Electrodeposition of Ternary Fe–W–H Alloys

Electrochemical deposition, structure, and properties of Fe–W alloys are studied. The alloys formed at a low current density are crystalline supersaturated solid solutions which are magnetic. At higher current densities, amorphous nonmagnetic alloys of the same composition are formed. As a result of treatment at 500–600°С, both amorphous and crystalline transform into a more equilibrium binary system, containing W solid solution in Fe and an intermetallic compound. The concentration of the remaining W in the solid solution was much higher than the equilibrium solubility. A considerable amount of hydrogen incorporates into the deposits. The deposits may be qualified as ternary Fe–W–H alloys. The incorporated hydrogen plays a crucial role in the transition from a crystalline state typical for specimens produced at low current densities to the amorphous or nanocrystalline deposits formed at higher current densities.


INTRODUCTION
Codeposition of tungsten, molybdenum, rhenium, etc. with ferrous metals has been known as "induced deposition" [1] for over 70 years, but recently the interest in these processes has been restored [2,3] owing to promising properties of the obtained materials. Though the mechanism of induced deposition is not yet clear [4], a number of alloys are formed by this way [5,6].
The materials of that kind are often nanocrystalline [7,8], a small size of crystals being one of the reasons of high corrosion resistance and hardness (stable up to 500°C and higher). The alloys Ni-W and Co-Mo are of interest as cathodic materials for water electrolysis because of low overpotential of hydrogen evolution at this surface.
The electrodeposition of tungsten alloys with iron, nickel, and cobalt was the subject matter of a number of publications [9][10][11][12][13][14][15], the majority of which were concerned with nickel-tungsten and cobalt-tungsten deposits. The deposition of an iron-tungsten alloy has been investigated to a far lesser extent although the deposits of this alloy can be obtained with a high current efficiency, high W content, and remarkable hardness. At recent works [14,15] very hard amorphous or nanocrystallinbe Fe-W deposits containing up to 25% W were obtained with subgrains size less than 10 nm.
The deposits of tungsten alloys with metals of iron group containing up to 50 wt % of W are deposited from the electrolytes in which W presents in the form of tungstenate ion. The deposits are very hard and have high internal stresses; they are brittle but wearresistant; they have also high corrosion resistance in various media [16]. The structure, properties, and composition of electrochemically reduced alloys, along with their current efficiency, are heavily dependent on the deposition conditions. As no such data are available for Fe-W alloys, the aim of this work was to create a stably working electrolyte for depositing a Fe-W alloy and examine its structure, composition, and properties as functions of the current density, solution composition, and other conditions. It happened to be possible to obtain both crystalline and amorphous deposits of the alloy by using the same solution at different current densities (amorphous or nanocrystalline objects can be deposited at a higher current density).
The obtained alloys were supersaturated solid iron-tungsten solutions. Some deposits simultaneously contained both crystalline and amorphous phases of an approximately identical composition. Thus, following an increase in the current density, an instant comes when the deposited alloy turns amorphous practically without a change in the relative content of components. Both metals have the same type (bcc) of the crystal lattice, that is why the formation of a solid solution could be expected. However, the equilibrium solubility of tungsten in solid iron amounts to 13 at % only at a temperature of 1532°C. Upon reducing the temperature, the solubility drops down to less then 1% at 700°C.
Thus, electrochemical deposition leads to the formation of a supersaturated solid solution, which exceeds the solubility near the melting point. Factors responsible for this phenomenon may be several. Formation of amorphous deposits requires a sufficiently high metal deposition rate to be maintained. On the one hand, amorphous alloys are formed at a very high rate of deposition from vapor or from melts. But on the other, amorphization may be due to the presence of considerable nonmetallic inclusions. It is known that, with the nonmetallic content in excess of 15%, amorphous phases are commonly observed. In can be assumed that the role of the amorphizer in the case under consideration is played by the co-deposited hydrogen which forms local micro-inclusions maybe of the hydride. As was pointed out in [17], amorphous alloys are frequently formed when one of the components has a high affinity to hydrogen (in our case this is iron) and the second one has no such property (tungsten). As was established in [18], the hydrogen content in amorphous alloys was much higher than that in crystalline ones.
