Pulse electrodeposition of adherent nickel coatings onto anodized aluminium surfaces

Pulse electrodeposition of adherent nickel coatings onto anodized aluminium surfaces

Applied Surface Science 330 (2015) 39–47 Contents lists available at ScienceDirect Applied Surface Science journal homepage: www.elsevier.com/locate...

3MB Sizes 0 Downloads 49 Views

Applied Surface Science 330 (2015) 39–47

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Pulse electrodeposition of adherent nickel coatings onto anodized aluminium surfaces Cédric Frantz a,∗ , Charlotte Vichery a , Johannes Zechner a , Damian Frey a , Gerhard Bürki a , Halil Cebeci b , Johann Michler a , Laetitia Philippe a a b

Empa—Swiss Federal Laboratories for Materials Science and Technology, Feuerwerkerstrasse 39, 3602 Thun, Switzerland RERO AG, Haupstrasse 96, 4437 Waldenburg, Switzerland

a r t i c l e

i n f o

Article history: Received 17 September 2014 Received in revised form 21 November 2014 Accepted 14 December 2014 Available online 18 December 2014 Keywords: Aluminium Anodization Electrodeposition Nickel Coatings

a b s t r a c t Aluminium is one of the mostly used elements in the industry because of its abundance and low weight. However, the deposition of a metallic coating requires performing the so-called zincate pre-treatment in order to allow the formation of inter-metallic bonds and thereby achieving sufficient adherence. In this work, porous anodic aluminium oxide (AAO) is used as an anchoring intermediate layer for nickel coatings. AAO is grown anodically in sulfuric acid and nickel coatings are deposited by potentiostatic reverse pulse electrodeposition onto as-anodized aluminium surfaces. The electrodeposition of nickel is initiated onto the electrochemically thinned barrier layer of AAO and pursued until the complete covering of the oxide. The electrochemical behavior of Watts and sulfamate baths is investigated by cyclic voltammetry for different barrier layer thickness, allowing to validate the thinning conditions and to determine the appropriate deposition potential of nickel. GD-OES measurements show that low duty cycles are necessary to achieve high filling ratio of the AAO. SEM micrographs show that a smooth uniform coating is obtained when nickel is deposited in presence of additives. © 2014 Elsevier B.V. All rights reserved.

1. Introduction It is well known that the surface of aluminium requires a specific pretreatment in order to electrodeposit adherent coatings [1]. The zincate process consists of the deposition of a thin Zn layer by galvanic displacement. This layer offers a protection against oxidation and thus prevents the formation of a native oxide dielectric layer which would otherwise exclude the establishment of intermetallic bonds between the electrodeposit and the substrate [2]. This process is widely used in industry for further electrodeposition of corrosion and wear resistant coatings. However, alternative processes can be required either when the presence of a zinc intermediate layer is prohibited or when the aluminium alloy composition induces a low reliability of the zincating step. For the last decades, anodization of aluminium has attracted many scientific interests because, in appropriate conditions, it leads to the growth of a porous oxide layer with arranged nanopores. More particularly, Masuda and Fukuda [3] have developed a

∗ Corresponding author. Tel.: +41 587656283. E-mail address: [email protected] (C. Frantz). http://dx.doi.org/10.1016/j.apsusc.2014.12.091 0169-4332/© 2014 Elsevier B.V. All rights reserved.

two-step anodization method which allows obtaining a high ordering of nanopores whose axis is perpendicular to the substrate. The pores are organized in a hexagonal close-packed cell which leads to very high pore density. Furthermore, the pore and cell dimensions can be tuned via the anodization conditions [4–7] and it has been shown that nanopores with a diameter smaller than 10 nm can be obtained in sulfuric acid [8]. The oxide film consists of two distinct parts: the porous oxide and the barrier layer. The thickness of the porous layer is dictated by the anodization duration whereas the barrier layer thickness is proportional to the anodization potential (1–1.2 nm/V) [9,10]. The electrochemical deposition of metals, semiconductors and polymers within nanopore arrays has been intensively studied for the fabrication of nanowires [11–13]. However, direct electrodeposition within anodic aluminium oxide (AAO) remains a challenge because of the presence of an oxide barrier layer at the nanopore bottom. High potentials are thus required in order to tunnel electrons through this dielectric layer. To facilitate the electron tunneling, Nielsch et al. [14] developed a method which allows thinning down the barrier layer by gradually decreasing the applied potential after the second anodization step. Further related studies have demonstrated the possibility to form branched

