Effect of doping of KDP crystal with amino acid l-arginine on the strength properties and character of laser damage

Graphical abstract


Introduction
KDP (KH 2 PO 4 ) group crystals which belong to a vast class of electrooptical and nonlinear optical media have found wide practical use in modern nonlinear optics and optoelectronics [1]. These crystals are applied in high-power laser systems with a peak power exceeding >1 GW/cm 2 emitting pulses with an energy of several kJ. This is due to a unique set of physical properties of group KDP crystals, in particular, their high transparency in a wide (ranging from the UV to near IR) spectral region, high laser damage resistance, good optical homogeneity combined with optical nonlinearity optimal for powerful high-energy laser systems. Moreover, the existing technologies make it possible to grow and mechanically treat large-size KDP crystals [2]. One of the methods for raising the efficiency of second harmonic generation (SHG) is introduction of organic molecules, e.g. amino acids, which possess high nonlinear coefficients, into the matrix of KDP crystal. As shown in [3], the efficiency of SHG for KDP with 0.3 wt% of L-arginine (L-arg) and KDP with 0.4 wt% of L-arg rises by a factor of 1.33 and 1.74, respectively, in comparison with that of pure KDP. On the other hand, in a number of studies [4,5] it has been revealed that doping with L-arg and other amino acids leads to certain decrease of the values of microhardness for KDP. It is well-known [6] that laser damage of KDP crystals essentially depends on their strength characteristics, and is followed by the appearance and propagation of microcracks caused by mechanical or heating pulse in the process of irradiation. Despite good optical properties, high structure perfection and the existence of the reproducible growth technology, the real laser-induced damage threshold in KDP crystals is essentially lower than the theoretically possible one [2]. Damage of the material under the action of laser pulses is one of the factors limiting the use of the crystals for generation of higher harmonics. Only few works [e.g. [6][7][8] are devoted to the regularities of brittle fracture in pure KDP crystals subjected to mechanical loading and laser irradiation. Nowadays, information about the damage processes in KDP doped with L-arg and other amino acids is absent. The present work reports the study of the strength properties of KDP crystals containing L-arg additions.

Material and investigation technique
KDP crystals, both pure and the doped, were grown from aqueous solution by the method of temperature reduction onto a point seed measuring 5 Â 5 Â 10 mm 3 . The content of L-arg in the mother solution (pH 4.0 ± 0.1) was 0.3, 0.4, 1.0 and 1.4 wt%. The solutions were filtered and then overheated during 24 h at T = 80°C. The relative solution supersaturation r was $2%, the rate of temperature lowering being 0.3°per day. To provide dynamical crystal growth conditions, the solution contained in the crystallizer was mixed at a rate of 70 rpm. The average crystal growth rates along a growth direction were V z = 2.0 mm per day and V x, y = 1.3 mm per day. The dimensions of KDP and KDP:L-arg (1.4 wt%) crystals were 79 Â 65 Â 54 mm 3 and 52 Â 42 Â 110 mm 3 , respectively.
At room temperature the structure of potassium dihydrogen phosphate is characterized by the space group I42d [9]. The morphology of KDP crystal is formed by tetragonal prism {1 0 0} and tetragonal bipyramid {1 0 1} faces (Fig. 1). All the grown crystals had well-developed prismatic and pyramidal growth sectors. The samples to be investigated were obtained from both growth sectors. We have previously shown that the doping in the range of 0.3-1.4 wt% concentrations lead to an increase of the efficiency of second harmonic generation by 1.33-2.53 times in the pyramidal sector, and in the 2.36-3.95 times in prismatic sector [10,11].
The study was performed on the mechanically ground and polished faces (0 0 1) and (1 0 0). After mechanical treatment the surface roughness was measured by Mod. TR_200 profilome-ter_profilograph. For each sample, the surface roughness average parameter Ra was determined by the results of four to five measurements. Standard deviation was 6.3% for the surface roughness parameter. The surface micro relief height of the crystals was 15 nm.
The mechanical properties of the crystals were investigated by the method of microhardness based on deformation by concentrated load. Indentation was realized using a PMT-3 device with the standard tetrahedral indenter, the loads ranged from 0.2 to 2 N. Most indentations were oriented along the diagonals of the square impression parallel to the [1 0 0] and [0 1 0] directions on the plane (0 0 1), and parallel to the [0 1 0] and [0 0 1] directions on the plane (1 0 0). The obtained indenter imprint and the radial cracks around it were measured by means of a computer-aided optical microscope Zeiss Axioskop 40A POL. The microhardness value was determined from the relation H V ¼ 1854P 1 =a 2 (kgf/mm 2 ), where P 1 is the indenter load (g), a is the imprint diagonal (lm). The crack resistance was estimated from the value of the threshold stress intensity factor R C for fracture according to the formula: Á P 1 l 3=2 (MPaÁm 1/2 ) [12], where l is the radial crack length from indentation centre to crack tip, E 1 , the Young' modulus (E 1 = 38.7 GPa) [8]. The measurements were realized on 10 imprints at each designated indentation load. The standard deviation in the measured value of mean microhardness was about 2%, whereas this deviation in the average crack length was found to be up to 10%. The samples were also irradiated by a single-mode pulsed YAG Nd 3+ laser at the wavelength k = 1.064 lm according to the scheme n -1 -n (90 pulses with the same energy, the samples were moved in the plane perpendicular to the laser beam), the radiation energy was 2.75 lJ, the pulse repetition frequency and pulse duration being 1 Hz and 10 ns, respectively. The criterion of laser breakdown was visually observed spark of high-temperature glow at the crystal damage. The measurement error was 10%. At loads of 0.5-2 N the microhardness of the planes (1 0 0) and (0 0 1) in the pyramidal sector of pure KDP crystal is practically the same. On the average, the microhardness of the samples cut out from the prismatic sector is somewhat higher than those obtained from the prismatic sector; thereat the hardness number for the plane (0 0 1) exceeds that for the plane (1 0 0) by $6-7%.