Therefore, along with direct measurements of the hydrogen content in the coatings, it was important to find factors influencing hydrogen evolution. As was found, one of the possible ways to exert such an influence is change in the cationic composition of the electrolyte. Most of the results presented in [19] were obtained while using citrate-ammoniate electrolytes containing ammonium and sodium cations. The cations of alkali metals may influence both the cathodic process, the structure and properties of the obtained deposits [20]. In this connection it was challenging to study the role of cations of alkali metals and ammonium in the overall process and its impact on the current densities corresponding to amorphous, nanocrystalline, and coarse-crystalline deposits. To this end, a chemical analysis of the deposits, the determination of hydrogen in them, and an X-ray diffraction (XRD) analysis of deposits were performed.

EXPERIMENTAL
The alloys were deposited from the electrolytes containing 0.22 M (or less) tungstate, 0.14 to 0.17 M iron(III) sulfate (8-10 g L -1 per Fe ions), 0.42 M ammonium and/or alkali metals citrates. The pH value was adjusted to 7.0 or 7.2 by adding respective hydroxides. A temperature of 50°C was chosen as corresponding to the highest current efficiency. Anodes were of stainless steel.
The samples were deposited potentiostatically in a range from -0.80 to -1.20 V (standard hydrogen electrode), or galvanostatically at the corresponding current densities from 5 to 100 mA/cm 2 . The deposition duration was determined in each particular case on the base of a necessity to get a sufficient mass of a substance for an analysis as well as a sufficiently high volume of hydrogen. The content of tungsten in the alloy, the concentration of ions Fe(III), and the overall concentration of iron in solution were determined by a chemical analysis [21]. The cell was furnished with a burette to collect cathodic hydrogen, with separated cathodic and anodic compartments [22]. The overall volume of collected hydrogen in each case was 4.5 to 7 mL. The alloys were deposited on both sides of copper or brass sheets of 0.75 × 1.0 cm.
The cathodic processes were studied by means of decomposition of overall potentiostatic polarization curves into partial ones using the data of the alloy and solution analysis and the measurements of the amount of evolved hydrogen.
Values of potentials are given vs. standard hydrogen electrode.
The XRD patterns of the deposited samples were obtained with a diffractometer using CuK α radiation.
Hydrogen incorporated into the samples during deposition was determined by the method of vacuum extraction at 500°C.
The magnetic and electric properties of the alloys were measured by usual procedures [23].

Compositions of Solutions and of Alloys
The solutions prepared with Fe(II) sulfate did not contain Fe(III), but after the addition of sodium citrate, the concentration of Fe(III) increased and amounted to 0.05 M. Then, after the exposure at room temperature during 24 h, the content of Fe(III) raised to approximately 0.08 M resulting from the oxidation of Fe(II) by air oxygen. The deposits from this solution, however, were of poor quality and unsatisfactory reproducibility of results was observed. But as a result of the pre-electrolysis of the solution at 20 mA/cm 2 for 2 h, the content of Fe(III) reached 0.12 M, which was more than 70% of the overall iron concentration. After that the solution was allowed to obtain bright and hard deposits of the alloy at a high reproducibility (with regard to the rate of deposition and the content of tungsten in the alloy). Transferring Fe(II) into Fe(III) ensures the obtaining of results that are practically the same with respect to the rate of the deposition and the composition of the iron-tungsten alloy, as in the case of the electrolyte with the initial insertion of no less than 0.12 M of compounds of Fe(III) into solution. This allows to claim that the direct discharge of iron into the alloy in the working conditions occurs from Fe(III) ions, rather than Fe(II). This is the distinctive feature of the process of deposition of alloys of tungsten with iron. So at calculations of the current efficiencies we assumed that the deposition of W is the six-electron process and the deposition of Fe needs three electrons per an atom.