40

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

nanopores during the thinning process and the hierarchy of multiple branched pores can be controlled via the shape and duration of the potential decrease [15,16]. Additionally, Winkler et al. [17] proposed a method to completely remove the barrier layer localized at the nanopore bottom by successively decreasing the potential, carrying out a chemical pore widening and, applying a cathodic polarization in potassium chloride. Alternatively, Jagminas et al. [18] reported an original process to decrease the barrier layer resistivity. It consists of doping the barrier layer by fluoride through a re-anodization in the presence of fluoride salts. The electrodeposition within as-anodized AAO has been investigated for many different materials by continuous deposition [19], AC deposition [20], and by pulse techniques [14]. It has been shown that continuous depositions may eventually damage the oxide layer either because of cathodic side reactions or by Ohmic heating [21]. Therefore, the elaboration of Ag, Ni, Fe, Bi, Co, Cu and Au nanowires has been investigated by AC electrodepositions since it offers the opportunity to rectify the barrier layer impedance [21–28]. Unfortunately, it has been proven difficult to control the nanowire morphology through AC electrodeposition. More efforts were then dedicated to the optimization of the method proposed by Nielsch et al. [14] which consists of performing a reverse pulse deposition onto the thinned barrier layer. Typically, the metal is deposited during a short cathodic pulse which is followed by a short positive polarization in order to discharge the barrier layer capacitance. Then, a pause can be applied in order to relax the diffusion layer. Sousa et al. [29] have shown that a uniform barrier layer thickness of 10 nm is the most appropriate for uniform nickel deposition. This method was also applied to the electrochemical deposition (ECD) of other materials such as silver [30]. However, the cathodic pulses were always applied in galvanostatic mode. Thus, it is very sensitive to active surface area modifications which may occur during the different stages of the nanopore filling. In this work, nickel was potentiostatically deposited within AAO templates by reverse pulse methods. The purpose was to reach more stable growth conditions over the filling of the different levels of the branched nanopores and the growth of the covering layer. The appropriate potential for nickel deposition was determined by voltammetry techniques for a diluted Watts bath and diluted sulfamate baths with or without additives. The resulting multilayered coatings were characterized by scanning electron microscopy (SEM) and Glow Discharge Optical Emission Spectroscopy (GDOES). 2. Experimental 2.1. Electrochemistry All electrochemical experiments were conducted in a twoelectrode electrochemical cell using a platinum counter electrode and a potentiostat/galvanostat Autolab 302N equipped with a voltage multiplier. The electrolyte temperatures were accurately controlled by a Julabo refrigerated/heating circulator F12-ED and validated with a glass thermometer before each experiment. 0.5-mm thick aluminium (99.999%) disks were purchased from Goodfellow. The substrates were first degreased in 1.25 M NaOH at 60 ◦ C for 5 min then neutralized in 5.55 M HNO3 before being electropolished in HClO4 :C2 H6 O = 1:3 at 10 ◦ C for 120 s at 20 V. Nickel electrolytes were prepared using boric acid and nickel sulfate and nickel chloride for the Watts or nickel sulfamate for the sulfamate bath. Sodium dodecyl sulfate (SDS) and saccharine were added as additives to the nickel sulfamate electrolyte when mentioned. NaOH (98%), HNO3 (70%), H3 PO4 (85%), H3 BO3 (99.5%), NiSO4 (99%), NiCl2 (98%), and Ni(NH2 SO3 )2 (98%) were purchased from Sigma

