TEXTURE AND ANISOTROPY BY FORMATION AND DECOMPOSITION OF NICKEL HYDRIDE

The effect of microstructure and crystal direction on the extent of phase transformation (EPT) of Ni 
into β-NiH by cathodic charging with H has been investigated by X-ray diffraction, EPT is controlled 
by the crystal direction in the case of heat-treated specimens. In the case of electrodeposited 
specimens, the imperfections of which are commensurate with those of cold-worked metals, EPT is 
controlled by both the crystal direction and the “dislocation-induced” anisotropy at the same time. 
The decomposition of β-NiH follows the Kolmogorov–Avrami–Johnson–Mehl kinetics. The microstructural 
anisotropy induces anisotropy in the rate of decomposition of β-NiH.


INTRODUCTION
It is well known that as a result of the cathodic charging of nickel with H in the presence of inhibitors, an instable at room temperature nickel hydride phase is produced (Janko 1960, Boniszewski et al. 1961, Baranowski 1964, Calbe et al.  1964, Borbe et al. 1980, Pielaszek 1985).Hydrogen reacts reversibly with nickel by means of the following reaction x Ni + HE NiHx where x =0.7-0.8.This is a typical example of a solid-state reaction.The formation of fl-NiH is preceded by the formation of tr-Ni, tr-Ni is a solid solution of about 5 at.%H in Ni and its lattice constant is equal to that of Ni (Pielaszek 1985).Above this mean content, there is an abrupt formation of fl-NiH, a-Ni and fl-NiH possess a FCC lattice: the lattice constant of fl-NiH is 0.3738 nmmit is about 6% greater than the lattice constant of Ni, which is 0.3524 nm (Janko 1960,  Pielaszek 1985).fl-NiH, formed within the surface layer of the textured electrodeposited Ni, has the same crystallite orientation as the Ni matrix (Majchrzak, Jarmolowicz 1964).
The kinetic analysis of the decomposition of fl-NiH was done with the help of volumetric (Baranowski 1959) and X-ray diffraction (Borbe et al. 1980, Pielaszek  1985, Janko 1962, Pielaszek 1972, Rashkov et al. 1982) methods.Most of these indicated the fl-NiH decomposition is expressed, as a rule, by a first-order reaction (Borbe et al. 1980, Pielaszek 1985, Baranowski 1959, Janko 1962,  I. TOMOV Pielaszek 1972), but there is also a case presented (Rashkov et al. 1982) in which the fl-NiH decomposition is described by another function.These results (Borbe   et al. 1980, Pielaszek 1985; Baranowski 1959, Janko 1962, Pielaszek 1972) are irreconcilable with the regularities in the kinetics which have been observed in the formation and decomposition of metal hydrides (Larsen, Livesay 1980, Douglas,  Northwood 1983).
In principle, the real structure determines the solid-state transformation rate by its influence both on the chemical decomposition rate (by means of the various defects of the crystal lattice), and on the rate with which the gaseous product reaches or leaves the reaction zone, i.e. the diffusion rate (Delmon 1969).If a   property is dependent on the structure, it should then be an anisotropic property, since "the structure anisotropy is a source of anisotropy of the properties" (Bunge 1988).It is known that the diffusion rate is to a great degree dependent on the crystallographic direction (Meyer 1968).
There have been no studies as yet on the effect of the microstructure (crystal size and microdeformations) on fl-NiH decomposition.Having in mind these considerations, the purpose of our study was to acquire information on the influence of microstructure anisotropy on the fl-NiH formation and decomposi- tion in the various crystal directions.To solve this problem we have employed a suitable X-ray diffraction method.With its help the respective experimental data of previous investigations were processed (Tomov et al. to be published).

