Influence of a magnetic field on the electrodeposition of nickel–iron alloys

Electrochimica Acta 52 (2007) 2785–2795

Influence of a magnetic field on the electrodeposition of nickel–iron alloys A. Ispas, H. Matsushima, W. Plieth, A. Bund ∗ Institute of Electrochemistry and Physical Chemistry, Technical University of Dresden, D-01062 Dresden, Germany Received 9 June 2006; received in revised form 27 October 2006; accepted 30 October 2006 Available online 1 December 2006

Abstract The electrodeposition of nickel–iron alloys is studied under the influence of a superimposed external static magnetic field. It is shown that the direction of the magnetic field with regard to the electric field affects the electrodeposition process (current efficiency, composition and morphology of the layers). Furthermore, the influence of the simultaneous action of natural and magnetically driven convection is discussed in function of the orientation of the working electrode. The electrochemical quartz crystal microbalance (EQCM) technique is used to determine the partial current due to the hydrogen evolution reaction (HER). It is observed that the magnetic field applied perpendicular to the electric field lines increases the rate of the HER, while no significant variation is observed in a magnetic field parallel to the electric field. © 2006 Elsevier Ltd. All rights reserved. Keywords: Ni–Fe electrodeposition; Magnetic field effects; Natural convection; Hydrogen evolution reaction; MHD

1. Introduction

2. Theory

The deposition of Ni–Fe alloys was classified by Brenner [1] to be an anomalous one, which means that the less noble metal (iron) is deposited preferentially compared to the more noble one (nickel). The electrodeposition of Ni–Fe alloys has been intensively studied [1–12] and some models have been developed during the last years to explain this behavior [2–7]. The motivation lies in the special properties of these alloys like high internal strength, hardness and special magnetic properties. Nevertheless there are still open questions concerning the deposition mechanism. One of them is the effect of an externally applied magnetic field during the electrodeposition process. The present paper addresses this question and thus extends the previous work of Chopart and co-workers [13] and Tabakovic et al. [14]. We focus on the magnetic field effects induced during the galvanostatic deposition of Ni–Fe alloys with relatively low iron content. Special regard is given to the hydrogen evolution reaction and the morphology of the deposited layers. Furthermore, the interplay between the natural convection and the superimposed static magnetic field is discussed.

It has been shown that a superimposed magnetic field can be used to increase the mass transport in electrochemical reactions or to change the properties of the electrodeposited layers [15–22]. In order to tailor the magnetic fields for electrochemical reactions (such as nickel–iron plating), it is important to understand the effects of a magnetic field in electrochemical processes. It is generally accepted that the main influence produced by an external magnetic field in electrochemical systems is due to the Lorentz force, fL , that acts on the moving ions, Eq. (1)

∗

Corresponding author. Tel.: +49 351 463 34351; fax: +49 351 463 39820. E-mail address: [email protected] (A. Bund).

0013-4686/$ – see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2006.10.064

 fL = j × B

(1)

where j represents the flux of ions (current density) and B is the external magnetic flux density. When the magnetic field is applied perpendicular to the electric field lines (i.e. electric current), fL is maximal, while in magnetic field parallel to the electric field, E, the Lorentz force should be negligible, but concentration gradient forces, fc , can still be active due to concentration gradients of paramagnetic ions, Eq. (2) [16–18]. χm B2  ∇c f∇c = 2μ0

(2)

2786

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

where χm is the molar magnetic susceptibility, c the concentration, and μ0 is the permeability of vacuum. The force densities fL and fc act on volume elements of the electrolyte and thus induce convection in the electrochemical cell. In most cases this convection increases the transport of electroactive species towards the electrode, and thus an increase of the limiting current density, jlim , occurs. It has been reported that depending on the geometry of the cell and the interplay of the magnetic forces with the natural convection jlim can also decrease [17,22]. The works of Iwakura et al. [23], Aogaki et al. [24,25], Shannon et al. [26], Arumugam et al. [27] show the role of the direction of B with regard to E. In the present work an Electrochemical Quartz Crystal Microbalance (EQCM) with damping monitoring (dissipative EQCM) was used for the in situ characterization of the electrodeposition process. The dissipative EQCM technique is based on an admittance spectrum of the quartz crystal near its resonance frequency. This resonance curve is a Lorentzian centered at the resonance frequency and with a full width at half maximum (FWHM) which is proportional to the damping of the quartz crystal. The mass change at the surface of the quartz crystal can be calculated from the shift of the resonance frequency using the Sauerbrey equation, Eq. (3). f = −