Aldrich whereas HClO4 (85%), C2 H5 OH (99.8%), H2 CrO4 (99%), SDS (97%), and saccharine (99%) were provided by Fluka. 2.2. Electron microscopy High resolution SEM images were acquired with a Hitachi S4800 FE-SEM with an acceleration voltage of 2 kV. Cross sections and tomography images were obtained by using a focused ion beam (FIB)/SEM Tescan workstation equipped with a Shottky Field emission gun electron beam. The FIB Lyra was operated with a gallium source at 30 kV and 15 pA. The 3D reconstruction was done by combining 36 slices with FIJI ImageJ Freeware software and the 3D image resolution is x = 6.7 nm, y = 6.7 nm, z = 19.8 nm. 2.3. X-ray diffraction X-ray diffraction patterns were recorded using a Bruker D8 Discovery diffractometer in Bragg-Brentano configuration, using the ˚ The structural coherence length was Cu K␣ radiation ( = 1.5418 A). evaluated through individual peak profile analysis using the Fullprof suite of programs and the Scherrer formula, after correction from the instrumental resolution function. 2.4. Filling ratio GD-OES measurements were performed with a Jobin Yvon JY 5000 RF with a pressure of 600 Pa and an output power of 12 W. When required, the nickel covering layer was minutely mechanically polished with SiC abrasive paper. Nickel has been calibrated using a pure nickel (99.999%) reference material. It thereby allowed determining the nickel quantity present in the AAO template by integrating the Ni signal over a specific time range. The uncertainty of the results was calculated by the uncertainty propagation formula. 3. Results and discussion 3.1. Anodization: structure of AAO grown from Al Porous AAO layers were prepared by a two-step potentiostatic anodization process at 25 V in 0.3 M sulfuric acid at 3 ◦ C. The first step consists of applying a long anodization of 8 h. It gives rise to an unordered porous oxide which is subsequently dissolved in a mixture of 0.4 M H3 PO4 and 0.2 M H2 CrO4 at 60 ◦ C for 1 h. This first step allows the formation of a honeycomb-like pattern on the aluminium surface. Afterwards, the pre-structured substrate was anodized for 30 min under the same conditions as the 1st anodization. This 2nd anodization leads to the formation of ordered nanopore arrays. Then the barrier layer was thinned down by exponentially decreasing the applied voltage through 60 steps of 20 s. Finally, the barrier layer thickness was homogenized by maintaining the last potential value for 600 s. It is well known that the barrier layer thickness is proportional to the anodization potential and it can then be monitored by adjusting the last potential value, later called “the thinning potential”. Fig. 1 shows the potential form applied for the 2nd anodization step and the resulting current density. The part (a) corresponds to the steady state anodization performed at 25 V. During the first instants, the barrier oxide layer is built up and the current density drops to its minimum value. Afterwards, the current density increases due to the formation of cracks in the oxide which serve as nucleation sites for the pore growth. During the part (b), the anodization potential was exponentially decreased through 60 steps of 20 s until it reached the selected potential value which corresponds to the targeted barrier layer thickness. During this stage, the average current density decreases exponentially. For

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

41

Fig. 1. Applied potential (left) and resulting current density (right) during the 2nd anodization process of pure aluminium in 0.3 M sulfuric acid at 3 ◦ C. The parts (a), (b), and (c) correspond to the 2nd anodization, the barrier layer thinning, and the barrier layer homogenization, respectively.

each potential step, the current density starts from a minimum value and logarithmically increases due to the formation of new cracks in the barrier layer which eventually induce the growth of new generations of nanopores with smaller diameters. This mechanism leads to the formation of a root-like porous structure at the nanopore bottom [15,16,25,31]. Finally, the chosen thinning potential is applied for 600 s (part (c)) in order to homogenize the barrier layer thickness. In this paper, the influence of the barrier layer thickness on nickel electrodeposition has been studied and several thinning potential (10 V, 7.5 V, 5 V and 2.5 V) have been investigated. Fig. 2 shows the surface of an AAO layer grown with the previously described two-step anodization. The porous oxide exhibits the hexagonal close-packed arrangement and presents an average pore diameter of 26.3 ± 2.5 nm, an interpore distance of 65 ± 1.4 nm. These characteristic dimensions convey a porosity of 14.8 ± 3% and a pore density of 2.7 ± 0.1 (×1010 ) pores cm−2 in ordered domains. 3.2. Voltammetry study for Ni electrodeposition The as-anodized aluminium disks were used as templates for the electrodeposition of nickel. It has been reported in the literature that direct current deposition onto the barrier oxide layer is feasible but requires high potential for electron tunneling and may eventually lead to the breakdown of the AAO template. To avoid that, one can apply AC [21] or reverse pulse galvanostatic depositions [14]. The latter one consists in applying a galvanostatic deposition pulse, a short anodic pulse in order to discharge the double layer capacitance, and a long pause for relaxing the diffusion layer. This method is claimed to limit possible cathodic side reactions at the pore bottom and to minimize local heating of the barrier layer by Joule effect. Here, the objective is to obtain a high filling ratio of the nanopores and to pursue the electrodeposition process until the formation of a homogeneous covering layer on the AAO surface. To achieve this, potentiostatic pulse depositions are preferred to galvanostatic methods since the active surface area evolves as the nanopores are progressively filled by nickel. Another advantage of potentiostatic methods lies in the high selectivity of electrochemical reactions, thus allowing avoiding hydrogen evolution. Foremost, cyclic voltammograms have been recorded in order to determine the appropriate deposition potential ranges with respect to the barrier layer thickness. A Watts and a Cl-free sulfamate baths have been studied. Nevertheless, the nickel concentration has been divided by a factor 10 compared to classical Watts and sulfamate baths in order to decrease the growth rate

