INTRODUCTION

Physical characteristics of lithium niobate (LiNbO3) crystals, which are among nonlinear optical materials of greatest practical importance, are commonly varied via changes in their stoichiometry (R = [Li]/[Nb]) or by doping [1]. Doping with impurities (Mg, Zn, Gd, In, Sc, and others) is an effective approach for reducing the photorefractive effect in LiNbO3 crystals [1, 2]. Their optical damage resistance rises stepwise at rather high (threshold) dopant concentrations [1, 2]. For example, the threshold for LiNbO3:Mg crystals is ~4.0–5.3 mol % Mg (~0.7–0.93 wt % Mg) [3–6]. However, such doping typically leads to chemical inhomogeneity of crystals.

Nonmetallic ions (such as B3+) are incapable of occupying the oxygen octahedra in LiNbO3 crystals. At the same time, our previous results demonstrate that, when used as a flux in crystal growth, boron oxide influences the structure and physicochemical characteristics of the melt, which makes it possible to obtain nearly stoichiometric LiNbO3 crystals from a congruent melt [7–9]. According to mass spectrometry data, boron concentration in LiNbO3:B crystals is extremely low: ~10–5 to 10–4 wt % [7].

Optical characteristics of LiNbO3 crystals can be significantly improved by codoping with Mg, an impurity capable of suppressing the photorefractive effect, and B3+, a nonmetallic element that markedly improves the optical homogeneity of LiNbO3 crystals. Masloboeva et al. [10] demonstrated the feasibility of growing optically and chemically homogeneous LiNbO3:Mg,B crystals from a homogeneously doped growth charge prepared from a mixture of Li2CO3 and Nb2O5:Mg,B as a precursor.

The objectives of this work were to grow optically and chemically homogeneous LiNbO3:Mg,B crystals from a growth charge prepared from a Li2CO3 + Nb2O5 + MgO + H3BO3 mixture and assess the chemical homogeneity, optical quality, optical damage resistance, and degree of unipolarity of the crystals.

EXPERIMENTAL

A granulated Mg–B codoped growth charge of congruent composition ([Li2O]/[Nb2O5] = 0.946) was prepared by a synthesis/granulation method from a Li2CO3 + Nb2O5 + MgO + H3BO3 mixture. Li2CO3 and Nb2O5 were prepared for synthesis in accordance with recommendations by Palatnikov et al. [11]. MgO was preannealed at ~850°C for 6 h. After mechanical activation of the starting materials for 24 h in a tumbling drum, the Li2CO3 + Nb2O5 + MgO + H3BO3 mixture was placed in a platinum crucible and slightly compacted. The granulation temperature and holding time were ~1235°C and 5 h. The mixture was heated at a rate of ~200°C/h. The percentage of boron in the growth charge and crystals was determined by inductively coupled plasma mass spectrometry (ICP MS) with an accuracy of ~10–6 wt %. Mg concentration was evaluated by atomic emission spectroscopy (ICPS-9000, Shimadzu) with an accuracy of ~10–3 wt %. The phase composition of the growth charge was determined by X-ray diffraction on a DRON-2 diffractometer (CuK radiation, graphite monochromator).

LiNbO3:Mg,B crystals were grown in the [001] direction by the Czochralski technique from 85-mm-diameter platinum crucibles in air, using a Kristall-2 furnace. The heating unit design ensured an axial temperature gradient of ~3°C/mm over the melt in combination with an extended isothermal region in the post-growth annealing zone. The translation rate was 0.6 mm/h, and the rotation rate was 12 rpm, which ensured a flat solidification front. To relieve the thermoelastic stress, the crystals were heat-treated at 1230°C for 15 h. The heating/cooling rate was ~50°C/h. To determine the dopant concentrations, plates ~0.8 mm in thickness were cut from the upper (cone) and lower (tail end) parts of each boule for the preparation of powder samples.

The rest of the LiNbO3:Mg,B crystal was converted to a single-domain state via high-temperature electrodiffusion annealing during cooling of the samples at a rate of ~20°C/h in the temperature range from ~1235 to 735°C in a dc electric field. The degree of conversion into a single-domain state (degree of unipolarity) was evaluated from amplitude–frequency response curves and by static d333 piezoelectric modulus measurements. The optical quality of the crystals was assessed by counting the number of scattering centers in a laser beam using a procedure described previously [12].