2f02 m Zq

(3)

where f represents the shift in the resonance frequency of the quartz crystal when a rigid mass density, m, is deposited on one side of the resonator, Zq the mechanical impedance of quartz and f0 is the resonance frequency of the unloaded quartz. A shift of 1 Hz in the resonance frequency of a 10 MHz quartz crystal corresponds to a mass change of 4.425 ng cm−2 . The hydrogen evolution reaction (HER) is a key feature for the electrodeposition of Fe group metals and their alloys, since it reduces the current efficiency, can modify the surface pH and the kinetics of metal reduction [9], or the electrocatalytic activity of the alloys [10]. EQCM can be used to detect the formation, growth and detachment of the gas bubbles, since the densities of liquids and gases are very different [28]. It has been shown in the case of Ni [20,29] or Fe deposition [30] that the EQCM is a sensitive technique that can be used to separate the partial current of the metal deposition from the partial current of the HER, jHER . The algorithm used to calculate

jHER from the EQCM data was presented in a previous paper [20]. Shortly, for a galvanostatic deposition, the total imposed current is the sum of the partial currents of the individual electrochemical reactions (HER, Ni, and Fe deposition). In the present study, jHER is obtained as the product of the total current density (jtotal ) and (1 − η), where η is the current efficiency for the alloy deposition. The latter is the ratio between the deposited (measured with the EQCM) and the calculated (from Faraday’s law) mass. The molar mass was calculated according to the composition of the alloy (determined ex situ by energy dispersive X-ray analysis). 3. Experimental The electrolyte contained 0.5 M NiSO4 ·6H2 O, 0.01 M FeSO4 ·7H2 O and 0.4 M H3 BO3 (both p.a. grade) in highly purified water. The pH was adjusted to 3 by adding sulphuric acid. The composition of the electrolyte was chosen based on previous studies [8,11] with special regard to achieve the maximum current efficiency. 10 MHz optically polished AT-cut quartz crystals (KVGNeckarbischofsheim, Germany) were used for the EQCM measurements. Two gold electrodes (100 nm thickness) were deposited by thermal evaporation on the opposite faces of the quartz crystal, on a thin adhesion Cr layer (5–10 nm). The electrodeposition was done in a three electrode cell, home made from Teflon (4 cm × 4 cm × 6 cm). One gold electrode of the quartz crystal was used as the working electrode, the reference electrode was a Saturated Calomel Electrode (SCE, Sensortechnik Meinsberg GmbH, Meinsberg, Germany) and the counter electrode was a platinum foil (area 3.7 cm2 and 0.15 mm thickness). In the following all potential values refer to the SCE. The active area of the working electrode was generally about 0.22 cm2 . After the deposition the exact value of the electrochemically active area was determined with the help of imaging software. This procedure ensured accuracy in the order of 1% for the determination of the current efficiency. The distance between the working and the counter electrode was 2 cm. A water-cooled electromagnet (VEB Polytechnik, Phylatex, Germany) could furnish a homogeneous magnetic field of maximum 740 mT (for a 4 cm gap), that was superimposed to the electrochemical cell (Fig. 1). An Advantest R3753BH network analyzer (Advantest, Tokyo, Japan) was used to record the resonance spectrum of

Fig. 1. Schematic representation of the experimental setup for the EQCM measurements in magnetic fields. WE: working electrode (Au electrode on the quartz crystal), B: direction of the magnetic field; NC: direction of the natural convection (driven by density gradients); MHD: direction of the magnetohydrodynamic convection (driven by the Lorentz force).