The overall measurements had shown that the number of parallel processes that occur on the cathode is equal to four. These are the deposition of tungsten, the deposition of metallic iron, the evolution of hydro-SURFACE ENGINEERING AND APPLIED ELECTROCHEMISTRY Vol. 55 No. 4 2019 GAMBURG, ZAKHAROV gen, and the incomplete reduction of Fe(III). The partial curves for these processes along with the overall curve E-i are shown at Fig. 1. No processes of partial W(VI) reduction were in evidence here. In a potential range from -0.83 to -0.91 V, the rate of the latter reaction Fe(III)/Fe(II) exceeded the rates of all other processes, but, at a further increase in the cathodic overpotential, the rate of this reaction alters inconsiderably, whereas other processes accelerate. In a potential range from -1.2 to -1.45 V (the corresponding current densities from 20 to 100 mA cm -2 ), the rate of the overall process is defined by the current density of the deposition of the alloy. It is precisely the region of potentials where one can obtain deposits of the alloy of a satisfactory quality with a maximum value of the current efficiency (up to 55%) and a constant content of tungsten (23, up to 25 at %). With further displacement into the region of negative potentials, the partial deposition rate of the alloy changed slightly, which is due to the presence of diffusion inhibition of both processes. For tungsten, even a small decrease in the discharge rate into the alloy was observed. This can result from the fact that the tungstate ion has a lower diffusion coefficient compared to Fe 3+ (in addition, a negative charge, which impedes its movement to the cathode). The deposits formed in this interval of potentials were non-uniform, with local formation of areas of black color in places with an elevated current density (along the edges).
The deposits obtained at a current density of 20 to 70 mA/cm 2 were bright, hard (microhardness 11-12 GPa), and well adhered to the brass substrate.  (Fig. 2); increasing or decreasing this concentration diminished it, therefore, for a further study, this particular concentration was chosen. The current efficiency somewhat increased at more high W content in the deposits and was highest at 50°C and 30 mA cm -2 . The data upon the dependence of {W} A on {W} S are presented at Table 1.
Comparing the overall current density vs. efficiency dependences, obtained in a potentiodynamic mode in the solution containing 0.22 M [W] S immediately after its preparation and after its aging, shows that the aging barely affects the process kinetics. The partial curves obtained for the hydrogen evolution, iron deposition, and tungsten deposition ( Fig. 1) all are characterized by the Tafel dependences with a common slope of about b = 2.3RT/αnF = 0.11 V per decade. For iron, this probably indicates a two-stage process limited by the first-electron transfer with an value of α being about 0.58 (temperature T at the measuring time was 323 K). As for tungsten, the obtained data point to no definite mechanism of the process. It is evident, nevertheless, that the processes of reduction of both Fe and W are coupled. The constancy of the alloy composition in a wide range of current densities is in favor of this hypothesis. It can be supposed   The said dependences and production of deposits with a definite composition and structure are typical only for a solution without agitation. Even at a slight agitation, current efficiency reduced sharply.

Role of Ionic Composition of Solutions
The rates of the four partial processes depended on the ionic composition of the electrolyte. The effect of cations on cathodic processes is significant. At potentials more positive than -1.05 V, the rates of all respective reactions insignificantly differ from each other. With increasing cathodic polarization, an increase is observed in the overall rate of the alloy deposition in the lithium electrolyte as compared with that of the sodium electrolyte. At the same time, the formation of Fe(II) in the sodium electrolyte proceeds more intensively as compared with that in the lithium solution. At all said potentials black and crumbling alloys had deposited.
The most important is the effect of cations of alkali metals and ammonium on the current efficiency and the tungsten content in the alloys. In the series NH > K + > Na + > Li + , the working region of current densities corresponding to satisfactory deposits become narrow. In the ammonium solution, this region extended from 10 to 50 mA/cm 2 , whereas, in the lithium electrolyte, it is impossible to obtain dense compact coatings of the alloy at all. The ammonium electrolyte is characterized by the constancy of both cur-4 + rent efficiency (45%), and the tungsten content in the alloy practically throughout the entire investigated range (23 at %). In electrolytes with other cations, depending on the current density, current efficiency passes through a maximum at 20-30 mA/cm 2 . As the current density increases, it sharply decreases (Fig. 3), and the tungsten content in the alloy decreases to less than 10 at %. The presence of potassium ions in the electrolyte leads to an increase in the fraction of partial reactions of hydrogen evolution and the formation of Fe(II), while in the sodium-containing electrolyte the ratio of the rates of all cathodic processes varies little over a wide range of potentials. These effects are probably associated with different degrees of hydration of ions, possibly with their association, and depend on the size of the ions. As will be described in more detail below, the current density at which the transition from crystalline to amorphous precipitates occurred depended on the anionic composition of the solution.