In photoinduced light scattering (PILS) and laser conoscopy experiments, we used a Nd:YAG laser (MLL-100, Changchun New Industries Optoelectronics, China, λ = 532 nm, I up to ~6.29 W/cm2). More detailed descriptions of the PILS and laser conoscopy characterization techniques and the block diagrams of the experimental setups used in this study can be found in Palatnikov et al. [13].

RESULTS AND DISCUSSION

For LiNbO3:Mg,B crystal growth, we prepared a granulated charge with a loose bulk density of 2.8 g/cm3. According to X-ray diffraction data, the growth charge corresponded to single-phase LiNbO3. The phases present were identified using the ICDD database (PDF 4, release 2020, card no. 04-009-8480). The growth charge contained ~0.67 wt % Mg and 0.009 wt % B, which roughly corresponded to dopant concentrations in the homogeneously doped growth charge used in a previous study [10]. We grew two LiNbO3:Mg,B crystals, 1 and 2, weighing 148.1 and 147.2 g, respectively, ~34 mm in diameter, with an ~20-mm-long cylindrical body. The total weight of the growth charge in the crucible was 1775 g. No more than 8% of the melt was used for the growth of the crystals.

Mg concentration in the melt (Cm) for the growth of the next crystal and the effective dopant distribution coefficient (Keff) were evaluated as described previously [13, 14]. The results are presented in Table 1. Analysis of the data in Table 1 indicates that direct solid-state synthesis of a growth charge from a Li2CO3 + Nb2O5 + MgO + H3BO3 mixture allows crystals with high chemical homogeneity to be grown: Mg concentration was shown to decrease by just ~0.01 wt % in going from the cone to the tail end. In the case of crystal growth from a melt containing ~0.67 wt % Mg (~3.9 mol % Mg), Kd was ~0.96–0.97.

Table 1. Magnesium concentration in the growth charge and LiNbO3:Mg,B crystals and distribution coefficient Kd
Full size table

In a study of LiNbO3:Mg crystals [6], it was shown that chemical homogeneity was only reached in the case of crystal growth from a melt containing ~4.6–5.3 mol % Mg, which was interpreted in terms of the structure and types of ionic complexes in the melt. If the melt contained <4.6 mol % Mg, attempts to grow chemically homogeneous LiNbO3:Mg crystals were unsuccessful [6]. Since the boron and magnesium dopants in the melt–crystal (LiNbO3:Mg,B) system have an additive effect, the Mg concentration in the melt at which crystals with a uniform dopant distribution can be grown is substantially lower (Table 1). Note that, according to the ICP MS results, boron concentration in the LiNbO3:Mg,B crystals was as low as ~10–5 wt %.

After conversion to a single-domain state, the end faces and base (x-cut) of the crystals were thoroughly polished. Figure 1 shows photographs of the LiNbO3: Mg,B crystals. High-speed evaluation of optical quality from the number of scattering centers in a laser beam showed that the density of scattering centers in crystal 1 was ~1.27 cm–3 (Table 2). In crystal 1, an imperfect region ~1 mm in thickness, with a rather high density of scattering centers, was detected ~1.5 cm from the cone. The imperfect region was located near a so-called “waist”—a local change in the diameter of the cylindrical body of the crystal—which probably originated from a short-term instability of growth conditions. At the same time, no scattering centers were detected in the bulk of crystal 2 (Table 2).

Fig. 1.
figure 1

Appearance of the LiNbO3:Mg,B crystals after conversion to a single-domain state and polishing.

Full size image
Table 2.   High-speed optical quality assessment results for the LiNbO3:Mg,B crystals
Full size table

Amplitude–frequency response (AFR) measurements showed that the main resonance peak in crystal 1 was located at a frequency of ~98.77 kHz and that the amplitude of the main resonance peak was eight times the background (Fig. 2a). In addition, the AFR curves had two secondary peaks at frequencies of ~117.93 and ~143.13 kHz, and one of them was a doublet. The AFR curve of crystal 2 had two resonance peaks (Fig. 2b). The main resonance was located at a frequency of ~95.25 kHz, and its amplitude exceeded the background by a factor of ~7.5. A side electroacoustic resonance peak was located at a frequency of ~142.99 kHz. Both resonances of crystal 2 had a regular shape in the form of sharp peaks (Fig. 2b). The results of the AFR measurements for the LiNbO3: Mg,B crystals indicate that they were rather successfully converted to a single-domain state. The presence of two side resonances in the AFR curve of crystal 1 and the doublet shape of the latter resonance seem to be due to the “waist” in the cylindrical body of the crystal and the presence of residual microdomains in this region of the crystal. The piezoelectric modulus d333, needed for assessing the degree of unipolarity of the crystals, was determined using an experimental setup for static piezoelectric effect measurements [15].