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

2787

the quartz crystal from which the frequency and the damping shift were obtained [20,31]. A potentiostat model 263A (EG&G Instruments, Princeton, NJ) was used for the electrochemical experiments. The topography of the deposited layers and the roughness were investigated with a contact mode AFM (Atomic Force Microscopy, PicoSPM, Molecular Imaging, Phoenix, USA), using triangular cantilevers with sharpened pyramidal tip, made of silicon nitride (Olympus Corporation, Tokyo, Japan). Energy Dispersive X-ray analysis (EDX, system Voyager III 3200, SUS Pioneer) yielded information about the alloy composition. Besides AFM, Scanning Electron Microscopy (SEM, model DSM 982 Gemini from Zeiss Oberkochen, Germany) was used to investigate the surface morphology of the alloys. A Siemens D500 X-Ray Diffractometer with Cu K␣ radiation (wavelength ˚ was used for the crystallographic characterization of 1.5405 A) some representative samples in reflection mode. The depositions were performed galvanostatically, with a constant total charge density of 4.5 C cm−2 , using a variety of current densities between −1 and −60 mA cm−2 (the exact values of the current densities are shown in Fig. 3). The thickness of the deposited layer was around 1.5 ␮m. Fifteen minutes before each measurement and during all depositions, the electrolyte was purged with Ar. Three orientations of the working electrode were investigated (Fig. 1). One time the quartz crystal was fixed at the bottom of the cell (in the following called orientation 1), and the other time it was fixed vertical, parallel to the gravitational field (orientations 2 and 3). In orientation 2 the magnetic field could be applied perpendicular to E and in orientation 3 it could be applied parallel to E. 4. Results and discussion 4.1. Cyclic voltammetry Cyclic voltammetric experiments (five cycles, 5 mV s−1 ) were carried out from the open circuit potential to −1.1 V, going first in the cathodic direction. In the potential region from −0.4 to −0.7 V the HER is the dominating reaction (Fig. 2). This fact can be easily verified from the time derivative of the EQCM data which corresponds to the partial current of the alloy deposition (Fig. 2b). The HER increases in the presence of the magnetic field because it is limited by mass transport and the MHD stirring decreases the diffusion layer thickness of the hydrogen ion. At around −0.75 V the alloy deposition starts. EDX analyses for the layers deposited at this potential yield a composition of 8 at.% Fe and 92 at.% Ni. Interestingly the partial current of the alloy deposition decreases in the magnetic field. This inhibition is visible in both the total current (Fig. 2a) and the EQCM data (Fig. 2b) and it occurred for both orientations of the working electrode (parallel to gravity and perpendicular to gravity). The dissolution peak is not affected by the magnetic field. Because of the high Ni concentration one can assume that the Ni reduction is over a wide potential range mixed controlled while the Fe reduction is diffusion controlled. Therefore, one can expect a stronger magnetic field effect in the case of Fe. As

Fig. 2. (a) Typical cyclic voltammogram (fourth cycle, 5 mV s−1 , orientation 1) and (b) time derivative of the EQCM signal in 0.5 M NiSO4 ·6H2 O, 0.01 M FeSO4 ·7H2 O, 0.4 M H3 BO3 (pH 3, 298 K). Black curves: no magnetic field and gray curves: B = 0.7 T.

the iron reduction current increases the surface concentration of intermediate Fe(I) species increases. According to the models of Matlosz and Landolt these Fe(I) species inhibit the Ni reduction and the net effect would be a decrease of the total cur-

Fig. 3. Iron content of the deposited alloy as a function of the total current density (orientation 2, 0 mT). Electrolyte composition as in Fig. 2.

2788

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

Table 1 Iron content in atomic percent of the deposited alloys in function of the direction of the external magnetic field and the total current density applied jtotal (mA cm−2 )

B = 0 T (orientation 1)

B = 0.7 T (orientation 1)

B = 0 T (orientation 2)

B = 0.7 T (orientation 2)

B = 0.6 T (orientation 3)

−25 −60

4.8 ± 0.2 2.8 ± 0.1

4.5 ± 0.4 2.2 ± 0.3

4.1 ± 0.3 2.9 ± 0.6

5.4 ± 0.2 3.1 ± 0.3

4.4 ± 1.2 3.0 ± 0.8

rent density which is in perfect agreement with the experimental data. 4.2. Effect of the total current density on the alloy composition First, the changes in the alloy composition in function of the total applied current density were investigated in absence of an external magnetic field. Between three and eight samples were prepared for a variety of current densities in the range −1 to −60 mA cm−2 (Fig. 3). The data in Fig. 3 represent the averaged values of each set of measurements. The standard error (shown as the error bar in Fig. 3) of the mean value was evaluated from the ratio between the standard deviation and the square root of the number of measurements. In the low current density range (−5 to −10 mA cm−2 ) the Fe/Ni ratio in the alloy (ca. 14:86 in terms of molar ratio for a current density of −10 mA cm−2 ) is higher than in the electrolyte (2:98). At higher current densities (−15 to −40 mA cm−2 ) the Fe/Ni ratio in the alloy decreases significantly (ca. 5:95 for a current density of −25 mA cm−2 ), but it is still higher than in the electrolyte. In passing we note that this is the expected behavior for a system showing anomalous codeposition (Fe is the less noble, Ni the more noble metal). The expected maximum of the Fe content [8,11,32] was not observed in the present study. This is because our current densities are beyond the current density where the maximum occurs. Several explanations have been given for the decrease of the Fe content at high current densities [1,2,4,8,33]. The simplest one is that both Ni and Fe reach their limiting currents and the deposition becomes normal again. For the electrolyte composition and the current densities used in this study the limiting current density of iron is much smaller than the total current density. The partial current of the iron reduction can be obtained from partial current of the alloy deposition (obtained from the EQCM data) and the known dependence of the alloy composition on the current density (Fig. 3). Using these data we find that at higher potentials the partial current of the iron reduction levels off well below 1 mA cm−2 .