of nickel nanowires; thereby limiting Joule heating and the ohmic drop in the barrier layer. The electrolytes consist of 0.73 M H3 BO3 containing 0.114 M NiSO4 and 0.019 M NiCl2 for the Watts bath and 0.154 M Ni(SO3 NH2 )2 for the sulfamate bath. The electrolytes were weakly stirred in order to maintain a homogeneous temperature of 45 ◦ C. The temperature has been chosen lower than the commonly used one in order to discard alumina hydration which could intervene during the deposition stage. Fig. 3 shows cyclic voltammograms recorded in the Watts bath at 1 V s−1 in a two electrodes setup using as-anodized AAO with different thinning potential as working electrodes. The cyclability behavior is discussed for a thinning potential of 10 V (Fig. 3a). The cathodic peak is attributed to nickel electrodeposition and the peak potential remains at about −12.5 V for every cycle. This broad peak presents a shoulder at about −10 V which does only appear during the first scan. It can be explained by the modification of the electrode surface as the nickel, which has been electrodeposited onto the barrier layer during the first cycle, was not dissolved during the reverse scan. This shoulder may thus be related to a cathodic side reaction which only occurs at the barrier layer/electrolyte interface. A voltammetric scan has then been performed in 0.73 M H3 BO3 but no significant faradic reaction is observed at the corresponding potential, meaning that the reaction involves nickel salts. Fig. 3b shows the influence of the barrier layer thinning potential on the voltammetric behavior. The peak potential is expected to be more negative as the barrier layer thickness increases since electrons require more energy to tunnel through it. It is indeed observed for the thinning potential of 10, 7.5, and 5 V for which the peak potential is of about −12.5, −11, and −8.5 V, respectively. However, the recorded peak potential becomes more negative again for a thinning potential of 2.5 V. This behavior may be related to an incomplete thinning of the dielectric barrier and alternate thinning procedures have been sought for this potential. Fig. 4 presents the chronoamperograms associated to two different thinning methods. Method 1 (Fig. 4a) corresponds to the one described previously. Method 2 (Fig. 4b) is similar to method 1 but the time of each thinning step has been doubled, thereby consisting of 60 steps of 40 s with an exponentially decreasing potential after which a homogenization step is performed at 2.5 V for 1200 s. For method 1, the current does not reach its maximum value neither during the short transition steps nor during the long homogenization stage, thus proving that the thinning process is not fully completed. Concerning method 2, even though the current does not reach its maximum value during the transition steps, it rapidly stabilizes during the homogenization step at its maximum value of about 0.12 mA cm−2 .

42

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

Fig. 2. SEM images of AAO surfaces grown onto pure aluminium by a two-step anodization at 25 V in 0.3 M H2 SO4 at 3 ◦ C.

Fig. 3. Cyclic voltammograms of the diluted Watts bath recorded at 1 V s−1 for AAO; (a) cyclability for a thinning potential of 10 V and (b) influence of the barrier layer thinning potential.

Fig. 5a compares the barrier layer thinning methods 1 and 2 by cyclic voltammetry. The shift in the peak potential (i.e. −10 V for method 1 and −6 V for method 2) confirms the previously stated incomplete barrier thinning through method 1. Thus, the ideal thinning process consists of steps which would be long enough for reaching the anodization steady state. That is to say, the current value must reach its maximum value for every potential step, ensuring than the minimum resistivity of the barrier layer is achieved. Fig. 5b compares the cyclic voltammograms recorded for the Watts and the sulfamate baths in AAO for a thinning potential of 5 V. It appears that, whatever the barrier layer thickness, the peak potential is slightly less negative for the sulfamate bath than for

the Watts bath. The slight increase in the cathodic current may be explained by the higher concentration of nickel cations. The peak potentials obtained for each condition are listed in Table 1. 3.3. Potentiostatic pulse deposition Fig. 6 shows a typical chronoamperogram obtained during the filling of an AAO template by reverse pulse deposition of nickel. The reverse pulse deposition period consisted of a cathodic pulse of 8 ms at −11 V, and an anodic discharge of 2 ms at 7.5 V which corresponds to the barrier thinning potential. The evolution of the cathodic current is representative of the different deposition

Fig. 4. Current transit curves recorded during the 2nd anodization with a thinning potential of 2.5 V with (a) the thinning method 1 and (b) with the thinning method 2.

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

43

Fig. 5. Comparison of the two different thinning methods for a thinning potential of 2.5 V with the diluted Watts bath (a) and comparison of the diluted Watts and sulfamate baths for a thinning potential of 5 V (b) by cyclic voltammetry at 1 V s−1 .