Structure of Deposits
According to XRD, samples deposited at current densities below 15 mA/cm 2 were usually crystalline, whereas, those produced at CD higher than 20 mA/cm 2 were mostly amorphous. The crystalline ones have a body-centered cubic crystal lattice typical for both iron and tungsten. The lattice period was 0.294 nm, which is intermediate between that of iron (0.2866 nm) and tungsten (0.3165 nm). The samples deposited at 20 mA/cm 2 had both crystalline and amorphous phases (Fig. 4), and those at 50 mA/cm 2 were completely amorphous. Assuming that Vegard's law is true for this system, it follows that [W] A = 24.7 at %, which corresponds to the data of chemical analysis. Thus, the crystalline Fe-W alloy is single-phased and represents a supersaturated solid solution of W in Fe (the equilibrium solubility of W in Fe is 4.5 wt % at 700°C).
The tungsten solubility in iron in the case of crystalline deposits happened to be much higher than the equilibrium value. This is due to a high overpotential of the process, which ensures an energy required for a corresponding distortion of the crystal lattice. During heat treatment of both amorphous and crystalline precipitates (up to 800°C in vacuum), a phase transition occurred; reflections from the saturated solid solution W in Fe and from the intermetallic compound Fe 2 W appeared on the powder X-ray diffraction patterns (Fig. 5). The crystalline deposits decomposed into a solid solution (with a lower concentration of tungsten as compared with the original) and intermetallic compounds at temperatures of about 600°C, while the amorphous ones began to decompose at a temperature of about 500°C.
The concentration of the solid solution was relatively low. Hence, applying Vegard's law, from the XRD data, the equilibrium solubility of W in Fe as 6 at % at the temperature of the transition was estimated. This value is substantially higher than that known from that published elsewhere for this temperature (1.4 at %) [24]. An analysis offers three explanations. Either this system fails to satisfy Vegard's law it the corresponding concentration range, or the formed solid solution remains oversaturated, or the literature data on the solubility of W in Fe are underrated. The most interesting result obtained in the study is, most probably, an opportunity to obtain both magnetic and nonmagnetic alloys from the same solution, at different current densities. The crystalline and amorphous deposits were ferromagnetic and nonmagnetic, respectively. The same phases crystallized from the amorphous deposit at temperatures exceeding 500°C. Annealing drastically altered magnetic properties as well. The Curie temperature of the alloys was nearly 720°C. Amorphous deposits during annealing converted into a crystalline shape before the Curie temperature was reached. Because of this, the deposit first acquired some magnetization, which passed through a maximum at 650°C, and then decayed to zero at the Curie temperature (Fig. 6, curve 1). When cooled, the alloy again behaved as a magnetic material, but the value is twice lower the starting value. Upon annealing amorphous precipitates (curve 2), the magnetization arose at 500°C (as a result of the decomposition of the amorphous phase to form Fe 2 W), passed a maximum at 650°C, and then decreased to zero at the Curie temperature, as in the case of crystalline precipitate. Upon reverse cooling, the amorphous alloy behaved similarly to the crystalline one.
The electric resistivity of the samples was no less than 0.3 mΩ cm and increased with current density. This value corresponds to the alloys with the highest resistivity, which confirms the conclusion on the heavy structural disordering of the deposits, especially at high current densities.
The crystalline alloys were obtained usually below 15 mA/cm 2 in the solutions of various compositions. In the solutions containing lithium, crystal deposits formed at the current density range studied. The similar deposits were analyzed for the hydrogen content by the method of vacuum extraction. Table 2 shows the dimensions of crystals D in the alloys (calculated by the corrected Scherer formula) along with the measured hydrogen contents in similar deposits.
