1 Introduction

Zinc is the most widely used metal in surface treatment to enhance the lifetime of ferrous material structures owing to its low cost [1] and ease of application [2, 3]. A large part of zinc coating is used to cover steel sheets in the automotive industry, oil and gas fields [4]. Ferrous structures, such as steel oil pipes, face air oxidation and require a metallic layer to be applied to their surface to avoid this risk of oxidation. A sacrificial coating such as zinc to the surface of steel structures oxidizes preferentially over steel because of its lower redox potential compared to that of steel. Zinc is the metal most often used as a sacrificial coating and has long been deposited on ferrous substrates using various techniques, namely galvanizing, followed by electroplating [5]. In view of the scale of standard potentials of zinc (Zn2+/Zn = − 0.76 V vs. HSE) and iron (Fe2+/Fe = − 0.44 V vs. HSE), zinc is less noble than iron [6, 7]. Zinc coatings tend to oxidise at a faster rate than iron coatings and form passive layers such as ZnO and Zn(OH)2, more commonly referred to as white rust, on its surface to protect metal underneath on contact with its environment. However, zinc presents fewer disadvantages, for example corrosion under humidity conditions, and increases the corrosion rate to temperatures above 55 °C. To improve the anticorrosive performances of pure zinc coatings, zinc-nickel (ZnNi) alloy coatings were developed [7,8,9] as sacrificial coatings by increasing the corrosion potentials in respect to pure zinc coating. These alloys are electrodeposited via an anomalous codeposition with a higher proportion of zinc (less noble species) in respect to nickel (more noble species) in the coating alloy according to the Brenner [10] classification based on two main types of electrochemical behaviour, the first being classified as normal and the second abnormal. In the former, there is a preferential deposition of the most noble metal while, in the latter, there is a preferential deposition of the less noble metal. The composition of the deposits differs from what is presented by the equilibrium considerations, and this classification developed on thermodynamic considerations has its limits.

The electrocrystallization process involves a competition between nucleation and crystal growth processes, more specifically between a lateral growth (bidimensional growth) and a surface normal growth (three-dimensional growth). Both processes can be controlled by the presence of additive agents and by the application of current densities. These phenomena correspond to the adsorption of the metallic atoms – called adatoms or adions—on the surface of the substrate. One of these two processes can be predominant and can cause gaps in the coating crystal structure. The nucleation is the first step of the coating formation and takes place when the charge transfer begins. This process allows the spontaneous formation of a three-dimensional layer in order to get a thickness of coating. The adsorption and the diffusion of adatoms after their stabilization lead to nuclei, spontaneously formed at the surface of the substrate and then allow the creation of new crystals. If imperfections, like rugosity created by sandblasting, are present in the initial surface of the substrate, the nucleation can be influenced. The defects then act as preferential growth sites and participate to grape formations until the full recovery of the substrate [11]. This stage can occur in two forms: instantaneous nucleation or progressive nucleation. Depending on the type of nucleation, the electrolysis current differs. From chronoamperometric curve data taken from the literature [12], the nucleation type and the growth pattern can be determined by plotting \(J\left( t \right) = f\left( t \right)\). Asseli and his co-workers [13] studied the nucleation step according to the bath composition like the zinc and nickel concentrations ratio [Ni2+]/[Zn2+] or to the pH, but also at different operating conditions such as temperature and deposition potential. Their study shows that the nucleation type changes by modifying the metal concentration ratio, from instantaneous to progressive three-dimensional nucleation through increasing the metal concentration ratio from 0.25 to 4.

In this present work, the influence of the surface preparation on zinc-nickel nucleation and growth was studied from acidic electrolyte. Particularly, the cathodic activation or the sandblasting of austenitic steel substrates was examined. The nucleation type was investigated by the theoretical Scharifker-Hills [14] and Palomar-Pardavé [15] models. The morphology, the crystal structure and the adherence of the alloy coatings were investigated.