SPECIMEN PREPARATION AND EXPERIMENTAL
Most of the models for this study were nickel coatings, deposited on 1 x 25 x 35 mm copper substrates.In an X-ray diffraction sense, the 50/m thick coatings behaved as bulk metal, since their thickness was greater than the effective depth of X-ray penetration (Tomov 1986a).The conditions of electrodeposition were selected in such a way, so as to obtain coatings with differing microstructure and texture.To this end we prepared matt, half-bright, and bright coatings with different surface morphologies.The nickel layers were not detached from their substrates.All specimens given in the tables are representative of 3-4 other specimens.
The cathodic charging of Ni with H was carried out at room temperature in an electrolysis bath consisting of: In H2SO4 and H2SeO3-10mg/l.The duration of cathodic charging was 30 min at a current density of Dk 1 A. dm--.We charged specimens which were in the as-fabricated state as well as such which were heat treated.The annealed Ni coatings were H-charged after electrolytic polishing.
The polishing was carried out in a 60% H2SO4 aqueous solution at about 35C and 40 A. dm -2 for 20 s.

EXPERIMENTAL METHODS
The approach to assessing the quantity of fl-NiH phase was determined by the specific features of a solid-state reaction.In principle, it starts at the surface and its occurrence is accompanied by the progress of the reaction interface in the specimen bulk (Delmon 1969).This results in the formation of a hydride layer over the matrix metal.Therefore the problem of assessing the extent of formation or decomposition of nickel hydride is linked to determining the thickness of this layer.
Expressing the Thickness of the Hydride Layer which has Formed Let us assume that the layer is of a homogeneous composition and that its thickness t is less than the effective depth of X-ray penetration.Then, in the case of a diffractometer, t in any/-direction is expressed by the equation of Friedman and Birks, 1946.
2NiH I where 0 is the Bragg angle, #NiH is the linear absorption coefficient of the hydride, I and I1 are the integrated intensities of the same diffraction line of Ni before hydrogen charging and at the moment of cessation of H charging (Figure la), respectively.Since the/-oriented fl-NiH crystallites are formed of/-oriented Ni crystallites (Majchrzak, Jarmolowicz 1964), t accounts only for the phase transformation of crystallites with ideal /-orientation, i.e. t is a measure of the extent of phase transformation (EPT) of Ni into fl-NiH in the respective/-crystal direction.
Expressing the Thickness of the Decomposed Hydride The decomposition process which starts at the specimen surface, results in the formation of a "sandwich" configuration: tr-Ni (superficial layer)--> fl-NiH---> tr- Ni (matrix) (Figure lb).Let us denote the thickness of decomposed hydride layer by td.Its lattice contraction brings about the generation of the tr-Ni superficial CI.
after H-charging at during decomposition of 13-NiH at : llt'} + I1:12 Figure I Schematic presentation of the formation (a) and decomposition (b) of a -NiH layer and the respective change in diffraction conditions; t-fl-NiH layer thickness; td and t'--thicknesses of the decomposed fl-NiH layer and the a-Ni layer which has been generated from it, respectively.
Diffracted integrated intensities, immediately after H-charging (r 0) 11, and in any other moment r r before complete decomposition 12.
I. TOMOV layer with a thickness of t'.The smaller thickness of t', compared to td, is due to the difference in the lattice parameters of t-Ni and fl-NiH.The layer with a thickness of t' causes diffraction contributions to the nickel lines, which we denote by Itc).Evidently the total diffracted intensity I2 is caused by contributions of the t-Ni substrate and a-Ni superficial layer.
Itt,) can then be expressed by the difference of I2 and the I1 intensity, measured after the cessation of H-charging: I,,) I2-I1 (2) It is well known (Cullity 1967), that the intensity of a t' thick superficial layer is linked to the intensity I, diffracted from an infinitely thick specimen through After substituting Eq. (2) in Eq. ( 4), we obtained sin 0

-(-
Since absorption is a property of the atom, the superficial layers with thicknesses of td and t' absorb X-rays approximately equally, and we may thus write exp( 2/zit'=exp( 2/Zin_.td]sin0/ sin0 /" In this case we ignore only the X-ray absorption of H evolved as the result of the decomposition process.It follows from Eq. ( 6) that /./Ni t'= It/NiH/d (7) After the substitution of Eq. ( 7) in Eq. ( 5), we obtain for the thickness of the decomposed fl-NiH layer sin 0 I td 2/H In I (I I) During the decomposition process 12 changes within the limits I<t<_ After the complete decomposition 12 I and then Eq. ( 8) becomes equivalent to Eq. ( 1).This is a simple way to check the reliability of Eq. ( 8).
The parameters t and td which are linked with the /-direction and thus additionally marked by an further on--they become t and td.Then the fraction of decomposed hydride F/ is defined as the ratio of the thickness of the decomposed layer td to the initial thickness t, i.e.

