laser radiation power of P ~ 1 and 90 mW. At low radiation power (~1 mW), the distortions of conoscopic patterns of doped crystals are associated only with their structural inhomogeneity, for example, due to uneven distribution of the doping component in the process of crystal growth. Additional distortions of conoscopic patterns appearing at high laser radiation power (90 mW) give information about the distortion of the crystal structure caused by the action of the laser beam (the photorefraction effect). The conoscopic pattern of crystal 1 obtained at low radiation power (1 mW) before the crystal annealing exhibits clearly anomalous optical biaxiality: a significant deformation at the center of the black cross in the form of vertical displacement of the cross fragments from its center, which corresponds to the direction of deformation of the optical indicatrix of the crystal (Fig. 2a (1)). Analysis of this conoscopic pattern also shows that the angles between the branches of the cross are different from 90° and there is some lucidity in its central part. The isochromatic curves retain their integrity but are elongated in the direction of displacement of the cross fragments and take the form of ellipses (Fig. 2a (1)). There is some blurring of the image, which indicates the optical heterogeneity of the sample under study. In addition, some anomalies are visible in the area of all branches of the cross. Such distortions of the conoscopic pattern are probably related to the structural nonuniformity of crystal 1. This inhomogeneity, apparently, is not related to the inhomogeneity of the impurity distribution, since the value of ΔC is insignificant for this crystal (C_u–C_l ~ 0.1 mol % ZnO) (Table 1). At the same time, the distortion of the conoscopic pattern can be caused by charged defects and mechanical stresses. When the laser radiation power is increased to 90 mW, an almost common conoscopic pattern of a uniaxial crystal is seen (Fig. 2a (1)). There is a symmetry at which the contrast of the Maltese cross remains intact at the center of the field of view, and the isochromatic curves are concentric circles centered at the exit point of the optical axis. There is a slight decrease in image contrast in the area of the upper left branch of the cross. But in general, the conoscopic pattern indicates the optical uniformity of the sample and its good optical quality. Obviously, when the laser radiation power increases from 1 to 90 mW, there occurs laser annealing (“healing”) of the defects in the crystal. After annealing of crystal 1, the conoscopic pattern obtained at laser radiation power of 1 mW is sharper and generally has the appearance characteristic of uniaxial crystals. Isochromatic curves (lines of the same phase shift) have the form of concentric circles with the center at the exit point of the optical axis, and the Maltese cross retains the minimum intensity within the entire field of view (Fig. 2b (1)). There is a slight elongation of the cross in the vertical direction, but no signs of abnormal optical biaxiality, as observed before annealing, were detected (Figs. 2a, 2b (1)). When the laser power is increased to 90 mW, a common conoscopic pattern of a uniaxial crystal is obtained (Fig. 2b (1)). Thus, crystal 1 after annealing is more optically homogeneous and has good optical quality.

Fig. 2.

Conoscopic patterns of LiNbO₃:ZnO crystals 1–3 (a) before and (b) after annealing in the short-circuited state. λ = 532 nm, P = 1 and 90 mW.

The conoscopic pattern of crystal 2 obtained at a radiation power of 1 mW before annealing is close to common for uniaxial crystals (Fig. 2a (2)). There is a slight deformation in the area of the branches of the cross. The conoscopic pattern obtained at high laser radiation power (90 mW) is also typical of uniaxial crystals (Fig. 2a (2)) and shows no significant distortions, except for certain blurring of the image, indicating the presence of optical inhomogeneity, which is likely associated with the inhomogeneity of the impurity distribution (ΔC = C_u – C_l ~ –0.6 mol % ZnO) (Table 1).