Experimental results
Doping of KDP with 1.4 wt% L-arg resulted in a diminution of the crystal's load-independent microhardness by $5-9% (this difference is not easily recognized) for the pyramidal sector and by $14-18% for the prismatic sector. The hardness number in the pyramidal sector for the face (1 0 0) of the doped crystal exceeded the values obtained for the face (0 0 1) by $6%. In the prismatic sector the hardness anisotropy of the planes (1 0 0) and (0 0 1) of the crystal containing the amino acid remained unchanged.
The character of the dependence of the microhardness on the concentration of L-arg for the pyramidal and prismatic sectors of KDP crystal is defined by the indenter loading. In particular, the experiments performed at an indenter load of 1 N have shown that doping with 0.3 and 0.4 wt% of L-arg leads to certain hardening of the crystal (Fig. 3-a). When the indenter load is 0.5 N the concentration dependence shows the hardness maximum only for the samples cut out from the pyramidal sector ( Fig. 3-b). In the samples obtained from the prismatic sector the hardness uniformly diminishes with the rise of the content of the amino acid in the crystal.
Presented in Figs Fig. 6 illustrates the dependence of the crack length l on the indentation load P applied on these investigated planes. At the same value of P the cracks propagate rather along the plane {2 2 1} than along (1 0 0) and (0 0 1).    (Fig. 7-a). The damage tracks stretched along h0 1 2i, observed from the plane (1 0 0) coincide with the crosses the investigated plane (1 0 0) and the planes {2 2 1} (Fig. 7-b). However, in the case when the damage occurs at local heating at the laser focus point, the anisotropy of crack propagation is much more essential than the one at local application of mechanical load. The planes of predominant crack propagation is {2 2 1}, the probability of laser-induced damage along the planes (1 0 0) and (0 0 1) is considerably lower. Table 2 presents the average values of laser-induced damage threshold in the pure and doped crystals for the pyramidal and prismatic growth sectors measured in the laser irradiation directions [0 0 1] and [1 0 0]. The value of laser breakdown threshold of KDP:L-arg (1.4 wt%) obtained for the prismatic sector is lower than the one of pure KDP crystal, although they are within the experimental error. However, for the pyramidal sector these values are higher. In pure crystals laser strength in prismatic growth sector is somewhat higher than in pyramidal one; in doped crystals an inverse picture is observed. The laser damage threshold of KDP crystals is limited by the presence of impurities. The higher crystalline perfection of pyramidal growth sector may be responsible for larger laser damage threshold in the said sector in doped crystals. For the plane (0 0 1) the laser-induced damage threshold for pure and doped KDP crystals is higher than that for the plane (1 0 0). This agrees with their lower microhardness values in the plane (1 0 0) in comparison with the one in the plane (0 0 1) in both growth sectors.