td F, (9) t
There are two reasons which determine the use of the above expressions (Eq. (1) and Eq. ( 8)) for the correct investigation of the anisotropy of the phase transition: (1) the orientation distribution of the two phases is analogous, since the "growth" of the fl-NiH crystallites during the H-charging is a result of the isotropic increase of the lattice parameter of te-Ni crystallites, as follows from results by Majchrzak and Jarmolowicz 1964, and   (2) the linear absorption coefficient of Ni and NiH are very close.The effective crystal size Di and the microdeformations ei were determined by the single line method (Langford 1978, Delhez et al. 1982)   /' ffc cos 0 (10) ei 4 tan 0 (11) where , is the wavelength, 0 is the Bragg angle, ffc and if6 respectively are the Cauchy (C) and Gaussian (G) components of the integral breadths of the pure diffraction profile (f).
The measurements of the 111,200, 220 and 311 diffraction lines of cr-Ni were performed with an X-ray diffractometer Philips with CuK,, radiation (LiF focusing monochromator).The specimen texture was determined from the poles figures, measured with a texture goniometer.The effective size and the microdeformations of the crystallites in the same directions were determined prior to the cathodic H-charging.To this end we use CoK e radiation isolated with the help of a LiF focusing monochromator.
The study was carried out at room temperature.The time for measuring the integrated intensity (area) for each line was 100 s, i.e. it is negligible compared to the experimentally established time of decomposition of fl-NiH.This leads to the fact that the ti values are not influenced by effects of the measurement technique.