increases to 5.4 ± 0.2 at.% (Table 1). A qualitatively similar behavior is observed in the current density range from −25 to −60 mA cm−2 . Looking at the data in Table 1 one might argue that the layers deposited in orientation 1 contained slightly more iron than those prepared in orientation 2 and this “effect” would be more pronounced for the smaller current densities. Notice however, that the accuracy of the EDX data is not sufficient to speculate about a possible mechanism. In the following we try to explain the significant changes of the polarization and the increase of the Fe content in orientation 2. In order to understand the effect of B (and its orientation) on the polarization one must take into consideration the natural convection in the cell. Natural convection is driven by density gradients which appear as a consequence of the electrochemical reaction. For a cathodic deposition metal ions are removed from the electrolyte near the electrode which causes a decrease of the density. Some theoretical studies were done in order to understand the effects of natural convection in electrochemical reactions. The limiting current of Cu deposition and the relative ionic mass coefficients for different electrochemical systems were calculated. The later can be used to determine the buoyancy force [34,35]. At the example of Cu it has been pointed out that the electrode arrangement with regard to the gravitational field can influence the nucleation behavior [36], the size of the crystals [37] and the pattern of the fractal growth [21]. If the anode was placed above the cathode (which corresponds to our orientation 1) more nuclei appeared due to a Rayleigh-B´enard type natural convection [36] and a decrease of the fractal dimensionality was observed [21]. Quraishi calculated the Sherwood number for different Hartmann numbers and different angles between the

4.3. Effect of the magnetic field on the alloy composition The potential transients for jtotal = −25 mA cm−2 are presented in Fig. 4. In the absence of a magnetic field the deposition potential, Edep , is −1.48 ± 0.02 V versus SCE and the Fe content is 4.8 ± 0.2.at%. If B is applied in orientation 1 Edep becomes more cathodic and the Fe content remains virtually the same. If B is applied in orientation 2 the polarization decreases (Edep becomes more anodic) and the Fe content

Fig. 4. Transients of the deposition potential for the orientation 1 (a) and for the orientation 2 (b). The black curves correspond to B = 0 mT and the grey curves stand for B = 740 mT (a) or 715 mT (b). The total current density was jtotal = −25 mA cm−2 .

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

Fig. 5. Normalized partial current density of the hydrogen evolution, jHER , in orientation 2 ( ) and orientation 3 () for a total current density of −60 mA cm−2 as a function of the magnetic flux density. The data have been normalized with respect to the value of jHER in absence of a magnetic field.

cathode and the gravity field when a magnetic field was superimposed [38]. His conclusion was that the interplay between the natural convection and the gravity cannot be described in a simple way. It has been reported [37] that the plating rate of Cu can be suppressed in the presence of a B field parallel to the E field, but promoted by a G field (gravity) parallel to the E field. When the working electrode is arranged parallel to the gravitational field, the mass transfer is expected to be enhanced because of additional convection induced by the density gradients [22,39]. The effect of natural convection on the flow and concentration profiles can be verified by optical methods [40–42]. Unfortunately those methods cannot be applied to our experimental setup (recessed electrode).

2789

In the presence of a B field magnetic forces will be active which induce magnetohydrodynamic convection (vector MHD in Fig. 1). The latter will interact with the flow profile of the natural convection (vector NC in Fig. 1). As a result the transport of electroactive species towards the electrode is modified and in a galvanostatic experiment the potential will shift [21,43]. As a rule of thumb it can be stated that increased mass transport will decrease the polarization (Edep becomes more anodic) and decreased mass transport will increase the polarization (Edep becomes more cathodic). The composition of the alloy will depend on the deposition potential [8,32] and we can expect changes of the composition, too. In orientation 1 the MHD convection is perpendicular to the natural convection and without numerical simulations it is hard to judge if the mass transport towards the electrode increases or decreases. According the potential data the mass transport should decrease because the polarization increases. In orientation 2 the MHD convection is parallel to the natural convection and we can expect an increase of the mass transport which is in perfect agreement with the data in Fig. 4 (polarization decreases when B is on). Actually the force balance should also contain the concentration gradient force, fc (Eq. (2)). In an electrodeposition reaction involving paramagnetic species it has the same direction as the concentration gradient (χm > 0) and thus tends to move volume elements with high susceptibility away from the working electrode. It has been shown that fc can counterbalance the natural convection [22]. In orientation 1 it has the opposite direction as the vector NC and indeed seems to counterbalance it partly (polarization increases when B is on). In orientation 2 the fc vector points away from the electrode and is perpendicular to NC and MHD. Again the flow profile

Fig. 6. Optical microscope images of the layers deposited at jtotal = −25 mA cm−2 , with the shapes of hydrogen bubbles. (A and D are for orientation 2, B is for orientation 3 and C is for the orientation 1). (A) B = 0, (B) B = 615 mT, (C) B = 740 mT, and (D) B = 715 mT.