Discussion
The dependence of microhardness H v on indentation load -''reverse indentation size effect" (Fig. 2) in [13] is attributed to the influence of the surface layer of the crystal which is the most essential at low indenter loads. At low loads the thickness of the   surface-adjacent deformed layer resulting from machining is comparable with the indenter penetration depth, so the imprint undergoes the action of tensile surface stresses which increases its size. As the penetration depth increases, the effect of the distorted zone adjoining the crystal surface on the hardness becomes less noticeable. At higher loads (0.5-2 N) when these stresses are overcome, and the indentor penetrates the undistorted material, the microhardness value is independent of the load. Hardness anisotropy of (0 0 1) and (1 0 0) planes, and difference in microhardness values of the samples cut from the pyramidal and prismatic sectors of pure KTP crystals may be connected with predominant incorporation of the metal ions contained in the solution, such as Fe 3+ , Cr 3+ , Al 3+ etc. into the prismatic growth sector [14]. The charge state of the tetragonal prism {1 0 0} and tetragonal bipyramid {1 0 1} surfaces are different. The faces of the tetragonal bipyramid consist of two layers of K + -cations and two layers of H 2 PO 4 À -anions. The positively charged ions K + define on the surface of the tetragonal bipyramid, the faces {1 0 1} are positively charged. The surface of the tetragonal prism is formed by phosphate-ions mixed with K + -ions, therefore they are electrically neutral [15]. In the process of KDP crystal growth the faces {1 0 1} adsorb different anions and reject cations, {1 0 0} faces capture both positive and negative impurities. The impurity ions impede dislocation motion and to some extent favor the rise of the crystal's hardness in the prismatic growth sector. Deformation of the surface (1 0 0) involves larger number of the slip systems (1/2h1 1 1i {1 1 0}, 1/2h1 1 1i{1 1 2}, 1/2h1 1 1i{1 0 1}) than it takes place at deformation of the surface (0 0 1) (1/2h1 1 1i{1 1 2}, 1/2h1 1 1i {1 0 1} [16]), and this leads to different degrees of their hardening. The observed differences in the character of concentration dependence of microhardness for the pyramidal and prismatic sectors (Fig. 3) is obviously connected with different character of the interaction between the impurity ions and L-arg molecules at high and low indenter loads.
The observed decrease of microhardness values of KDP crystals doped with L-arg in comparison with those of pure KDP seems to be bound up with weakening of the interatomic bonds in the crystal lattice due to incorporation of 1.4 wt% of L-arg. As shown in [10,11], L-arg molecules enter into the prismatic and pyramidal growth sectors of KDP crystal that is caused by the ability of the amino acid molecules to form hydrogen bonds with the growing crystal face {1 0 0}, and to interact electro statically with the positively charged face {1 0 1}. Presented in Fig. 8  . The ionic bonds RAO are of two types -practically parallel and practically perpendicular to the direction [0 0 1] (Fig. 8b).
Doping of KDP crystal with the amino acid increases the lattice parameter a in comparison with the one of pure KDP by 2.45 Á 10 À4 Å and 2.42 Á 10 À4 Å, respectively, for the samples cut out from the prismatic and pyramidal sectors. Thereat, the parameter c of the doped crystal diminishes with respect to that of the pure crystal: Dc = À8.0 Á 10 À5 Å (the prismatic sector) and Dc = À9.5 Á 10 À5 Å (the pyramidal sector) [11]. Thus, both growth sectors show ''extension" of the elementary cell along the direction a and ''compression" along the direction c (Fig. 8) which decrease the lattice energy due to transformations in the system of hydrogen bonds and changes in electrostatical interaction between the ions RAO.
KDP crystals pure and doped with L-arg are characterized by high propensity to brittle fracture ( Table 1). The fracture toughness of these crystals is of the same level as in the optical materials such as the crystals of lithium-gadolinium borate -0.41 MPaÁm 1/2 and   [17]. As shown earlier [6,7], perfect cleavage is not characteristic of KDP crystals, however, they show microcleavage under the indentor parallel to the planes (0 0 1), (1 0 0) and {2 2 1}. The pattern observed for the doped crystals is analogous. Well-developed radial cracks propagate from the corner of the indentations for loads exceeding 20 N. At loads of 0.5-2 N linear plots of l against P (Fig. 6) have a slope of nearly 2/3, consistent with l $ P 2/3 [12]. At low loads the slope of these plots is not exactly 2/3, as expected from fracture mechanics concept. The surface stresses caused by mechanical treatment lead to expansion of surface traces of the cracks. Studies of the bulk laser-induced damage of optical materials have usually been directed towards measuring its threshold, and only few researchers have studied bulk damage morphology [18]. The post-exposure pattern of the damaged surfaces of KDP is similar to the one of crack formation at deformation under concentrated load. Certain disagreement between the obtained values of mechanical and laser damage resistance (Tables 1 and 2) seem to be due to different conditions of the formation of cracks in the crystal caused by concentrated load and laser pulses. The morphology of optical breakdown [19] in the range of nanosecond pulses is represented by melted zone with cracks, since damage is formed under the conditions of local heating at the laser focus point.

Conclusions
KDP crystals pure and doped with L-arg (0.3, 0.4, 1.0, 1.4 wt% of amino acid in the solution) were grown by the method of temperature reduction. It is shown that at loads of 0.2-2 N pure KDP and KDP:L-arg crystals show ''reverse indentation size effect" connected with the influence of the distorted zone of the crystal surface. For the crystals containing 0.3 and 0.4 wt% of L-arg the microhardness values rise. Doping of KDP with 1.4 wt% of L-arg leads to loss of hardness of both faces by $5-9 % and $14-18 % in the pyramidal and prismatic sectors, respectively. Brittle fracture of the pure and doped crystals has the same regularity, the planes of easy crack propagation are {2 2 1}, (1 0 0) and (0 0 1). The crack resistance of KDP crystals doped with 0.3 and 0.4 wt% of L-arg remains unchanged. The presence of 1.4 wt% of L-arg diminishes the coefficient of stress intensity. The pattern of crack formation in the zone of laser radiation action is similar to the one of brittle fracture obtained at local load application. The plane of predominant crack propagation is {2 2 1}. Strength characteristics of the doped crystals vary insignificantly all over the range of the investigated concentrations of L-arg. Thus doping with amino acid does not promote cracking under mechanical or thermal exposure in the process of generation of higher harmonics.