FORMATION OF fl-NiH
EPT of Nickel into Nickel Hydride If the t values are equal in all/-directions of a given specimen, then we have a case of homogeneous phase transformation.In reality, however, the 6 data given in Table 1 indicate that the phase transformation has occurred to a different degree in the ( 111), ( 100), ( 110), ( 311) directions in all specimens, i.e. there is an anisotropy of EPT of Ni into fl-NiH.Evidently 6 is considerably greater in the (111) and (100) directions than in the (311) and (110) directions."There are two clearly manifested trends of this nonhomogeneous phase transformation.The t Because of the presence of a texture it was not possible to measure the 220 lines of te-Ni in all cases, which is the reason for the limited number of tl0 determinations.EPT anisotropy of the matt nickel layers (not counting only specimen No. 2 with a (210) texture) can be presented most generally by the inequalities: tll > tl00 > t311 > tllo (12) while for the bright, half-bright and matt (with a (210) main texture component) nickel coatings the following inequalities are valid.
tl0o > tll > t311 t11o (13) The circumstance that the anisotropy of the phase transformation is described by two inequalities (( 12) and ( 13)) gives us grounds to assert that EPT depends both on the crystal direction and the microstructure anisotropy.The main problem is how to distinguish the influence of the different structure parameters on EPT.This requires the next step of our study to be linked to obtaining information on the effective crystal size and microdeformations.
Effect of Microstructure Anisotropy on EPT The microstructure parameters were determined prior to the specimen charging with H. Table 2 lists the values of the effective crystal size Di (upper lines) and microdeformations ei (lower lines).Di of all matt coatings (specimens Nos.1-4) in all crystal directions are considerably greater than the Di of the bright coatings (specimen No. 6).Moreover, these results indicate that Di is anisotropic.A specific characteristic of the grain size anisotropy is that Di is largest in the direction of the main texture component (Surnev, Tomov 1989).This is confirmed in this case too: the effective crystal size is greatest in the (100), (110), ( 100) and (111) directions for specimens Nos. 1, 4, 5, 6, respectively." The results indicate the specificity of the grain size anisotropy for each concrete growth texture, and that it does not correlate with the anisotropy of f It was impossible to measure the effective crystal size Di in the (210) and ( 211) directions with CoKt-radiation because of technical considerations.
Table 2 Effective size Di (upper lines) and microdeforma- tions ei (lower lines) of the crystallites in some (hkl) directions before cathodically charging with hydrogen of the Ni electrodeposited layers.X-ray diffraction measurements with CoKt-radiation.LiF focusing monochromator. No.
Di + AD [nm}--upper  phase transformation as expressed by the two series of inequalities ( 12) and ( 13).
But it is also evident that EPT falls continuously with the increase in Di, i.e.
from bright to matt coatings (compare Table 1 with Table 2).
The papers (Prelaszek 1985, 1972) discuss the role of dislocations in fl-NiH formation.We shall on the basis of our experimental results attempt to assess the role of microdeformations on the EPT.By origin they are linked mainly with the tension fields around dislocations in the crystal interior (Bever et al. 1973), which is the reason the microdeformations to be used as a measure for assessing the dislocation density (Williamson and Smollman 1956).The dislocation density in electrodeposited metals (Hinton et al. 1963; Hofer et al. 1965) is commensurate with that of cold-worked metals (Read, Shokley 1952), for which it is 10 to 10 2 dislocations per cm2.The values which we have obtained for the dislocation density from Di and ei data (see Table 2), according to Williamson and Smollman 1956, are within the limits 8x 101 to 2x 10 2 dislocations per cm2.We established the highest dislocation density for bright coatings (specimen No. 6), and the lowest--for matt coatings (specimen No. 1).Because of the existing analytical connection between the microdeformations and dislocation density, the two terms will be used interchangeably.
The microdeformation values increase in the order from matt to bright coatings.Within the set of matt coatings this order depends in addition on the texture.The ei distribution has an anisotropic character.Thus for most of the specimens the results reveal a trend for relatively higher microdeformations in the (100) direction compared to the other crystal directions.This trend is especially clearly expressed for specimens Nos. 2, 5, 6 (Table 2) for which e are very high.In fact, inequalities (13) are valid for specimens Nos. 2, 5, 6.These facts undoubtedly confirm that the dislocation-induced anisotropy causes anisotropy of EPT and that the dislocations strongly favour the phase transformation of Ni into fl-NiH.They are in agreement with the assumption made by Atreus et al. 1980, Latanision 1983 that the dislocations act as H traps and in thus support the model for the role of dislocations for the formation of fl-NiH (Pielaszek 1985, 1972).
The question arises what changes would occur with anisotropy of phase transformation, if the dislocation-induced anisotropy is eliminated.This deter- mined the next step in our study to be to investigate the influence of the crystal structure obtained as a result of annealing processes (recovery, recrystallization, grain growth) on the EPT of Ni into fl-NiH.
EPT of Ni into fl-NiH in Thermally Treated Specimens The thermal treatment of the above specimens was carried out at 520C in an argon atmosphere (purified of 02) for five hours.The data on the recrystallization texture obtained from the analysis of the pole figures are presented in Table 3, column (uvw).In fact, the specimens now possess a new structure, and this is reflected in their notation by the addition of a '-mark.In all cases was observed the (100) recrystallization component, but for specimens Nos. 2, 4 were also observed the ( 110) and ( 210) main components of recrystallization texture, respectively.As a result of the recrystallization some crystallites had grown considerably, the consequence of which was that they had their own detectable reflections on the pole figures.According to Hirsch 1956, it may be considered as established that the annealed metals contain 10 6 to 10 8 dislocations per cm2, which means that the dislocation density of our specimens is 3-4 orders of magnitude lower than that of their nonannealed state.This determines a reduced permeation of H in Ni crystallites as has been discussed for an analogous case by Paatsch 1988.As a rule the annealing process leads to the onset of a similarity in the texture, crystal size, and dislocation density.This similarity reflects on a common for all specimens decrease in ti compared to its values for the nonannealed state.Thus stands out the sharp decrease in t0o for specimens Nos. 2, 5, 6 which in some cases is 2-4 times less than t0o for the nonannealed state of the same specimens.By means of the experiments with annealed specimens for which the transport of H toward the interior is hindered, because of the lowered dislocation density and diminished grain boundaries area (because of the increased crystallite size), the promoting role of these two structure parameters on EPT is again revealed.