2790

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

resulting from this interaction is hard to predict but from the polarization data we can conclude that it increases the mass transport. For orientation 3 there will be no macroscopic MHD convection (Fig. 1) because j and B are parallel. The steady state deposition potentials varied just by 4–10 mV, and no clear tendency was found (notice that in orientations 1 and 2 Edep varied by more than 60 mV, Fig. 4). Thus it is not astonishing that the iron content remained constant within the limit of error (Table 1). 4.4. Hydrogen evolution reaction In orientation 2 jHER increased with B (Fig. 5). This increase can be explained on the basis of the mechanisms that govern the reduction reactions [20]. The alloy deposition is (mostly) activation controlled, while the HER is mass transport controlled. A superimposed magnetic field perpendicular to the electric field decreases the diffusion layer thickness and thus increases the mass transfer. Therefore, the primary effect of the magnetic field is to facilitate the hydrogen evolution reaction, (see also the current–potential curves). No significant effect on jHER was observed when the magnetic field was applied parallel to the electric field lines (Fig. 5). In this case, no (macroscopic) Lorentz forces appear, and the mass transfer process is not influenced. Optical micrographs taken after the deposition showed traces of the hydrogen bubbles, some of them being quite large (approximately 50 ␮m). They have circular shapes if no MHD convection was present, i.e. for B = 0 and for B applied in orientation 3. For the orientation 2 they have a typical “tail” indicative for the MHD flow around the bubble (Fig. 6). This effect has been already described by other authors [13,44]. The interplay between the natural convection and the MHD effect can be seen in the orientation 1. Even if the Lorentz force tends to align the “tail” of the hydrogen bubble, it does not succeed completely. The final direction of the tail is also affected by the natural convection. Fig. 7A is a typical example for the damping changes of the quartz crystal (in terms of changes of the FWHM of its resonance curve, see theory part) during the deposition in orientation 1. The periodic oscillations seen in Fig. 7A can be explained with interference effects of compressional waves emitted by the quartz crystal. These waves are reflected by the surface of the counter electrode and depending on the distance between the source (quartz surface) and the reflector (counter electrode) interact constructively or destructively. In other words, the volume between the working and the counter electrode acts as a resonance cavity and the dimension of this cavity becomes smaller by the electrodeposition process. A similar behavior has been found in the case of evaporation of liquids above a quartz crystal [45]. A detailed analysis of these oscillations is beyond the scope of this paper. The aperiodic oscillations of the damping in orientation 3 (Fig. 7B, black and dark grey curves) have a different origin. In orientations 2 and 3 the quartz is fixed vertically and the reference electrode between the quartz and the counter electrode acts as a spoiler for the compressional waves. It is well known that the damping of

Fig. 7. Variation of the damping of the quartz crystal during the electrodeposition of NiFe layers with jtotal = −25 mA cm−2 . (A) black curve stands for B = 0 mT, and grey curve stands for 740 mT (orientation 1). (B) black curve stands for B = 0 mT, and grey curve stands for 615 mT (orientation 3), and the light grey curve stands for 715 mT (orientation 2).

the quartz crystal increases with the viscosity of the adjacent medium. As the viscosity of hydrogen is much smaller than that of the aqueous solution the detachment of the hydrogen bubbles is accompanied by an increase in the damping, while a decrease in the damping indicates that new hydrogen bubbles are formed. The increasing tendency of the damping with the time indicates an increasing surface roughness of layers. Rough layers deposited on the quartz crystal induce a larger damping compared to smooth layers. This can be explained with the generation of turbulences above the rough surface which leads to an increased dissipation of energy. The qualitative information about the roughness of the deposited layers given by EQCM can be compared to the quantitative information given by AFM. The gold electrode deposited on the polished quartz has a typical roughness of 4–5 nm (RMS value determined by AFM). The growth of the layer will increase this value up to 25 nm (see below). For orientation 2 (Fig. 7B, light grey curve) the damping stays constant during the whole deposition process. That means that smooth layers are formed which is in accordance with ex situ information (SEM, AFM, see below).