2 Experimental

The electrolyte was a concentrated zinc-nickel chloride solution prepared with zinc chloride and nickel chloride salts (VWR Chemicals products 99%), potassium chloride (VWR Chemicals products 100%) and boric acid (Sigma Aldrich 99.5%) whose pH was adjusted to 5.4 with hydrochloric acid or potassium hydroxide. These acid electrolytes contained 0.46 M of ZnCl2; 0.51 M of NiCl2,6H2O; 4 M of KCl and 0.16 M of H3BO3, in presence of 0.6 M of glycine to fully chelate nickel species and to form Ni(Gly)2+ complex and in presence of 25 mL L−1 of mixture of triblock PPG-PEG-PPG and PEG1000 molecules in a ratio of 1:4. This volume concentration corresponds to 3 g L−1, the concentration chosen in the middle of the 0.5–6 g L−1 range that was studied in previous works [16,17,18,19]. In our chemical conditions of the zinc-nickel electrolyte, the complexes of zinc and nickel are respectively ZnCl3− and Ni(Gly)2+ due to the non-formation of ZnCl+, ZnCl2 and ZnCl42−, and to the non-chelation of nickel by chlorides. The conditional pKd of the Ni(Gly)2+ species is higher than those of Zn(Gly)2+ and Zn(Gly)22+. So only the nickel glycinate complexes are possible and this is confirmed by a change in color of the electrolytic solution. The concentrations of metal species in the electrolyte are verified by inductively coupled plasma optical emission spectroscopy (ICP-OES) and X-ray fluorescence (XRF) measurements.

2.1 Electrochemical measurements

All the electrochemical measurements were carried out with a BioLogic potentiostat, controlled by the EC-Lab interface. A three-electrode cell was used, containing a platinum counter electrode, a reference electrode of the Ag/AgCl/KCl type (saturated) and a working electrode of 1 cm2 (sample) to investigate the effects of surface preparation on nucleation and the growth of zinc-nickel on a steel substrate. The substrate is an austenitic steel in Fig. 1, rich in chromium and nickel. The steel substrate composition was 31wt% nickel, 28wt% chromium and 41wt% iron. Also this substrate is the most corrosion material, its protection by a zinc-nickel layer is required in an agressive environnement for petroleum applications. Linear sweep voltammetries were carried out at 20 mV.s−1 and at T = 40 °C in order to select the deposition potential for the plotting of the current-transient curves.

Fig. 1
figure 1

EDX analysis spectrum of austenitic steel substrate

Full size image

2.2 Surface preparation and synthesis of coatings

Different surface preparations (as-machined or sandblasting followed by a cathodic activation (CA) at different current densities) of austenitic substrate were tested before electroplating in Table 1. Each as-machined substrate was polished with carbide silicon paper (Si-C 2000) leading to a final roughness not exceeding 0.3 μm (AsM samples) and all sandblasting was achieved with white corundum F80 to provide an average roughness of 2.2 μm (SB samples). Every cathodic activation was made at 0, 50 and 150 A dm-2 for 1 min in an industrial acid solution (20% of H2SO4 at 97% and 7.5% of (NH4)2SO4 at 99%), and the metallic piece (cathode electrode) is then activated by the release of hydrogen evolution. Each test was repeated three times.

Zinc–nickel coatings were carried out at 4 A dm−2 (consistent with industrial conditions) and at 40 °C with an appropriate agitation in order to deposit 10 μm with a platinum-plated titanium as anode electrode. The nucleation is analyzed at a potential of −1.30 V vs Ag/AgCl, chosen to have only a charge transfer control.

Table 1 Preparation process of austenitic steel substrate
Full size table

2.3 Characterization material

The roughness of the austenitic susbtrate is measured after sandblasting using a Mitutoyo RJ210 roughness tester. The morphology was examined by Scanning Electron Microscopy (SEM) SUPRA 40 JEOL with an acceleration voltage of 15 kV at working distance of 10 mm in Secondary Electron (SE) analysis. The microgeometry of the surface was calculated with a LEICA microscope DCM 3D. The thickness of the deposits is measured by X-ray fluorescence (XRF). The alloy composition was determined by a XL3T X-ray Fluorescence (XRF) from ThermoFischer Scientific. For the phase identification, X-ray diffraction (XRD) patterns of deposits were recorded in the angle interval 30–100° (2θ) using a D8 Bruker Advance diffractometer, equipped with a CuKα tube. The crystallite size determination was recorded by using the standard Scherrer equation by viewing full width at half maximum FWHM where K equal 0.89. The preferential orientation of the dominant phase of each sample was determined via the texture coefficient (Tc) using the equation (Eq. 1).

