Influence of Growth Conditions on Mechanical Properties of K₂Ni_XCo_(1−X) (SO₄)₂·6H₂O Crystals

Abstract

K₂Ni_XCo_(1−X) (SO₄)₂·6H₂O (KCNSH) mixed crystal is a promising material for solar blind optical filters, combining high transparency in the ultraviolet range with effective suppression of the visible spectral region. Increasing the mechanical strength of these crystals is important to enable them to be machined in the manufacture of optical elements. A comprehensive study of the inhomogeneities and crack resistance of KCNSH crystal as a function of the growth conditions was carried out. The influence of the radial and mosaic inhomogeneity, as well as other structural defects, on the crack resistance of the crystals was analyzed. To assess the crack resistance of crystals, the parameters c_a (crack length), c/a (the ratio of crack length to the size of the indentation), and K_C (fracture toughness) were used. The correctness of the obtained results was analyzed. The conditions for growing KCNSH crystals with the best crack resistance were determined on the basis of the results of the study. It is shown that growing the mixed crystals using the temperature difference technique with a peripheral solution supply into the shaper provides the best crystal quality.

1. Introduction

The advantage of UV-C range in diagnostic equipment is extremely low level of background noise due to the almost complete suppression of solar radiation in this spectral region by the Earth’s ozone layer. The technology for recording the radiation with wavelengths of 250–280 nm is called “solar-blind” (SBT) and is intensively developing across the world. SBT devices are mainly used to detect sources of electric corona discharge and flame. If the residual noise is close to 10⁻¹⁸ W·cm²/Hz^1/2 and transmission in the working range is 20–70%, it becomes possible to register single photons, as well as to measure their spatial and temporal characteristics [1]. Such registration of the optical signal is called “monophotonic technology” (MPT). It opens up new fields for a number of unique technical solutions. For example, the MPT-based remote system for monitoring of power lines provides not only remote registration of current leak but also comprehensive quantitative diagnostics of equipment. Such technology is also realized in systems for automatic landing of aircraft and collision prevention, as well as systems for detecting forest fires and their targeted extinguishing from aircraft [2,3].

A mandatory element of such devices is an effective band filter transparent in the 250–280 nm range and opaque in the other ones. Such zone filtering allows one to maintain a high signal-to-noise ratio and to provide an unique sensitivity of the device. At present, crystals of hexahydrates of cobalt and nickel sulfates are used as optical filters in SBT devices. These are crystals of α-hexahydrate of nickel sulfate α-NiSO₄·6H₂O (α-NSH) [4] and Tutton’s sulfo-salts (M₂¹⁺M²⁺ (SO₄)₂·6H₂O, where M⁺ is an alkali metal or ammonium, M²⁺ is Co²⁺ or Ni²⁺) [5,6,7,8,9].

The characteristics of MPT devices can be improved mainly by increasing the efficiency of signal filtering and noise suppression. To achieve this aim, it is necessary to increase the transmission in the operating spectral range and/or increase the absorption in the nonoperating ones. The abovementioned crystals of hexahydrates of nickel and cobalt sulfates have unwanted transmission bands in the visible spectral range around 490 nm (green band) and 710 nm (red band), respectively. This adversely affects the efficiency of radiation filtering.

This problem can be solved by creating an optical filter based on mixed crystals (solid solution) of nickel and cobalt salts, where hexahydrate complexes of nickel and cobalt ions mutually suppress the parasitic bandwidth of each other. This will increase the efficiency of radiation filtration and, as a result, the sensitivity of the devices and their range of action. Mixed crystals of nickel–cobalt sulfate duodecimhydrate [10], cobalt–nickel–ammonium sulfate hexahydrate (ACNSH) [11], and cobalt–nickel–potassium sulfate hexahydrate (KCNSH) [12] are suitable for these purposes. All these crystals have similar transmission spectra; they are transparent in the range of 250–280 nm, while green and red transmission bands are significantly suppressed. However, all the crystals described in [10,11,12] had low transmission in the UV range due to the scattering of radiation by numerous structural defects.

A common problem for water-soluble mixed crystals is an enormous quantity of defects, which strongly limits their practical application. There are several fundamental reasons for this problem.