These data exhibit the determining role of hydrogen in the formation of amorphous deposits. In all cases hydrogen content in the deposits of about 2.5 cm 3 /g was found above which the formation of crystalline coatings was not exhibited (In the case of lithium-containing solutions, the amount of hydrogen had never exceed 3.3 cm 3 /g). The critical concentration of hydrogen depends slightly on the solution composition and deposition conditions and in some cases it can be higher than 3 cm 3 /g (2-3 at %). A correlation exists between the type of the deposit and the current efficiency of hydrogen reaction (Fig. 2): at current efficiency higher than 5% more than 90% alloys were crystalline. The incorporated hydrogen was only a fraction (less than 10%) of the total amount of evolved hydrogen.  Thus, the transition from the crystalline phase growth to the formation of amorphous deposit occurs at a certain content of hydrogen reached at various current densities, potentials, etc. depending on the electrolyte composition, its temperature, and on other conditions. The maximum amount of hydrogen in the crystalline deposits is approximately 2-3 cm 3 /g.
Thus, the transition from the growth of the crystalline phase to the formation of an amorphous precipitate occurs at a certain hydrogen content, which can be achieved at different current densities, potentials, etc., depending on the composition of the electrolyte, its temperature and other conditions. In mixed and ammonium electrolytes, crystalline precipitates form at low current density (up to 15-20 mA/cm 2 ), and potassium ions increase this range of current density to 30 mA/cm 2 . The reaction rate of hydrogen evolution in solution with potassium is lower, and therefore amorphization of precipitation occurs at higher current densities and potentials, when hydrogen evolution accelerates. In solutions containing lithium, crystalline precipitates were formed over the entire studied range of current densities. Thus, neither a high deposition rate of the alloy, nor a sufficiently negative deposition potential of the alloy are probably sufficient conditions for the transition from coarse to nanocrystalline and amorphous depositd. A more important role is played by hydrogen, which is included in the crystal lattice.
The following problem is worth discussing: an increased content of hydrogen is the reason for amorphous deposit formation or, conversely, the amount of incorporated hydrogen increases after the transition to the amorphous phase had happened. It is supposed that the effects are interrelated and, as a result, the said transition proceeds by a chain mechanism: a degree of hydrogenation gradually increases at higher current densities; this corresponds to the following rather general considerations.
Assume that the surface coverage of a deposited alloy with hydrogen atoms depends on the current density and is θ(i). The flux of hydrogen incorporation into the metal lattice is equal to the sum of exchange flux J 1 H ads ↔ H abs existing also in the absence of metal deposition (with the rate constants k + and k -), and an additional flux of hydrogen adds atoms capture J 2 = k inc iθ, i being a partial current density of metal deposition and k inc is the dimensionless coefficient. Since the metal deposition rate is i/nF, the content of hydrogen in the deposit (atomic fraction) is C d = θ(k + + k inc i)/((1 -θ)k -+ i/nF) [23]. Here the amount n is an average number of electrons accounted for one atom in the alloy, i.e. n ≈ 3.7 for the alloys studied (ca. 23 at % of tungsten). This means that the concentration C d differs from the equilibrium one existing at i = 0 and depends on the coverage θ.
In the simplest case, the concentration of hydrogen in the deposit depends on the current density in the same manner as the coverage does. On the other hand, it is known that θ can depend on the current density in different manners. If hydrogen is removed by the mechanism of electrochemical desorption, which is most probable for the system under investigation, a potential shift into the negative direction leads to an increase in the coverage (in the case that the exchange current of discharge stage is higher than that of the stage of electrochemical desorption). If hydrogen is removed by the mechanism of recombination, then a similar effect takes place. An increase in the surface coverage of a steel cathode with hydrogen from approximately 0.2 to 0.6 (with a decrease in the potential from -0.3 to -0.6 V) was observed experimentally, for example, in [25]. Consequently, in those cases, the content of hydrogen in the deposits will increase with increasing total current density, and this was experimentally observed. At a certain current density depending on the cationic composition of electrolyte, its pH value, temperature, and other conditions, when the necessary surface coverage θ with adsorbed hydrogen is reached, the critical value of hydrogen concentration in the deposit will be also reached. At this concentration in the initially formed very thin coating layer, high stresses arise and a considerable number of the crystal lattice defects (steps and kinks on them, screw dislocations, etc.) are formed. These defects serve as the sites for the formation of many incoherent nuclei, which coalesce and form amorphous boundary zone rather than a single grain. As a result of the formation of more friable structure, the rate of hydrogen absorption by the deposit abruptly increases further promoting amorphization.