Fig. 6. Potentiostatic reverse pulse deposition of nickel at −11 V in an AAO template for a thinning potential of 7.5 V. The electrolyte consisted of 0.73 M H3 BO3 , 0.154 M nickel sulfamate, 0.7 mM SDS, and 10.9 mM saccharine, was slightly stirred, and thermostated at 45 ◦ C. The cathodic deposition pulses and the anodic discharge pulses are highlighted in blue and red, respectively (For interpretation of the color information in this figure legend, the reader is referred to the web version of the article.).

Table 1 List of the peak potentials determined for the Watts and the sulfamate bath with different barrier layer thickness. Thinning potential [V]

10 7.5 5 2.5 (method 1) 2.5 (method 2)

Peak potential [V] Watts bath

Sulfamate bath

−12.5 −10.8 −8.5 −10 −6

−10.5 −9.5 −7.8 −8.3

stages during the filling and covering of the AAO template. The first part corresponds to the deposition of nickel within the nanopores and occurs at low current density. Once the first nanopores are completely filled, the current density increases because of the progressive growth of three dimensional clusters on the AAO surface. The current density eventually reaches a maximum value which corresponds to the maximum active surface area. Then, the current density slightly fluctuates due to the simultaneous growth and

coalescence of isolated nickel clusters. Isolated islands eventually merge to form a homogeneous covering film on the AAO template and the current density stabilizes to a lower value as the active surface area tends to approach the geometric surface area of the substrate. The anodic part of the figure corresponds to the anodic discharge and may thus evolve with, inter alia, the microscopic surface area of the nickel deposit. Fig. 7 shows cross-sectional views of the AAO layer filled with nickel as described after Fig. 6. The AAO thickness is of about 2.6 ␮m and is homogeneous all along the FIB cut. The zoom (right) shows the AAO/Al interface and branched nanopores which were formed during the thinning stage. This portion of the AAO spreads on about 350 nm. Only a small fraction of the nanopores are completely filled and contributed to the growth of the top coating. Alternatively, one can see that the nucleation occurred homogeneously in all branches, thereby demonstrating the homogeneity of the barrier layer under the corresponding thinning conditions. Tomography images (Fig. 7c and d) show that almost all of the pores are filled on the first micron. However, the ratio of filled nanopores dramatically decreases over the next 500 nm. Therefore, the partial filling

44

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

Fig. 7. (a and b) Cross-sectional SEM micrographs of AAO template with a barrier layer thinned down with a potential of 7.5 V and filled with nickel by reverse pulse deposition at −11 V. (c and d) 3D views obtained by FIB/SEM tomography from the region highlighted by the red rectangle; the observation orientation is indicated by the arrows (For interpretation of the color information in this figure legend, the reader is referred to the web version of the article.).

is related to different growth rates, given in length per time unit, of individual nanowires. This may result from diffusional limitations (i.e. contribution and competition of hemispherical diffusion layers at the porous layer top surface), as discussed hereafter. 3.3.1. Filling ratio Fig. 8 schematizes the linear and hemispherical parts of the Ni2+ diffusion layer obtained after a time t of Ni electrodeposition within an AAO template. The two defects of the AAO structures which are represented (i.e. pore defect and cell defect) may affect the growth rate in individual nanopores once the bulk electrolyte contributes to mass transfer. On the one hand, the nanopores may differ in size and shape. Let us consider a single cell of surface, SC , in the middle of which lies a pore of surface, SP . For long electrolysis time, the hemispherical layer can be approximated to a linear diffusion layer whose surface SD corresponds to the cell surface. In this case, the deposited mass per time unit would be dictated by the diffusion cell in the bulk electrolyte but it would be arranged in a pore of smaller diameter. The growth rate would then be proportional to the ratio SD /SP , i.e. the average growth rate is inversely proportional to the porosity of the template. On the other hand, the pore ordering can be disrupted and thus brings about cell defects, as it is easily observed at the boundary of ordered domains. These defects affect the interaction with neighboring hemispherical layers thus forcing geometric readjustments of the approximated diffusion cell surface for the defective cell and for all of the surrounding ones. This eventually modifies the ratio SD /SP , thereby leading to uneven growth rates in the nanopores. The influence of the barrier layer thickness, the deposition potential, the pulse time, and the duty cycle (i.e. pulse time divided by the whole period) have been investigated in order to optimize the filling ratio. Indeed, Trahey et al. [32] have demonstrated that the filling ratio can be dramatically increased by performing pulse deposition; the periodical relaxation of the diffusion layer allowing preventing diffusional limitations. Thus, the filling ratio is expected