• According to the Gibbs phase rule, at constant pressure, a ternary system (water and two salts) has two degrees of freedom: temperature and concentration. As a result, at any fixed temperature, a continuous series of saturated solutions of various compositions and crystals in equilibrium with them exists. This leads to the fact that the solution composition may change during the crystal growth; in addition, it may be different in different areas of the crystallizer bulk.
• Generally, the distribution coefficients of isomorphic components are different and differ from unity. Thus, in the course of crystal growth, the solution is enriched with one component and depleted of the other. As a result, the zonal inhomogeneity of mixed crystals becomes many times more intense compared with single-component crystals. An additional problem is the presence of an initial transient zone, which forms at the initial stage of crystal growth due to the formation of a diffusion layer, as described in Tiller’s work for a melt with an impurity [13]. In this area, the crystal composition changes are especially noticeable.
• The distribution coefficients of the components also differ in the different growth sectors. As a result, in mixed crystals, the sectorial inhomogeneity is much more noticeable in comparison with that in single-component crystals.
• A specific process called the isomorphic substitution reaction develops upon contact of a mixed crystal with a saturated solution whose composition does not meet the equilibrium conditions [14,15]. As a result, the crystal surface turns into a chaotic mosaic of regions where the processes of crystal dissolution and the growth of a new phase occur simultaneously. Being immured in the crystal bulk during growth, they lead to the formation of mosaic inhomogeneity, which is characteristic for multicomponent crystals only [16].

The high inhomogeneity of the composition of mixed crystals leads to the generation of high internal stresses. They are sources of defects such as cracks, inclusions, and dislocations. The reason for the formation of inclusions is the loss of stability of growth steps during movement in a nonuniform elastic stress field [17,18,19]. Inclusions relax the elastic stresses [20], but they are also the scattering centers and can lead to the formation of dislocations [17].

The KCNSH crystal is the most studied and technologically advanced among the three abovementioned cobalt–nickel sulfates. It has been shown that the sectorial inhomogeneity of these crystals can reach 12 wt.% [21], zonal inhomogeneity can reach −8 wt.% [22], radial inhomogeneity can reach −4.4 wt.% [23], and mosaic inhomogeneity can reach −2.7 wt.% [23]. Moreover, a direct relationship between the value of sectoriality and the number of inclusions has been demonstrated [21].

To eliminate the sectorial inhomogeneity, it has been proposed to grow these crystals in a shaper in order to provide the growth of only one growth sector [24]. To suppress the zonal inhomogeneity, a method of solution feeding according to a special law has been developed [25]. The suppression of these strongest compositional inhomogeneities had a noticeable positive effect. Transmission of 87% was achieved at a wavelength of 250 nm, which indicates the absence or an insignificant number of inclusions in the crystals.

Today, the KCNSH crystal is the best material for solar-blind optical filters in terms of its optical properties. However, the obtained crystals are very fragile and often crack during mechanical treatment (cutting, grinding, and polishing with abrasive materials) [23]. This indicates the rather high level of residual internal stresses in the crystals, probably due to the radial and mosaic inhomogeneity, which has not received sufficient attention. In [23], the conditions for the elimination of or significant reduction in these inhomogeneities were determined. It was shown that mosaic inhomogeneity is suppressed by supercooling above 2 °C. Moreover, the radial inhomogeneity practically disappeared when a swirling flow of 55–135 cm/s was created in the shaper.

This paper represents an important step in determining the optimal growing conditions for mixed KCNSH crystals ensuring their mechanical strength. Its purpose is to verify whether the conditions found in [23] actually make it possible to increase the resistance of crystals to crack formation and make them suitable for the mechanical processing in the manufacture of optical filters.

It should be noted that we are not aware of any studies on the mechanical properties of mixed water-soluble crystals, with the exception of KCNSH. As indicated above, the reason is apparently the high imperfection and the lack of potential practical use of such crystals. As for the KCNSH crystals, their microhardness was investigated in [26], and their resistance to cracking was studied depending on the direction of growth and the growth rate in [27]. Since the standard strength test methods applied for metals are inapplicable to fragile anisotropic water-soluble crystals, the development of a method for determining their crack resistance is a new problem.

In the present article the dependence of the crack resistance of KCNSH crystals on supercooling and the parameters of the solution flow are discussed. Several different parameters are analyzed in terms of the development of the method for determining the crack resistance of crystals.

2. Materials and Methods

2.1. Growing of Mixed KCNSH Crystals

The raw materials for growing KCNSH crystals were NiSO₄·7H₂O (chemically pure grade), CoSO₄·7H₂O (chemically pure grade), and K₂SO₄ (reagent grade). KCNSH crystals were grown from solutions with the mole ratio of isomorphic components KCSH:KNSH = 1:2. Solutions were prepared by dissolving the salts in water according to the following reaction:

K₂SO₄ + 0.67NiSO₄·7H₂O + 0.33CoSO₄·7H₂O → K₂Co_0.33Ni_0.67(S0₄)₂·6H₂O + H₂O.