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

2791

Fig. 8. Scanning electron micrographs for jtotal = −25 mA cm−2 , orientation 1 (A and B) and orientation 2 (C and D). (A) and (C) are the deposited layers without an external magnetic field, (B) and (D) are the deposited layers in a magnetic field applied perpendicular to the electric field lines, B = 740 mT (B), respectively 715 mT (D).

4.5. Morphology of the deposited layers Scanning electron microscopy (SEM) showed some degree of porosity for the layers prepared in the absence of the magnetic

field (Fig. 8A and C). Atomic force microscopy (AFM) indicated a pore depth of (at least) some 10 nm and a pore diameter of some hundred nm. With increasing current density the porosity decreases (Fig. 9). If the layers were deposited in a B field

Fig. 9. Scanning electron micrographs for jtotal = −60 mA cm−2 , orientation 1 (A and B) and orientation 2 (C and D). (A) and (C) are the deposited layers without an external magnetic field, (B) and (D) are the deposited layers in a magnetic field applied perpendicular to the electric field lines, B = 740 mT (B), respectively 715 mT (D).

2792

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

Table 2 Surface roughness (Rq ) of the deposited layers, expressed in nm jtotal (mA cm−2 )

B = 0 T (orientation 2)

B = 0.7 T (orientation 2)

B = 0.6 T (orientation 3)

−25 −60

8.6 ± 1.4 10.2 ± 0.5

7.8 ± 0.5 9.7 ± 0.3

7.4 ± 1.1 9.5 ± 0.3

there were no pores. The pore formation is a consequence of the HER. Small hydrogen bubbles stay trapped long enough at the surface to cause the formation of voids. Under the influence of the MHD flow the residence time of the bubbles at the surface is shorter (they are swept away) and pore formation does not occur. In summary it can be concluded that a superimposed B field perpendicular to the electric field lines increases the uniformity and the homogeneity of the layers. The surface roughness of the layers, Rq , was calculated from the AFM height data as the standard deviation of the Z values for a scan size of 6.5 ␮m × 6.5 ␮m. It has been shown that Rq increases with the scan size and approaches a limiting value [46,47]. In a separate series of measurements we made sure that our scan size is large enough to obtain the saturation value of Rq . In the range of current densities and B fields investigated here Rq decreased slightly with B (Table 2), independent of the magnetic field direction. This is in agreement with the disappearance of the pores seen in the SEM. As an additional fact, it was observed that the layers deposited between −1 and −5 mA cm−2 were very rough and had a poor visual appearance. In orientation 3 no macroscopic Lorentz forces can appear. However, a model proposed by Aogaki predicts the generation of Lorentz forces on a micrometer scale close to the electrode. The theory is based on the interaction of non-equilibrium fluctuations around growing nuclei and the parallel B field [24,48]. The final effect is a local Lorentz force around the growing nuclei. As these forces act locally, the roughness of the layer on the macroscopic scale (i.e. the saturation value of Rq ) will only be slightly affected. Aogaki [24] and Matsushima et al. [49] reported the formation of specific morphologies under these conditions and different scaling behavior of Rq . Also for B and j parallel the tendency to form pores reduced (Fig. 10). However, it seems that at higher current densities the

amount of pores increases again. This finding is in accordance with the Rq values (Table 2). AFM investigations showed that smaller grains are formed in the presence of a magnetic field. This applies to the perpendicular and the parallel orientation (Fig. 11). For the perpendicular orientation of B and j the grains seem to be elongated in one preferential direction (Fig. 11C). This could be a direct consequence of the MHD convection, which would favor the growth of the grains in the direction of the Lorentz force. Such an explanation is in perfect agreement with literature [49]. The local microscopic MHD turbulences, which appear for B and j parallel, seem to induce a more or less circular symmetry of the grains (Fig. 11B). 4.6. X-ray diffraction (XRD) characterization It is well known that the texture of electrodeposited Ni and Fe layers can depend on the current density and the hydrogen evolution reaction [44,50,51]. As it has been discussed above the samples deposited in orientation 1 at jtotal = −25 mA cm−2 showed a change of the morphology if plated in a magnetic field. Two samples deposited at 0 T and two at 0.74 T (B perpendicular to E) were chosen for the XRD investigations (Fig. 12). Four peaks could be seen for the range of 2 Theta angles from 40◦ to 80◦ . Three of these peaks can be attributed to the phase NiFe (Taenite-Awaruite, cubic system, space group 225-Fm3m, JCPDS # 23-297). The peak at 68◦ can be attributed to the quartz substrate. We calculated the orientation index, M(h k l), according to the standard procedure [51]. One can see that the amount of grains with (1 1 1) and (2 0 0) texture increase slightly in the magnetic field whereas the amount of (2 2 0) textured grains decreases (Fig. 12B). Tabakovic reported that the growth of small grains with (2 0 0) texture was enhanced by the application

Fig. 10. Scanning electron micrographs of the deposited layers in orientation 3. (A) jtotal = −60 mA cm−2 and (B) jtotal = −25 mA cm−2 .