to be more important for lower duty cycle. GD-OES has been used in order to evaluate and compare the filling ratio obtained for different electrodeposition conditions. Nickel has been calibrated using a pure nickel (99.999%) reference material and gave a proportionality factor of 7.3 ± 0.7 [×10−8 g (V s)−1 ]. The AAO layer thickness and porosity has been considered identical for all the investigated conditions, thereby permitting to calculate the filling ratio just by integrating the amount of nickel present in the oxide. Two ideal GD-OES profiles are shown in Fig. 9. Fig. 9a has been recorded during the sputtering of an AAO/Ni composite with negligible amount of overgrown nickel outside of the nanopores. Fig. 9b has been obtained for an AAO/Ni composite covered by a thin nickel layer which has been smoothed through a careful mechanical polishing. Experience has shown that it is mandatory to have a really smooth and homogeneous surface in order to distinguish the interface between the top nickel coating and the nickel nanowires. Once the nickel top layer has been totally removed, the nickel signal drops down to a low value which corresponds to the small fraction of nickel nanowires which have reached the nanopore extremity. Afterwards, the nickel content may again increase while the analysis is pursued deeper in the composite since the ratio of filled nanopores may become more important, as seen on the crosssectional views (Fig. 7). Once all of the oxide has been sputtered, the aluminium signal increases and then stabilizes due to the high sputtering rate of aluminium compared to the one of alumina. After integrating the quantity of nickel present in the composite, the filling ratio is evaluated by dividing the volume of nickel nanowires by the total volume of the nanopores. Table 2 lists the evaluated filling ratio for a barrier layer which has been thinned down with a potential of 7.5 V. The deposition of nickel has been carried out with the Watts bath by reverse pulse deposition which consisted of a cathodic pulse of 8 ms and an anodic pulse of 2 ms at 7.5 V. The results show that the best filling ratio is achieved for a deposition potential −11 V, which is of about the potential value of the cathodic peak observed on the

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

45

Fig. 8. Schematic cross-sectional representation of the diffusion layer of Ni2+ during the electrodeposition of nickel in porous aluminium oxide.

Fig. 9. GD-OES profiles recorded during the sputtering of AAO layers filled with nickel without (a) and with (b) a thin smooth nickel top layer.

Table 2 Influence of the deposition potential on the filling ratio obtained with the Watts bath, a pulse time of 8 ms, a duty cycle of 80%, and a barrier layer thinning potential of 7.5 V. Deposition potential [V] Filling ratio [%]

−10 28 ± 6

−11 (peak) 40 ± 10

−12 13 ± 4

−13 4±2

Table 3 Influence of the barrier layer thinning potential on the filling ratio obtained with the Watts bath, a pulse time of 8 ms, a duty cycle of 80%, and a deposition potential corresponding to the cathodic peak. Thinning potential [V]

Deposition potential [V]

Filling ratio [%]

10 7.5 5 2.5 (method 1) 2.5 (method 2)

−12.5 −11 −8.5 −10 −6

38 ± 10 40 ± 10 34 ± 7 23 ± 5 41 ± 8

voltammogram (Fig. 3). For lower over-potentials, the diminution in the filling ratio may result from uncompleted nucleation onto the barrier layer whereas it may be linked to hydrogen evolution and/or dendritic growth for higher over-potentials. Table 3 presents the filling ratio evaluated by GD-OES for different barrier layer thinning steps. The deposition of nickel has been carried out in the Watts bath by reverse pulse deposition. A cathodic pulse at the corresponding cathodic peak potential was applied for 8 ms and subsequently followed by an anodic discharge of 2 ms at the thinning potential. It appears that the filling ratio is not notably

Table 4 Influence of the pulse parameters on the filling ratio with the Watts bath, a barrier layer thinning potential of 7.5 V, and a deposition potential of −11 V. Cathodic pulse time [ms]

Duty cycle [%]

Filling ratio [%]

8 20 20 20 20 40 8

80 20 8 4 2 4 4

40 ± 10 55 ± 20 76 ± 17 89 ± 20 95 ± 17 83 ± 16 70 ± 14

affected by the barrier layer thickness and ranges from 34% to 41%. However, it has to be noted that the thinning procedure is of great importance. Indeed, when comparing the two methods for a thinning potential of 2.5 V, the filling ratio obtained after method 1 is much lower than the other ones. As discussed previously (after Fig. 4), steady state anodization has not been established when the barrier thinning was performed with method 1. This may result in inhomogeneity in the barrier layer impedance which could affect the nucleation and growth rate of nanowires.The influence of pulse parameters on the filling ratio has been investigated by GD-OES and the results are reported in Table 4. The barrier layer was thinned down with a potential of 7.5 V and the corresponding peak potential has been applied during the cathodic pulses. It appears that the filling ratio is not significantly affected by the deposition pulse time in the studied range. However, short pulses are expected to result in a lower faradic efficiency due to the bigger proportion of the double layer charging. Thus, longer pulses of 20 ms have been