(1)

The solutions were saturated at 40–43 °C and filtered through the track membrane with a pore diameter of 0.2 μm.

KCNSH crystals were grown in shapers using the temperature difference technique with feeding. The growth technique was described in detail in [19]. The volume of the crystallizer was 1 L, while the total volume of solution in the growth system was 4.5 L. The diameter of the cylindrical shaper was 30 mm. Plates of mixed KCNSH crystals 28 mm in diameter were used as seed crystals. They were mounted in shapers, and the crystals were grown top-down. The supply of the solution to the shaper was carried out in two ways: upright supply of the solution to the center of the crystal face and peripheral supply tangentially to the wall of the shaper to create a “swirling” flow. Supercooling of solution was 1.3–2.5 °C. Solution velocity varied from 10 to 175 cm/s.

2.2. Study of Mechanical Properties of KCNSH Mixed Crystals

The samples were (110) crystal plates cut using a thread saw and polished with diamond powder to optical quality. Microhardness H and crack length c were measured in the vicinity of the center of the sample in two areas located at a distance of 10–15 mm from each other. In these areas, there were no defects visible by microscope.

The microhardness and crack formation were studied under indentation by concentrated load. Indentation by the Vickers pyramid and measurement of the diagonals d of the recovered indentation were carried out using a standard attachment to the Neophot-21 microscope. The diagonal of the recovered indentation d was measured immediately after indentation. The crack length c was measured at 500× magnification using an optical microscope. The samples were placed in such a way that the measured diagonal d of the indentation made an angle of 45° with [001]. Figure 1 shows the photo of a real indentation and the scheme of indentation.

The microhardness H of each crystal was calculated after averaging over a complete sample (10 indentations) using the following formula for the Vickers pyramid [28]:

H (kgf/mm²) = 1.854 P (kgf)/d_st² (mm²),

(2)

where P is the applied load, and d_st is the statistical average for a local or complete sample. The numerical coefficient corresponds to the geometry of the Vickers indenter. In Equation (2), the units of measurement are indicated in square brackets. In this paper, the values of P and H are given in N and GPa.

A widely used method for determining the fracture toughness K_C (critical stress intensity) by the length of cracks c formed in the corners of the indentation was used to compare the crack resistance of crystals [29].

$$K_C=(0.016\pm 0.004)\sqrt{E/H}(P/c^{3/2}),$$

(3)

where E is Young’s modulus. The crack length c for each crystal was determined by averaging over two samples of 20 cracks (five indentations in each of the two crystal areas) and then over a complete sample. The measurement error was considered as a standard statistical error of the average value δc; in the case of microhardness δH = 2H (δd/d_st), δd is the standard error of the sample average. In most cases, δH was 2–3%, while δc was 5–7%.

In addition, crystals were compared by the crack length c and the parameter c/a, where a = d_st/2 (Figure 1b); c/a clearly demonstrates the ratio of the crack length to the size of the indentation.

On the one hand, for the crystals under consideration, the application of a load P > 1 N often leads to brittle destruction of the area around the indentation, thus preventing correct measurement of microhardness and crack length. On the other hand, for a correct assessment of K_C, the c/a ratio should be ≥2 [29]. The measurements were carried out at indenter loads of P = 0.34, 0.58, and 0.93 N for five indentations in each series. As shown below, both requirements were met at the selected loads.

2.3. X-Ray Projection Topography

The study of the real structure of mixed KCNSH crystals was carried out using Lang X-ray projection topography (FSRC “Crystallography and Photonics”, Moscow, Russia) [30]. In this method, a collimated X-ray beam in the form of a vertical strip equal in height to the size of the sample is used (Figure 2). Scanning allows to get a picture of the real structure of the entire sample. The slit prevents the direct beam from hitting the photographic plate. The characteristic radiation Mo_ka1 with a wavelength λ = 0.0709 nm was used. The samples were (001) plates about 1 mm thick. Photographic plates for nuclear research P-50 with an emulsion thickness of 50 microns were used for the image registration.

2.4. Study of Optical Transmission Spectra of KCNSH Mixed Crystals

To measure the transmission spectra of samples in the ultraviolet and visible ranges, an automatic two-beam spectrophotometer “Cary 300 UV/Vis” was used. It allows recording the spectra in the 200 to 900 nm wavelength range. Samples were 10 and 20 mm thick cylinders with two parallel polished faces. To improve the accuracy of the results, the spectrophotometer was calibrated before each measurement.