The above pattern is qualitative, because, at present, the values of most parameters determining the above described processes are unknown. Nevertheless, a more complete pattern of the process remains to be studied.
The correlation of the deposit structure with the current efficiency (by hydrogen) may also confirm the alternative hypothesis: iron hydroxide precipitating in the near-electrode layer (the pH value of electrolyte is slightly higher than that of iron hydrate formation) plays a role in the mechanism of deposit structure formation. Hence the deposits have to contain some oxygen [9] but much less compared to hydrogen. It can be supposed that this hydrogen facilitates preservation of the emerging nonequilibrium structures, hindering crystallization processes as a result of the local formation of inclusions of a hydride (on the base of iron but, more likely than not, mixed-similar to iron-titanium hydride). According to the authors' data, the content of hydrogen in the iron lattice reaches in some cases 14 at % (ignoring tungsten, which is little inclined toward interaction with hydrogen). This is a quantity that is close to typical concentrations of amorphizers. As is known, a similar effect caused the formation of a finer structure in some cases; for example, when nickel is deposited at pH higher than 6.0 and solution buffering is not sufficient.
CONCLUSIONS (1) The process of the electrochemical deposition, structure, and properties of Fe-W alloys are studied. It is established that if the solution is prepared on the base of compounds of Fe(II), then deposition of smooth and compact deposits of the alloy begins only after a larger fraction (at least 2/3) of these compounds converts into Fe(III) as a result of oxidation by air oxygen and anodic oxidation.
(2) Four parallel processes occur on the cathode: the deposition of iron and tungsten, the formation of products of reduction of Fe(III) to Fe(II) and the evolution of hydrogen, partial current density depending in a complicated manner on different conditions.
(3) The content of tungsten in the obtained alloys raises when its concentration in the solution becomes higher reaching the limit at W concentration in solution about 0.22 M. It is weakly dependent on the current density and amounts to nearly 23 at %. The current efficiency is about 40 to 50%.
(4) Cations of alkali metals and ammonium have a substantial influence on the cathodic processes and on the structure of deposits thus obtained. The presence of ammonium ions in solutions leads to a relative constancy of the current efficiency and the tungsten content in the alloy in a wide range of deposition conditions.
(5) Alloys obtained at low current densities were magnetic crystalline oversaturated solid solutions of W in Fe. Higher current densities yield amorphous nonmagnetic alloys of the same composition. The alloys have a very high resistivity (nearly 300 μΩ cm).
(6) After a treatment at 500-700°C, both amorphous and crystalline alloys transform into a more equilibrium binary system comprising a solid solution of W in Fe and an intermetallic compound Fe 2 W. The remainder concentration of W in the solid solution is higher than equilibrium solubility at 700°C.
(7) A high alloy deposition rate and a sufficiently negative potential of its deposition are conditions not sufficient for transition from large-crystal to nanocrystalline and amorphous deposits. Here, a more important role is played by hydrogen codeposited into the crystal lattice.
(8) A considerable amount of hydrogen (up to 10 at %) incorporates into the deposits so that the deposits may be qualified as ternary Fe-W-H alloys. The amount of hydrogen essentially depends on the current density, potential of deposition, and on the ionic composition of electrolyte, primarily on the content of alkali metals and ammonium ions.
(9) Incorporated hydrogen plays a crucial role in the transition from a crystalline alloy state, which is typical of specimens produced at low current densities, to amorphous or nanocrystalline deposits formed at higher current densities. In the presence of ammonium ions, the transition occurs at a much lower current density than in the presence of potassium ions.
(10) The transition from crystalline iron-based solid solutions to the nanocrystalline ones occurs at a hydrogen content exceeding 3 cm 3 (under standard conditions) per 1 g of alloy, or 2.3 at %. The corresponding current density and deposition potential depend on the solution composition and other conditions.
(11) It is assumed that the transition occurs as a chain process: in a thin juvenile deposited layer containing high amount of hydrogen, high stresses arise; as a result, the concentration of defects increases accelerating hydrogen incorporation.