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

2793

Fig. 11. AFM images of the layers deposited at jtotal = −60 mA cm−2 : (A) 0 mT, (B) 615 mT (orientation 3), and (C) in 715 mT (orientation 2).

of a magnetic field [14] which seems in perfect agreement with our results. The magnetic field increases the HER thus increasing the specific adsorption of some inhibiting species (like Hads or hydroxides). It can also lead to a preferential growth of the

grains if their easy axis of magnetization is in the direction of the applied field. Both causes can be responsible for the texture effects described above. Furthermore, the additional convection induced by the MHD effect can be invoked to be responsible for the texture evolution [51].

2794

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795

when B was applied perpendicular to E and round grains were generated in the case that B was applied parallel to E. The traces of the hydrogen bubbles were circular in the absence of a B field and for B and E parallel. They were elongated in the case B was perpendicular to E, due to the effect of the MHD convection. Even if the morphology of the layers on the nanometer scale seems to be influenced by the presence and the direction of a magnetic field, the roughness of the layers (on the micrometer scale) was not dramatically influenced. This proves the importance of scaling size in the discussion of the roughness and morphology effects. Acknowledgments This work was financially supported by the Deutsche Forschungsgemeinschaft (DFG) within the Collaborative Research Center 609 “Electromagnetic Flow Control in Metallurgy, Crystal Growth and Electrochemistry”. References

Fig. 12. XRD patterns (A, black curves: no magnetic field, gray curves: B = 0.74 T) and orientation index (B, () (1 1 1), (䊉) (2 0 0), () (2 2 0)) for samples deposited in orientation 1 at jtotal = −25 mA cm−2 .

5. Conclusion The influence of an external static magnetic field on the electrodeposition of Ni–Fe alloys was studied. It was shown that the magnetic field influences the hydrogen evolution reaction. If the magnetic field was applied perpendicular to the electric field an increase in the partial current due to the hydrogen evolution was obtained, that was attributed to the MHD effect. For B and E parallel no significant effect on the hydrogen reduction was found. This proves once more the role of Lorentz forces in electrodeposition processes. The orientation of the working electrode with regard to the gravity field was proven to be important. Specific morphologies are generated and different behavior of the potential curves and Fe content of the alloys were noticed in function of the orientation of the working electrode. From these results it could be concluded that the magnetically driven convections are in the same order of magnitude as the natural convection. Furthermore, the macroscopic and microscopic MHD effects induce specific morphologies of the deposited layers and of the shape of the hydrogen bubbles. Elongated grains that presumably grow in the direction of the Lorentz force were obtained