3. Results and Discussion

3.1. Mechanical Properties of Mixed KCNSH Crystals

Concentrated load indentation is widely used for the study of brittle materials such as ceramics, glasses, and brittle polymer systems [29,31,32]. Indentation methods are perhaps the only way to quantitatively compare the mechanical properties of brittle single crystals. Having extremely low plasticity during active deformation, such single crystals, as a rule, do not form yield plateaus and spontaneously fail at very low strain values of ε < 1% [33,34].

The conditions for crystal growth and the parameters of their inhomogeneities are presented in

Table 1. Data on inhomogeneities of KCNSH crystals in various growth conditions.

Crystal	Supercooling ΔT, °C	Growth Rate, mm/day	Solution Velocity, cm/s	Solution Supply	Inhomogeneity, $\mathsf{\Delta }{\mathit{x}}_{\mathit{N}\mathit{i}}.\mathbf{a}т.\%$
					Radial	Mosaic
1	2.1	0.7	175	center	4.4	0.15
2	1.6	0.33	85	center	4	0.28
3	1.3	0.18	170	periphery	1.6	0.34
4	1.9	0.5	135	periphery	0.5	0.15
5	1.7	0.37	55	periphery	0.17	0.17
6	1.5	0.25	55	periphery	2	0.19
7	2.2	0.82	10	periphery	2.1	0.1
8	2.5	1	10	periphery	1.8	0.05

.

A comparison of the measured parameters in two areas gave an estimate of the uniformity of the samples. For comparison, the microhardness and crack formation in single-component KCSH and KNSH crystals, which are the end-members of the studied isomorphic series, were studied.

The measurements of microhardness showed that its values at different loads differed by no more than 5%.

Table 2. Microhardness H and crack length c for (110) plane of KCNSH crystals at the load of 0.58 N.

Crystal	Hi, GPa	Ha, GPa	ΔH, %	ci, μm	ca, μm	Δc, %	c/a
KCSH	1.105 1.222	1.161	−4.82 +5.25	45.25 40.13	42.69	+6.00 −6.00	2.79
KNSH	1.240 1.304	1.272	−2.52 +2.52	32.75 34.75	33.75	−2.96 +2.96	2.31
1	1.253 1.311	1.281	−2.19 +2.34	43.25 42.63	42.94	+0.72 −0.72	2.95
2	1.243 1.262	1.253	−0.80 +0.72	39.00 48.25	43.63	−10.61 +10.59	2.96
3	1.331 1.429	1.367	−2.63 +4.54	34.13 39.79	36.25	−5.85 +9.77	2.57
4	1.059 1.089	1.074	−1.40 +1.40	36.00 43.50	39.75	−9.43 +9.43	2.50
5	1.256 1.308	1.281	−1.95 +2.11	35.88 36.50	36.19	−0.86 +0.86	2.48
6	1.325 1.349	1.336	−0.82 +0.97	38.88 40.88	39.88	−2.51 +2.51	2.79
7	1.175 1.213	1.194	−1.59 +1.59	44.75 41.88	43.31	+3.32 −3.30	2.87
8	1.281 1.377	1.328	−3.54 +3.69	42.38 40.50	41.44	+2.27 −2.27	2.89

shows the values of the microhardness H_i and the crack length c_i in each area, as well as the average values of H_a, c_a, and c_a/a at a load of P = 0.58 N. It is possible to estimate how much the two sample areas differed from each other according to ∆H = (H_i − H_a)/H_a and ∆c = (c_i − c_a)/c_a (Table 2).

Cracks that occurred during the indentation of crystals of all compositions were formed at very low loads; in most cases they were straight and outcrop close to the corners of the indentation. A halo of the fracture zone was formed around the indentation (Figure 1a), indicating the presence of cracks hidden under the surface. Thus, we suppose that penny-like cracks outcropped after indentation. This was important for the selection of Equation (3) to estimate the fracture toughness K_C according to the results of indentation.

There are no data on the value of E for KCSH, KCSH, and KCNSH crystals in the literature; however, this difficulty can be overcome. All crystals are similar in term of lattice parameters; moreover, they are oriented in the same way relative to the applied load and exhibit a similar microhardness. Therefore, for a rough estimate of K_C, we took the value E = 13H₀, where H₀ is the average value of microhardness for all measured results for each sample (

Table 3. Microhardness H₀, fracture toughness K_C, and its standard deviation S for the (110) plane of KCNSH crystals.