$$T_{c} \left( {hkl} \right) = \frac{{I\left( {hkl} \right)/I_{0} \left( {hkl} \right)}}{{\left( {1/N} \right)\Sigma _{N} I\left( {hkl} \right)/I_{0} \left( {hkl} \right)}},$$
(1)

where \({I}_{0}\left(hkl\right)\) is the standard intensity of the (hkl) plane given by the JCPDS, \(I\left(hkl\right)\) is the observed intensity of the (hkl) plane and \(N\) is the number of diffraction peaks. The adhesion of the electroplated alloy is quantified with a tribometer Anton Paar TBN model through a scratch test on coating, by applying a load of 10 N with an amplitude of 10 mm, thanks to a spherical indenter on which a tungsten carbide ball (diameter 6 mm) is at 1 Hz for 100 cycles and 2500 cycles at 3.14 cm s-1. The images of the scratch test were carried out with the LEICA optical microscope.

3 Results and discussion

3.1 Effect of surface preparation on HER and ZnNi reduction

Figure 2 shows the linear sweep voltammetries relative to the surface preparation. These curves are obtained with an electrolyte support (electrolyte in absence of metallic species) and with zinc-nickel electrolyte from the OCP. The dash curves display the potential for hydrogen production in the electroplating process. These curves reveal that the activation or the sandblasting of steel substrate produce hydrogen gas at more anodic potentials. The hydrogen evolution reduction (HER) appears at − 1.05 V vs. Ag/AgCl on AsM substrate (black dash curve) while it appears at − 0.70 V vs. Ag/AgCl on AsM-150, SB and SB-150 substrates (red, blue and green dash curves respectively). Figure 2 also serves as evidence that the current densities reached related to the reduction of zinc and nickel simultaneous are higher on sandblasted substrates (blue and green curves), due to the higher surface with the rugosity, compared to the as-machined substrates (black and red curves) while the metal reduction occurs at the same potentials: E = − 1.10 V vs. Ag/AgCl. Furthermore, it should be noted that the HER (dihydrogen gas production) can occur before the metal reduction on the sandblasted substrate between − 1.10 ≤ E ≤ − 0.70 V vs. Ag/AgCl. The dash curves in blue and green overlap with the solid curves in blue and green related to the HER production.

Fig. 2
figure 2

Sweep linear voltammograms  of electrolyte behaviour for different substrate preparation. Dash curves (electrolyte support) and solid curves (zinc-nickel electrolyte)

Full size image

3.2 Nucleation and crystal growth

3.2.1 Nucleation at − 1.15 V vs. Ag/AgCl on as-machined substrate activated

Figure 3a shows the chronoamperometric curve at E = − 1.15 V vs. Ag/AgCl for 60 s on a steel activated and un-sandblasted surface. The experimental current-transient curve is supported by SEM images, in Fig. 4, to preview the nuclei characteristic of the first stages of electrodeposition. The Scharifker and Hills model [14] made it possible to analyse the transient current-time curves by exploiting the kinetics of electrodeposition. This model offers two types of three-dimensional nucleation; instantaneous, or progressive, with hemispherical growth controlled by the diffusion of nuclei. Instantaneous nucleation is characterized by a random and simultaneous nucleation on a surface, leading to seeds of identical size (Eq. 2). Progressive nucleation is a non-simultaneous random nucleation whose seeds are of different sizes (Eq. 3), where \(I_{{\max }}\) is the maximum current on the transient current-time curve and \(t_{{\max }}\) is the time corresponding to \(I_{{\max }}\).