The conoscopic patterns of crystal 2 after annealing are significantly more defective than before. At low laser power, the upper left branch of the cross is significantly blurred (Fig. 2b (2)). Some blurring and a more pronounced speckle structure of the image are also seen in the area of the lower left branch of the cross. In the area of both branches in the right half-plane, there are ruptures of the first isochromatic curve, as well as additional interference bands against the background of the main conoscopic pattern. The blurring is also seen on the periphery the right quadrant of the conoscopic pattern (Fig. 2b (2)). When the laser power is increased to 90 mW, some signs of abnormal optical biaxiality become visible in the area of the Maltese cross, including slight lucidity at the center of the cross, the shift of its fragments in the horizontal direction from the center, the difference from 90° of the angles between the cross branches, and the ellipticity of isochromatic rings (Fig. 2b (2)). There are also anomalies in the right half-plane of the conoscopic pattern, including decrease in contrast and clarity, blurring, and additional interference anomalies on the lower branch of the cross in the area from the fourth to the ninth isochromatic curve (Fig. 2b (2)). The results obtained indicate a significantly greater optical inhomogeneity of crystal 2, which underwent the annealing stage in the short-circuited state. Obviously, this inhomogeneity is due to peculiarities of the response of the crystal with a high concentration inhomogeneity to high-temperature annealing. Apparently, the annealing of such a crystal and, especially, the increase of the laser radiation power do not lead to “healing,” but to recharging and an overall increase in the number of charged defects. This is likely due to the nonequivalence of the activation energy of polar clusters in a compositionally inhomogeneous crystal.

Conoscopic patterns of crystal 3 before annealing observed at both low and high laser power have significant anomalies (Fig. 2a (3)). The blurring of the image is significant, and the sharpness and contrast are low. The conoscopic pattern at a laser power of 1 mW corresponds to a uniaxial crystal; however, the lower half-plane of the image has a blurred appearance with indistinguishable details (Fig. 2a (3)). When the power of the laser radiation is increased to 90 mW, the distortion of the conoscopic pattern of sample 3 increases (Fig. 2a (3)). There are signs of abnormal optical biaxiality: the Maltese cross is elongated in the vertical direction and isochromatic curves have the form of ellipses. The branches of the cross are also deformed. There is a significant blurring and reduction of the contrast of the image (Fig. 2a (3)). Probably, this deformation of the conoscopic pattern is associated with a greater scattering efficiency and photorefractive sensitivity of this crystal, which is confirmed by the results of the PILS study (Fig. 1 (3)). After annealing in the short-circuited state, the conoscopic patterns of crystal 3 at both low and high laser power acquire a more distinct and contrasting appearance (Fig. 2b (3)). This is due to the peculiarities of annealing of crystals having a small concentration inhomogeneity (ΔC = C_u – C_l ≈ –0.07 mol % ZnO (Table 1)). At the same time, the conoscopic patterns obtained at both low and high laser power still have minor signs of anomalous optical biaxiality (Fig. 2b (3)): the divergence of the Maltese cross, although without its rupture, in the vertical direction, which corresponds to the direction of deformation of the optical indicatrix of the crystal. Owing to the higher photorefractive sensitivity of this crystal, the increase of the laser power to 90 mW transforms the isochromatic curves into ellipses with a small-to-large semiaxis ratio of 0.9 : 1 (Fig. 2b (3)). It should be noted that the conoscopic patterns of crystal 3 (homogeneous doping) have a more blurred appearance and contain a greater number of defects than the conoscopic patterns of crystals 1 and 2 (direct doping). This allows us to conclude that LiNbO₃:ZnO crystals grown using homogeneous doping at an overall high compositional uniformity have worse optical quality than the crystals grown using direct solid-phase doping. At the same time, when the homogeneous doping method is used for growing LiNbO₃:ZnO crystals, the distribution coefficient K_d is significantly higher than in the case of the direct solid-phase doping method (Table 1). This allows growing LiNbO₃:ZnO crystals with a higher concentration of impurities at a lower ZnO concentration in the melt (Table 1). Therefore, the method of homogeneous doping essentially makes it possible to obtain LiNbO₃:ZnO crystals with a higher concentration of ZnO.