Crystal	H0, GPa	KC, MPam1/2	S/KC, %
KCSH	1.174	0.123	10.0
KNSH	1.262	0.167	8.4
1	1.281	0.119	-
2	1.253	0.116	-
3	1.323	0.141	8.8
4	1.161	0.135	6.2
5	1.254	0.142	6.3
6	1.260	0.135	9.4
7	1.242	0.121	4.7
8	1.314	0.128	7.0

). The coefficient “13” is the most typical for the materials considered in [29]. Thus, Equation (3) was transformed into the following expression:

$$K_C=0.016\sqrt{13H_0/H_a}(P/c_a^{3/2}).$$

(4)

Equation (4) allows one to estimate K_C from the obtained data. The K_C values were calculated for all values of load and averaged for each sample. The K_C values and the standard deviations S divided by the average K_C are shown in Table 3.

3.2. Refinement of Growth Conditions for Obtaining Highly Perfect KCNSH Crystals

To clarify the growth conditions for shaped KCNSH mixed crystals, we matched the data on their crack resistance to inhomogeneities. Figure 3 shows the dependences of parameters c/a, c_a, and K_C on the value of the radial inhomogeneity. As for the mosaic inhomogeneity, because of its small value, it did not significantly affect the crack resistance of crystals: for five out of eight crystals, the mosaic inhomogeneity was an order of magnitude smaller than the radial one.

Figure 3 shows that all three plots adequately reflected a decrease in the crack resistance of crystals with an increase in their radial inhomogeneity. However, there was a certain difference between them. In order to apply the techniques for optical processing of KCSH and KNSH to KCNSH crystals, it was necessary to ensure that the mechanical parameters of mixed crystals were comparable to the parameters of single-component crystals. Parameters c_a and K_C (Figure 3b,c) for crystals #3–6 and #8 fell into the intervals between the values for KCSH and KNSH, whereas only crystals 3–6 had parameter c/a between that for KCSH and KNSH (Figure 3a). To assess the correctness of these results, we can consider the physical meaning of the parameter c/a.

The semi-diagonal (hereinafter referred to as the size) of the indentation is $a~1/\sqrt{H}$; therefore, $c/a~c\sqrt{H}$. Accordingly, when assessing crack resistance using the parameter c/a, we take into account not only the crack length, but also the microhardness of the crystal. The microhardness of a material characterizes its resistance to deformation and destruction. The microhardness of KCNSH crystals of the same composition varies and, as a rule, it depends on the load. If the formation of cracks in two crystals begins at the same load, then the crack length will be longer for the crystal with the higher microhardness. Thus, the crack length is not a valid parameter for assessing crack resistance of material.

Since the size of the indentation a is, in a way, related to Young’s modulus E, the crack resistance of the crystal can be adequately assessed by the fracture toughness parameter K_C calculated using Equation (3). However, we do not know the elastic characteristics of KCNSH crystals; thus, we are forced to use the approximation E ≈ 13H₀; the values of K_C calculated by Equation (4) do not take into account the variations of Young’s modulus.

In this situation the c/a parameter seems to be the most suitable for assessing crack resistance. In addition, from the data [31], it can be seen that this parameter correlates with the value of fracture toughness K_C. Accordingly, we concluded that only crystals #3–6 were comparable in terms of crack resistance to KCSH and KNSH crystals.

Analyzing the results, it should be noted that crystals #1 and #2 turned out to be the most heterogeneous, as shown in the previous studies [35,36]; they were grown by supply of the solution into the center of the shaper directly onto the crystal surface. They also had the worst crack resistance evaluated by all three parameters (Figure 3). Therefore, such a mode of supplying solution is not acceptable for obtaining perfect mixed crystals. Peripheral supply tangentially to the wall of the shaper makes it possible to obtain significantly more homogeneous crystals. Nevertheless, not all conditions yield good results: crystals grown at the initial flow velocity of 10 cm/s (at the outlet of the tube) (Table 3, crystals #7 and #8) turned out to be rather heterogeneous.

In our previous studies, it was shown that the kinetic growth mode occurs at a flow velocity of 37 cm/s [36]. This was achieved in most of our experiments. At a flow velocity of 10 cm/s, the kinetic growth mode was not achieved. Thus, a boundary diffusion layer was formed near the crystal face; this layer was characterized by a high concentration gradient normally to the crystal surface. It is also necessary to take into account that crystals #7 and #8 were grown at the highest supercooling level and their growth rate was the highest. Under these conditions, the morphological instability of the growing face led to the capture of inclusions and the formation of other defects. The photos of crystals #7 and #8 (Figure 4) confirm these conclusions. This is an additional argument in favor of the facts that (1) crystal #8 is not suitable for machining, and (2) the positive assessment of its crack resistance according to the criteria of c_a and K_C is wrong.