Fig. 2.
figure 2

AFR curves of the low-frequency part of the electromechanical spectrum of crystals (a) 1 and (b) 2.

Full size image

Figure 3 presents Qp(F) measurement results. The piezoelectric moduli d333 of crystal 1 (8.1 pC/N) and crystal 2 (8.6 pC/N) approach standard values (d333 ≥ 8 pC/N according to Blistanov et al. [16]). The d333 of crystal 1 is somewhat smaller than that of crystal 2. This seems to result from the presence of residual microdomains in the small “waist” region. Thus, the growth process was very sensitive to stability of its conditions, which was reflected in both the shape of the AFR curve and the d333 piezoelectric modulus (Figs. 2, 3). In the initial state, mechanical stress applied to a LiNbO3: Mg,B crystal induces a small polarization charge due to a weak native unipolarity of a polydomain crystal (Fig. 3, curve 3).

Fig. 3.
figure 3

Qp as a function of F for crystals 1 and 2 after conversion to a single-domain state; (3) originally polydomain crystal 2.

Full size image

The PILS results for the LiNbO3:Mg,B crystals showed that they had rather high optical quality (Fig. 4). It is clearly seen in Fig. 4 that, even at a comparatively high excitation intensity (~6.29 W/cm2), there was no photorefractive response, the PILS indicatrix did not open, and no laser beam destruction occurred. This points to suppression of the photorefractive effect in the LiNbO3:Mg,B crystals 1 and 2.

Fig. 4.
figure 4

PILS patterns of the LiNbO3:Mg,B crystals (λ = 532 nm, I = 6.29 W/cm2).

Full size image

Macroscopic optical homogeneity of the LiNbO3: Mg,B crystals was assessed by laser conoscopy (Fig. 5). In spite of some distortion, mainly due to the large thickness of crystals 1 (~38 mm) and 2 (~40 mm) and the associated optical distortion accumulation during laser beam propagation, the conoscopic patterns of the LiNbO3:Mg,B crystals showed that they had rather high optical quality. It is worth noting that, to examine conoscopic patterns, use is commonly made of plates 1 to 3 mm in thickness. In the present conoscopic patterns, the “Maltese cross” is slightly elongated in the horizontal (crystal 1) or vertical (crystal 2) direction, corresponding to the optical indicatrix distortion direction, with no discontinuity in the center of the cross. Moreover, isochromes retain the shape of concentric circles, without acquiring ellipticity (Fig. 5). The anomalies correspond to a small optical biaxiality due to the optical distortion accumulation during laser beam propagation. Raising the laser output power to 90 mW caused no further distortion of the conoscopic patterns. This suggests that the LiNbO3: Mg,B crystals had no photorefractive response, as supported by the present PILS results (Fig. 4). The conoscopic patterns of crystal 1 are more imperfect than those of crystal 2, suggesting that crystal 2 has better optical homogeneity. This is due to the presence of an imperfect region in crystal 1 around the “waist”—a local change in the diameter of the cylindrical body of the crystal, which originated from a short-term instability of growth conditions.

Fig. 5.
figure 5

Conoscopic patterns of crystals 1 and 2 (λ = 532 nm, Р = 1 and 90 mW).

Full size image

CONCLUSIONS

A technologically feasible process has been designed for the preparation of a LiNbO3:Mg,B growth charge with preset dopant concentrations by solid-state synthesis. We have found conditions for the growth of optical-quality LiNbO3:Mg,B crystals with a highly uniform dopant distribution.

Using AFR and d333 piezoelectric modulus measurements, we have demonstrated a high degree of unipolarity of the LiNbO3:Mg,B crystals. The present PILS results for the LiNbO3:Mg,B crystals confirm that they have high optical damage resistance, and the laser conoscopy results demonstrate that they have rather high optical homogeneity.

Thus, LiNbO3:Mg,B crystals can be regarded as a new optical material for laser light conversion.