Thus, the study of the optical homogeneity of LiNbO₃:ZnO crystals obtained using different doping methods showed that the increase in the laser radiation power does not lead to significant distortions of conoscopic patterns, as is observed for crystals with a high photorefractive response [17]. This is in good agreement with the results of the PILS study, according to which photorefractive response for the studied LiNbO₃:ZnO crystals is insignificant (Fig. 1). Therefore, anomalies in the conoscopic patterns (deformations, ruptures of isochromatic rings, presence of additional interference bands) of the studied LiNbO₃:ZnO crystals are caused by the presence of charged structural defects and distortions of the optical indicatrix caused by mechanical stresses and compositional inhomogeneity of the crystals.

To evaluate the mosaic structure of LiNbO₃:ZnO crystals, an analysis of the obtained X-ray patterns was performed. Figure 3a presents typical scattering patterns in the region of the reflection (300) from the XZ planes of the LiNbO₃ crystal of the congruent composition and crystal 1 (Table 1). Figure 3b shows typical scattering patterns in the region of reflections (006), (0012) from the XY plane of crystal 1 grown using the direct doping method (Table 1).

Fig. 3.

Typical scattering patterns in area of (a) reflection (300) from the plane XZ and (b) reflections (006), (0012) from the plane XY of (1, 2) LiNbO₃ crystal of congruent composition and (3–8) LiNbO₃:ZnO crystal 1. Measurements were carried out (1, 3, 5, 7) without crystal rotation and (2, 4, 6, 8) with rotation.

Figure 4 shows typical scattering patterns in the region of the reflections (110), (220), (330) from the YZ plane of crystal 3 grown using homogeneous doping (Table 1).

Fig.4.

Typical scattering patterns in area of reflections (1, 2) (110), (3, 4) (220), (5, 6) (330) of LiNbO₃:ZnO crystal 3 obtained using the homogeneous doping method. Measurements were carried out (1, 3, 5) without crystal rotation and (2, 4, 6) with rotation.

As is seen from the X-ray diffraction pattern obtained without rotating the samples, the mosaic blocks are very large and it is possible to assess their disorientation in the mosaic crystal (Figs. 3, 4). The rotation results in averaging the scattering intensities from the disoriented regions of the crystal. Table 2 presents the calculation results on the average size of the blocks and the angles of their disorientation in the crystallographic directions [100], [001], [110] for a LiNbO₃ crystal of congruent composition and crystals 1–3. The average block size was calculated using the method of moments [14–16].

Table 2.   Structural characteristics of a congruent LiNbO₃ crystal and LiNbO₃:ZnO crystals calculated for reflections (300); (006), (0012); (110), (220), and (330). The crystallographic directions are [100], [001], and [110]

Table 2 presents the calculated values of interplanar distances d₃₀₀, d₀₀₆, d₀₀₁₂, d₁₁₀, d₂₂₀, d₃₃₀; average size of blocks 〈D₁₀₀〉, 〈D₀₀₁〉, 〈D₁₁₀〉; and the maximum and minimum disorientation of the blocks ($\alpha _{{\min }}^{^\circ }$, $\alpha _{{\max }}^{^\circ }$) in the directions [100], [001], and [110] for studied LiNbO₃ crystals of the congruent composition and crystals 1–3 grown using the direct and homogeneous doping methods.