Among crystals #3–6, crystal #5 was the most uniform in the radial direction (Figure 5a); it also demonstrated the best crack resistance. Crystal #6 (Figure 5b), grown under similar conditions (the tangentially supply of solution at V = 55 cm/s), showed a rather low resistance to cracking (c/a = 2.79), like the KCSH crystal, and a rather high heterogeneity (Table 1). The crystal #6 contained a large number of dislocations originated from the seed (Figure 5b); this reduced its crack resistance. Apparently, a large number of dislocations also led to the simultaneous existence of several vicinal hillocks on the crystal surface. This was evidenced by numerous vicinal–sectorial boundaries and bands of zonal heterogeneity. Generally, crystal #6 was also very homogeneous except for a narrow area at its periphery. It is possible that, at this area of measurement, one of the vicinal–sectoral boundaries came to the surface of the crystal.

Crystal #4 was close in c/a to crystal #5 and very homogeneous; it also demonstrated a fairly high resistance to cracking (Table 2, Figure 3a). Its X-ray topogram is shown in Figure 5c.

Summarizing the above, it is preferable to grow mixed crystals in shapers using the temperature difference technique with a peripheral supply of solution at velocities from 55 to 170 cm/s. Supercooling should be conducted at 1.7–2.0 °C.

It should be noted that the selected range of flow velocities allows one to avoid solution feeding for the initial transient mode, since, in this case, the kinetic growth mode is achieved, and the diffusion boundary layer is almost absent. As a consequence, an initial transition area is absent in crystals, which significantly simplifies the growth process. However, the developed recharge technique for the initial transient mode can be useful if crystal growth is carried out with slow stirring of the solution or without stirring, i.e., in static mode, as was done, for example, in [37].

3.3. Optical Elements Made from KCNSH Crystals

Under the above selected growth conditions, KCNSH shaped crystals suitable for the manufacture of optical elements with a diameter of 25–32 mm were grown (Figure 6).

It should be noted that, for the fabrication of optical filters from mixed KCNSH crystals, all operations (cutting, grinding, polishing) were carried out according to standard techniques used for KCSH and KNSH crystals. None of the samples were cracked or damaged in any way. The optical transmission spectra of UV filters are shown in Figure 7. All filters exhibited excellent transparency in the UV range (86–91%), as well as effective suppression of unwanted transmission in the visible range.

Figure 7b shows the optical transmission spectrum of the KCNSH crystal, the sample thickness was 20 mm. It can be seen that a twofold increase in the thickness of the crystal samples led to the almost total suppression of radiation in the nonoperating range of the spectrum (350–900 nm).

Taking into account the relatively high dehydration temperature of KCNSH [38] and the obtained optical characteristics of these crystals, it can be concluded that KCNSH crystals grown from solution with a ratio of isomorphic components KCSH:KNSH = 1:2 using the temperature difference technique with solution feeding are the best material for optical elements in solar-blind technology devices at present.

4. Conclusions

A comprehensive study of the inhomogeneity of the composition and structure of the mixed KCNSH crystal and its crack resistance dependent on the growth conditions was carried out. It is shown that the crack resistance of these crystals was determined by both radial and mosaic inhomogeneity, and it depended on the presence of structural defects in the crystals (dislocations, inclusions, boundaries). Due to its small value, the mosaic inhomogeneity in the crystals under study does not have a noticeable effect on the crack resistance.

To assess crack resistance, the parameters c_a (crack length), c/a (ratio of crack length to indentation size), and K_C (fracture toughness) were used. It is shown that all these parameters properly reflected a decrease in the crack resistance of KCNSH crystals with an increase in their radial inhomogeneity. However, the most correct assessment of crack resistance was obtained using the c/a parameter.

According to the obtained results, the preferable conditions for growing homogeneous KCNSH crystals with the maximum resistance to cracking were selected: growth using the temperature difference technique from solutions with a component ratio of KCSH:KNSH = 1:2 and supercooling conducted at 1.7–2.0 °C with peripheral solution supply at velocities of 55–170 cm/s.

The optical elements made of KCNSH crystals have better optical characteristics in comparison with KCSH and KNSH crystals used as materials for UV filters at present.