Talfle 3 Thickness ti of nickel hydride (fl-NiH) in some (hkl) crystal directions of cathodically charged with hydrogen nickel coatings after heat treatment.
At---error due to counting statistics.(uvw)--indices of recrystallization texture.The EPT anisotropy in the investigated crystal directions is described by the inequalities (12) for all thermally treated specimens.Therefore after annealing, that is after eliminating the influence of the dislocation-induced anisotropy, it becomes evident that the main factor which affects EPT is the crystallographic direction.This is one of the most important results of our study.
DECOMPOSITION KINETICS OF fl-NiH AT ROOM TEMPERATURE It is well known that the kinetic behaviour exhibited by hydriding and dehydriding processes (Larsen et al. 1980, Douglas et al. 1983, Rudman 1979) is similar to that of many of the phase transformations of the nucleation and growth type, i.e. the summary transformation process is described by the equation of Kolmogorov (1937), Avrami (1939), Johnson and Mehl (1939)   F/() 1 exp (14 where F() is the -NiH volume fraction corresponding to the/-direction, k is a number, the value of which is dependent on the geometry linked with the rate-controlling process, is the reaction rate time constant.The last has a clear physical sense--this is the time i, for which the decomposed volume fraction F(ri) has a constant value.
e-1 F/(i) e as follows from ( 14).The values of the parameters zi and k were determined by computer fitting of the F/() relations ( 14) to the experimental data.The values of the k exponent are within the interval 1.5-4, the lower values pertaining to recrystallized large crystalline specimens and the greater values referring to unannealed fine crystalline such.
An illustration for the decomposition kinetics of an unannealed half-bright coating is given by Figure 2. Since the decomposition process proceeds in the two investigated by us (111) and (100) directions at the same rate, Figure 2 shows the kinetic curve only for the (100) direction.An example for a decomposition in the same directions of NiH formed in the recrystallized specimen No. 3' (obtained after heat treatment of specimen No. 3) is presented in Figure 3.The results we have presented unambiguously show that the decomposition process both for annealed and unannealed specimens are controlled by the Avrami, Johnson-Mehl kinetics.Our results do not confirm the occurrence of a reaction of the first order.It is evident, however, that the decomposition process in the annealed specimen proceeds at a different rate in the ( 111) and (100) directions.
The last fact has been linked to the microstructure anisotropy in the specimen, which has appeared as a result of its H-charging.
A qualitative assessment of the microstructure in the (111) and (100) directions of specimen No. 3' can be made from the profile of the 111 and 200 diffraction lines, measured after the decomposition of fl-NiH (Figure 4).The value of the microdeformations in crystallites with an (111) orientation is about 1.5 10-3, while the microdeformations in the (100)-direction are below the No. 5).F(r)--volume fraction of nickel hydride which has decomposed in the (100) direction.
sensitivity of this measurement method.Pielaszek has also established that there are no isotropic deformations generated during the hydride formation- decomposition cycle (neither are there any deformations in the (100) direction).
This information reveals in a qualitative aspect the effect of the microstructure anisotropy on the anisotropy of the decomposition rate.This phenomenon can, however, be assessed quantitatively for all specimens.
Anisotropy of the fl-NiH Decomposition Rate The fl-NiH decomposition rate can be derived by differentiating Eq. ( 14) For the studied directions of all specimens this rate can be assessed by means of its value Qma at the point of inflection.Qmax is the NiH volume fraction, undergone transformation per unit time, at the moment the maximal process rate is achieved.These values are presented in Table 4.The different Qmax in the various /-crystal directions for a given specimen are evidence that there is an anisotropy in the decomposition rate, which can depend on both crystal direction and Ni microstructure.
To evaluate the effect of microstructure alone on the dehydriding rate Qi(r), Figure 5 compares kinetic data on the (100) crystal direction for specimens with considerably differing microstructures (Nos. 1, 5, 6--see Table 2).The differential kinetic curves Qi(r) were calculated according to Eq. ( 16), in which the parameters ri and k were determined by fitting Eq. ( 14) to the respective experimental data.The life-time of the hydride phase is greatest for specimen No.
6, which has the highest microdeformations and the smallest crystallites (the greatest area of grain boundaries).It is known that the defects and grain boundaries are sites which can act as H traps (Atrens et al. 1980, Latanision et al.  1983).On the basis of diffusion studies, a list is given by Atrens et al. 1980 of the approximate values of the interaction energies between the trap and H atom for different trap types (point, line and planar trapping defects).The binding energy of trapping is lowest for point defects, and highest for dislocations, i.e. the last are the traps with the greatest depth.From such a point of view, the different differential curves in Figure 5 can be explained .by the varying density and efficiency of the traps.This effect reflects the expected microstructure influence (form and size of the crystallites and density of their defects) on the chemical reaction rate (Delmon 1969), which in fact, means on the hydride phase stability.
The intrinsic parameter for characterizing the NiH stability is the reaction time constant ri.It has been defined from the condition which follows from Eq. (15).
The r values characterize the hydride stability at room temperature (Table 4).In the case of unannealed large crystalline specimens Nos. 1 and 3, the r values are   considerably lower than for fine crystalline coatings with great microdeforma- tions, i.e. for the specimens Nos. 5 and 6.