46

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

Fig. 10. SEM views of the covering nickel layer deposited by reverse pulse deposition in AAO with (a and b) the Watts bath, and with the sulfamate bath (c) without or (d) with additives. The reverse pulse period consists of an 8 ms deposition pulse at the corresponding cathodic peak potential followed by a 2 ms anodic pulse at 7.5 V.

preferred for investigating the effect of the duty cycle. It is thereby confirmed that the filling ratio is dramatically increased for low duty cycle values; the highest filling ratio having been obtained for a duty cycle of 2% and a cathodic pulse of 20 ms. 3.3.2. Characteristics of the nickel coatings Fig. 10 shows SEM micrographs of the covering nickel layer grown onto AAO whose barrier layer has been thinned down with a potential of 7.5 V. Three different electrolytes have been investigated: the Watts bath (a and b), the sulfamate bath (c), and the sulfamate bath with 0.7 mM SDS and 10.9 mM saccharine as additives (d). Ni electrodeposition has been performed at 45 ◦ C by reverse pulse deposition which consisted of an 8 ms deposition pulse at the corresponding cathodic peak potential followed by a 2 ms anodic pulse at 7.5 V. No pause was included in the period since a fast process was sought. The covering nickel film grown from the Watts bath (Fig. 10a and b) presents many cracks on the surface which may result from internal stresses in the deposit. The use of a Cl-free sulfamate bath has then been preferred in order to limit the

residual stress in the coating (Fig. 10c). The resulting coating does not present any cracks. It is mainly made of big agglomerates but small islands are also visible on the surface; which is characteristic of a progressive 3D nucleation and growth process. The use of additives has then been considered for obtaining a smooth coating surface (Fig. 10d). Indeed, the adding of 0.7 mM SDS and 10.9 mM saccharine allows the electrodeposition of a crack-free coating with a very smooth surface. The nucleation and growth process can be compared to a progressive 2D mode, as comforted by micrographs of Fig. 11. Fig. 11 shows Ni coatings electrodeposited onto AAO template for two different times after filling the nanopores. Directly after the filling of the AAO template (a), i.e. after 8000 reverse pulse periods, nickel overgrowths appear on the surface and may merge with neighboring ones to form small islands. They may vary in size since the growth rate dispersion in the nanopores can either delay or accelerate their appearance. After 10,000 reverse pulse periods (b), most of these islands have started to merge together and began to form a covering nickel layer. It can be noticed that

Fig. 11. Overview of the formation of the covering Ni layer by reverse pulse deposition from the sulfamate bath with additives; after 8000 (a) and 10,000 (b) reverse pulse periods. The reverse pulse periods consist of an 8 ms deposition pulse at the corresponding cathodic peak potential followed by a 2 ms anodic pulse at 7.5 V.

C. Frantz et al. / Applied Surface Science 330 (2015) 39–47

this coalescence step occurs with negligible thickness variation, meaning that the lateral growth is strongly favored over the vertical growth. It thereby confirms that the nucleation and growth process can be likened to a progressive 2D model. XRD analysis showed the (2 0 0) and (2 2 0) planes on this sample and the application of the Scherrer equation highlighted a nanocrystalline nickel with grain sizes of 27.6 ± 0.9 and 32.6 ± 1 nm, respectively. 4. Conclusion In this work, Ni electrodeposition has been successfully performed within thin AAO layers by reverse pulse deposition in potentiostatic mode. The voltammetric behavior of diluted Watts and sulfamate baths has been investigated in AAO for different barrier layer (BL) thinning method and potential. The presence of a cathodic side reaction has been evidenced at a slightly less cathodic potential than the reduction of Ni2+ but just occurs at the BL/electrolyte interface before it has been covered by Ni nuclei. It has been demonstrated that the Ni deposition potential logically becomes less negative as the BL thickness decreases. Also, it has been shown that the procedure employed to thin down the BL is of major importance for achieving a homogeneous thickness; i.e. the time of each step should be long enough to reach steady state anodization conditions. Finally, the filling ratio of AAO by nickel has been evaluated by GD-OES by integrating the amount of Ni present in the porous layer. The results show that the BL homogeneity is crucial for achieving high filling ratio whereas its thickness does not induce significant changes. The highest filling ratio have been obtained for a deposition potential which matches to the Ni2+ /Ni0 cathodic peak potential and for low duty cycle value allowing periodic relaxations of the diffusion layer. Concerning the Ni covering layer, it has been found that the use of a diluted sulfamate bath with SDS and saccharine as additives allows the formation of a smooth and crack free coating. These Ni/AAO composite materials also exhibit interesting mechanical properties [33]. Acknowledgements The authors would like to thank the Commission for Technology and Innovation (CTI) of the Swiss Confederation and RERO AG for supporting this research.