It can be concluded on the basis of the analysis of the obtained data that the size of the blocks in the direction [001] is almost twice as large as the corresponding size in the direction [110] (Table 2). Regardless of the doping method, the block size in the direction [100] and [001] grows with the concentration of zinc in LiNbO₃:ZnO crystals. In the direction [100], this increase is very significant (Table 2). Crystal 2 has the maximum value of the block size in the direction [100], 〈D₁₀₀〉 = 4594 Å. Of all the studied LiNbO₃:ZnO crystals, the block sizes in all crystallographic directions are also maximal in crystal 2 (Table 2). At the same time, it can be concluded that, for all studied LiNbO₃:ZnO crystals, the block size differs to the smallest degree in the directions [001] and [110] and to the largest degree in the direction [100]. A crystal is considered structurally homogeneous if the size of its blocks in any crystallographic direction is the same and the block disorientation angle does not exceed ~20′. From the viewpoint of the block size, crystal 1 has the smallest spatial anisotropy, 〈D₁₀₀〉 : 〈D₀₀₁〉 : 〈D₁₁₀〉 ≈ 1 : 1.8 : 1.1; crystal 2 has the highest spatial anisotropy, 〈D₁₀₀〉 : 〈D₀₀₁〉 : 〈D₁₁₀〉 ≈ 4.9 : 1.86 : 1. In respect to the structural homogeneity, crystal 3 has an intermediate position, 〈D₁₀₀〉 : 〈D₀₀₁〉 : 〈D₁₁₀〉 ≈ 2.9 : 2 : 1 (Table 2). The minimum disorientation of blocks α_min is approximately the same for all studied LiNbO₃:ZnO crystals (Table 2). The maximum disorientation of α_max in the direction [100] is also almost the same and exhibits a slight increase with increasing zinc concentration (Table 2). Overall, the block disorientation is maximal in crystal 3. For this crystal, α_max > 20′ in all crystallographic directions (Table 2). The disorientation of blocks is the smallest in crystal 1 (Table 2). In addition, it should be noted that local distortions due to the presence of zinc cations in the structure do not lead to fluctuations of interplanar distances (d₃₀₀, d₀₀₆, d₀₀₁₂, d₁₁₀, d₂₂₀, d₃₃₀), i.e., to microstresses in LiNbO₃:ZnO crystals (Table 2).

The influence of the size of blocks and their disorientation on the macroscopic physical properties of LiNbO₃:ZnO crystals cannot be unequivocally characterized, especially considering a possible significant dependence of these properties on the direction in the crystal. At the same time, the results of X-ray diffraction studies correlate well with the PILS and laser conoscopy data. In particular, crystals 2 and 3 have the largest spatial anisotropy of the mosaic structure, crystal 3 is characterized by maximal disorientation of blocks, and crystal 1 is the most structurally homogeneous (Table 2). At the same time, the PILS patterns of crystals 2 and 3 before and after annealing indicate a greater scattering capacity (i.e., optical and structural heterogeneity) compared to crystal 1 (Fig. 1). This is especially pronounced for crystal 3 grown using the homogeneous doping method (Fig. 1).

CONCLUSIONS

We have performed a comparative study of the photorefractive properties and optical and structural homogeneity of LiNbO₃:ZnO crystals ([ZnO] ~ 5.4–6.4 mol % in the melt) grown by the Czochralsky method from a charge synthesized using precursors Nb₂O₅:ZnO obtained by homogeneous doping by zinc at the extraction stage of Nb₂O₅ (crystal 3) and a LiNbO₃:ZnO charge obtained by direct solid-phase doping (crystals 1 and 2).

It is shown that crystal 1 obtained by direct doping is characterized by high compositional homogeneity and has the highest structural and optical homogeneity. Crystals 2 and 3 have a greater spatial anisotropy of the mosaic structure, and the disorientation of the blocks is maximal in crystal 3. The optical inhomogeneity is especially strong for crystal 3 obtained by homogeneous doping; i.e., at overall high compositional uniformity, this doping method gives LiNbO₃:Zn crystals with worse optical and structural uniformity as compared to the direct solid-phase doping.

It is established that high-temperature annealing in the short-circuited state of the LiNbO₃:ZnO crystal with high concentration inhomogeneity leads to an increase in its optical inhomogeneity, and annealing of compositionally homogeneous crystals leads to “healing” of charged defects and overall improvement of optical characteristics.

It is shown that, when LiNbO₃:ZnO crystals are grown using the homogeneous doping method, the distribution coefficient K_d is significantly higher than when using the direct solid-phase doping.