[1] A. Brenner, Electrodeposition of Alloys. Principles and Practice, vol. 1, Academic Press, New York, London, 1963. [2] H. Dahms, I.M. Croll, J. Electrochem. Soc. 112 (1965) 771. [3] S. Hessami, C.W. Tobias, J. Electrochem. Soc. 136 (1989) 3611. [4] M. Matlosz, J. Electrochem. Soc. 140 (1993) 2272. [5] D. Landolt, Electrochim. Acta 39 (1994) 1075. [6] T. Krause, L. Arulnayagam, M. Pritzker, J. Electrochem. Soc. 144 (1997) 960. [7] W. Plieth, G.G. Georgiev, J. Solid State Electrochem. 9 (2005) 859. [8] J. Horkans, J. Electrochem. Soc. 128 (1981) 45. [9] N. Zech, D. Landolt, Electrochim. Acta 45 (2000) 3461. [10] G. Lu, P. Evans, G. Zangari, J. Electrochem. Soc. 150 (2003) A551. [11] V.C. Kieling, Surf. Coat. Tech. 96 (1997) 135. [12] P.L. Cavalloti, B. Bozzini, L. Nobili, G. Zangari, Electrochim. Acta 39 (1994) 1123. [13] K. Msellak, J.-P. Chopart, O. Jbara, O. Aaboubi, J. Amblard, J. Magn. Magn. Mater. 281 (2004) 295. [14] I. Tabakovic, S. Riemer, V. Vas’Ko, V. Sapozhnikov, M. Kief, J. Electrochem. Soc. 150 (2003) C635. [15] T.Z. Fahidy, Prog. Surf. Sci. 68 (2001) 155. [16] C. O’Reilly, G. Hinds, J.M.D. Coey, J. Electrochem. Soc. 148 (2001) C674. [17] N. Leventis, X. Gao, J. Am. Chem. Soc. 124 (2002) 1079. [18] S. Ragsdale, K.M. Grant, H.S. White, J. Am. Chem. Soc. 120 (1998) 13461. [19] A. Bund, S. Koehler, H.H. Kuehnlein, W. Plieth, Electrochim. Acta 49 (2003) 147. [20] A. Bund, A. Ispas, J. Electroanal. Chem. 575 (2005) 221. [21] G. Hinds, F. Spada, J.M.D. Coey, T.R. Ni Mhiochain, M.E.G. Lyons, J. Phys. Chem. B 105 (2001) 9487. [22] N. Leventis, A. Dass, J. Am. Chem. Soc. 127 (2005) 4988. [23] C. Iwakura, T. Edamoto, H. Tamura, Denki Kagaku 52 (1984) 596. [24] R. Aogaki, Magnetohydrodynamics 39 (2003) 453. [25] R. Morimoto, A. Sugiyama, R. Aogaki, Electrochemistry 72 (2004) 421. [26] J.C. Shannon, Z.H. Gu, T.Z. Fahidy, J. Appl. Electrochem. 29 (1999) 577. [27] P.U. Arumugam, E.A. Clark, I. Fritsch, Anal. Chem. 77 (2005) 1167. [28] M.W. Carr, A.R. Hillman, S.D. Lubetkin, M.J. Swann, J. Electroanal. Chem. 267 (1989) 313. [29] A. Lachenwitzer, O.M. Magnussen, J. Phys. Chem. B 104 (2000) 7424. [30] E. Juzeliunas, K. Juettner, Electrochim. Acta 43 (1998) 1691. [31] A. Bund, G. Schwitzgebel, Electrochim. Acta 45 (2000) 3703. [32] P.C. Andricacos, C. Arana, J. Tabib, J. Dukovic, L.T. Romankiw, J. Electrochem. Soc. 136 (1989) 1336. [33] N. Zech, E.J. Podlaha, L. Landolt, J. Electrochem. Soc. 146 (1999) 2829.

A. Ispas et al. / Electrochimica Acta 52 (2007) 2785–2795 [34] D.A. Bograchev, V.M. Volgin, A.D. Davydov, Russian J. Electrochem. 41 (2005) 1197. [35] V.M. Volgin, A.D. Davydov, D.A. Bograchev, Russian J. Electrochem. 41 (2005) 925. [36] M. Morisue, Y. Fukunaka, E. Kusaka, R. Yshii, K. Kuribayashi, J. Electroanal. Chem. 559 (2003) 155. [37] Y. Oshikiri, M. Sato, A. Yamada, R. Aogaki, Jpn. J. Appl. Phys. 43 (2004) 3596. [38] M.S. Quraishi, T.Z. Fahidy, Electrochim. Acta 25 (1980) 591. [39] Y. Oshikiri, M. Sato, A. Yamada, R. Aogaki, Electrochemistry 71 (2003) 154. [40] Y. Awakura, Y. TAkenaka, Y. Kondo, Electrochim. Acta 21 (1976) 789. [41] M. Ota, S. Izuo, K. Nishikawa, Y. Fukunaka, E. Kusaka, R. Ishii, J.R. Selman, J. Electroanal. Chem. 559 (2003) 175.

2795

[42] K. Nishikawa, Y. Fukunaka, T. Sakka, Y.H. Ogata, J.R. Selman, J. Electroanal. Chem. 584 (2005) 63. [43] I. Ohno, M. Mukai, Electrodep. Surf. Treat. 3 (1975) 213. [44] O. Devos, A. Olivier, J.-P. Chopart, O. Aaboubi, G. Maurin, J. Electrochem. Soc. 145 (1998) 401. [45] A. Bund, G. Schwitzgebel, Anal. Chim. Acta 364 (1998) 189. [46] H. Matsushima, Y. Fukunaka, Y. Ito, A. Bund, W. Plieth, J. Electroanal. Chem. 587 (2006) 93. [47] W. Schwarzacher, J. Phys. Condens. Matter 16 (2004) R859. [48] R. Aogaki, J. Chem. Phys. 103 (1995) 8602. [49] H. Matsushima, T. Nohira, Y. Ito, Electrochem. Solid State Lett. 7 (2004) C81. [50] D.Y. Li, J.A. Szpunar, Electrochim. Acta 42 (1997) 37. [51] H. Matsushima, T. Nohira, I. Mogi, Y. Ito, Surf. Coat. Techn. 179 (2004) 245.