CONCLUSION
The phase transformation of Ni into fl-NiH is anisotropic.Our studies indicate that the anisotropy of EPT in polycrystals, consisting of crystallites with lattice imperfections, is controlled by two factors: (1) crystallographic direction, i.e. orientation of the crystallites versus surface of charging with H.
In the case of polycrystals, consisting of "perfect" crystallites which have grown during the annealing process, the anisotropy of the phase transformation is controlled by one main factor--the crystallographic direction.
The discussion so far indicates that the mean EPT value should depend not only on dislocation density, but also on the texture (through the distribution of the occurring crystal directions) and on the mean crystallite size (through the area of the grain boundaries).
The different EPT observed in the different crystal directions is in essence an indication for the occurrence of an orientation dependent process during the charging of Ni with H.The orientation dependence of EPT of Ni into fl-NiH undoubtedly is a consequence of the differing rates of permeability of H in the relevant crystal directions of a Ni polycrystal.
The kinetic analysis of the fl-NiH dehydriding indicates that the transforma- tion rate is dependent on the structure of the medium where the chemical reaction proceeds.We have to emphasize that the grain boundaries and dislocations may have a two-fold action on the phase transition: as a shortened diffusion paths in the H transport, and as H traps by chemical reaction.
The cases of fl-NiH decomposition which we have discussed do not concern kinetics of a first order reaction, such as has been reported in.Such an order of kinetics have some of the homogeneous reactions whose elementary chemical act is carded out with the same probability in all sites of the reaction system.The reaction rate is then proportional to the concentration of one of the reacting components.This is not the case, however, of a solid-state reaction, which was the subject of our study.

Figure 2
Figure 2 Decomposition kinetics of fl-NiH, formed in electro-deposited half-bright nickel (specimen in the kinetics of decomposition of fl-NiH, recrystallized after thermally treating Ni (specimen No. 3').F(t)--volume fraction of the hydride which has decomposed in the (111)-A and 100)-El directions.

Figure 4
Figure 4 Profile of the diffraction lines 111 and 200 of specimen No. 3', after the decomposition of fl-Nin.

Table 1
Thickness ti of nickel hydride (fl-NiH) in some (hkl) crystal directions of cathodically charged with hydrogen nickel coatings in as-fabricated state.At---error due to counting statistics.(uvw)--indices of growth texture.

Table 4
Maximal decomposition rate Qmax (volume fraction per second) and decomposition rate time constant ri of fl-NiH in the (111) and (100) directions of specimens with different textures.