47

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33]

W.G. Zelley, J. Electrochem. Soc. 100 (1953) 328–333. G. Qi, X. Chen, Z. Shao, Thin Solid Films 406 (2002) 204–209. H. Masuda, K. Fukuda, Science 268 (1995) 1466–1468. J.P. Osulliva, G.C. Wood, Proc. R. Soc. A 317 (1970) 511. K. Nielsch, J. Choi, K. Schwirn, R.B. Wehrspohn, U. Gosele, Nano Lett. 2 (2002) 677–680. G.E.J. Poinern, N. Ali, D. Fawcett, Materials 4 (2011) 487–526. G.D. Sulka, in: A. Eftekhari (Ed.), Nanostruct. Mater. Electrochem., Wiley-VCH, Weinheim, Germany, 2008, Chapter 1. E. Moyen, L. Santinacci, L. Masson, W. Wulfhekel, M. Hanbuecken, Adv. Mater. 24 (2012) 5094–5098. R.P. Simon Wernick, P.G. Sheasby, The surface treatment and finishing of aluminium and its alloys, ASM Int. (1987). F. Keller, M.S. Hunter, D.L. Robinson, J. Electrochem. Soc. 100 (1953) 411– 419. G.E. Possin, Rev. Sci. Instrum. 41 (1970) 772. J.D. Klein, et al., Chem. Mater. 5 (1993) 902–904. C.R. Martin, Science 266 (1994) 1961–1966. K. Nielsch, F. Muller, A.P. Li, U. Gosele, Adv. Mater. 12 (2000) 582–586. W. Cheng, M. Steinhart, U. Goesele, R.B. Wehrspohn, J. Mater. Chem. 17 (2007) 3493. C.T. Sousa, A. Apolinario, D.C. Leitao, A.M. Pereira, J. Ventura, J.P. Araujo, J. Mater. Chem. 22 (2012) 3110–3116. N. Winkler, J. Leuthold, Y. Lei, G. Wilde, J. Mater. Chem. 22 (2012) 16627– 16632. A. Jagminas, M. Kurtinaitiene, R. Angelucci, G. Valincius, Appl. Surf. Sci. 252 (2006) 2360–2367. J. Choi, G. Sauer, K. Nielsch, R.B. Wehrspohn, U. Gosele, Chem. Mater. 15 (2003) 776–779. D. Fern, B.R. Best, G.R. Jessup, Alcan Int Ltd., patent EP0368470 A1, 1988. N.J. Gerein, J.A. Haber, J. Phys. Chem. B 109 (2005) 17372. A.J. Yin, J. Li, W. Jian, A.J. Bennett, J.M. Xu, Appl. Phys. Lett. 79 (2001) 1039– 1041. K.H. Lee, H.Y. Lee, W.Y. Jeung, W.Y. Lee, J. Appl. Phys. 91 (2002) 8513–8515. G. Ali, M. Ahmad, J.I. Akhter, K. Maaz, S. Karim, M. Maqbool, S.G. Yang, IEEE Trans. Nanotechnol. 9 (2010) 223–228. Z. Wu, Y. Zhang, K. Du, Appl. Surf. Sci. 265 (2013) 149–156. N. Ji, W. Ruan, C. Wang, Z. Lu, B. Zhao, Langmuir 25 (2009) 11869–11873. J. Jong-Hyun, K. Sun-Hong, Y. Choi, K. Sung-Soo, Phys. Status Solidi C 4 (2007) 4429–4432. K. Wada, T. Hatano, K. Uchida, J. Appl. Electrochem. 9 (1979) 445–455. C.T. Sousa, D.C. Leitao, M.P. Proenca, A. Apolinario, J.G. Correia, J. Ventura, J.P. Araujo, Nanotechnology 22 (2011) 315602. G. Sauer, G. Brehm, S. Schneider, K. Nielsch, R.B. Wehrspohn, J. Choi, H. Hofmeister, U. Gosele, J. Appl. Phys. 91 (2002) 3243–3247. L. Zaraska, E. Kurowska, G.D. Sulka, M. Jaskula, J. Solid State Electrochem. 16 (2012) 3611–3619. L. Trahey, C.R. Becker, A.M. Stacy, Nano Lett. 7 (2007) 2535–2539. J. Zechner, G. Mohanty, C. Frantz, H. Cebeci, L. Philippe, J. Michler, Surf. Coat. Tech. (2014), http://dx.doi.org/10.1016/j.surfcoat.2014.08.086.