$$\left( {\frac{I}{{I_{{\max }} }}} \right)^{2} = \frac{{1,9542}}{{\frac{t}{{t_{{\max }} }}}}\left[ {1 - \exp \left( { - 1,2564\left( {\frac{t}{{t_{{\max }} }}} \right)} \right)} \right]^{2}$$
(2)
$$\left( {\frac{I}{{I_{{\max }} }}} \right)^{2} = \frac{{1,2254}}{{\frac{t}{{t_{{\max }} }}}}\left[ {1 - \exp \left( { - 2,3367\left( {\frac{t}{{t_{{\max }} }}} \right)^{2} } \right)} \right]^{2}$$
(3)
Fig. 3
figure 3

a Chronoamperometric curve of ZnNi on AsM-150 at − 1.15 V vs. Ag/AgCl and b Scharifker–Hills plots

Full size image

The interpretation of Scharifker and Hills plot in Fig. 3b shows that the nucleation of zinc-nickel electrodeposition would be instantaneous at this potential applied on activated AsM substrate. As illustrated in Fig. 4, after 7 s of electrodeposition, it depicts the homogeneous distribution of hemispheric nuclei with the identical size over the whole surface of the substrate characteristic of the instantaneous nucleation. At 13s, nucleus entanglement takes shape with nuclei of fairly similar size. At 17 s, the diffusional aspect of the electroactive species is the limiting factor in the growth of the coating; from this moment onwards, the agglomeration of the nuclei begins to make the deposit grow.

Fig. 4
figure 4

SEM images of ZnNi nucleation on AsM-150, at − 1.15 V vs. Ag/AgCl for three times (7, 13 and 17 s)

Full size image

3.2.2 Nucleation at − 1.30 V vs. Ag/AgCl on different substrate preparations

3.2.2.1 Chronoamperometric curves

Figure 5 shows the chronoamperometric curves at E = − 1.30 V vs. Ag/AgCl applied for 15 s. Figure 5a displays the impact of the cathodic activation on as-machined substrates at three current densities (0, 50 and 150 A dm−2).

Fig. 5
figure 5

Chronoamperometric curves of ZnNi nucleation step on different substrate preparations; a on as-machined and b on sandblasted surface, at − 1.30 V vs. Ag/AgCl

Full size image

The cathodic activation involves a change of the charge double layer by increasing the electrical capacity of the double layer at the interface electrode/electrolyte. Moreover, the cathodic activation implies an increase of the cathodic current density (Jmax) on AsM activated in Fig. 5a for the zinc–nickel electrodeposition by decreasing the time (tmax) needed to reach it. The cathodic activation treatment leads to an increased cathodic current density of zinc–nickel electrodeposition (Jmax) by decreasing the time (tmax). Figure 5b shows that the current-transient curves on sandblasted substrates remain similar with the cathodic activation without changing or affecting the cathodic current density.

3.2.2.2 Scharifker–Hills model

The experimental data of zinc–nickel nucleation for each surface preparation also correlate with the theoretical three-dimensional instantaneous nucleation (black curve in Fig. 6a, b). The deviation of experimental curves compared to the theoretical curves can be attributed at the HER throughout the nucleation of zinc-nickel deposits. These results exhibit that the nucleation step of zinc-nickel is hemispheric three-dimensional instantaneous at − 1.30 V vs. Ag/AgCl, like at − 1.15 V vs. Ag/AgCl determined previously, for each surface preparation, as-machined or sandblasted with a different current density for cathodic activation. Asseli et al. [13] have reported the same results. They reported that the zinc-nickel nucleation type is instantaneous, and that the spherical geometry of an identical size is characteristic of an immediate activation with a constant number of nuclei, regardless of the temperature and of the pH of the electrolyte containing a molar ratio [Ni2+]/[Zn2+] close to 1.

Fig. 6
figure 6

Scharifker–Hills plots for zinc-nickel nucleation on different substrates at − 1.30 V vs. Ag/AgCl

Full size image
3.2.2.3 Palomar-Pardavé model

An exploitation made with the Palomar-Pardavé model [15] (Eq. 4), based on the Scharifker and Hills models, made it possible to simulate the theoretical curves having the optimal correlations with the experimental curves, and thus the extraction of the parameters of nucleation – namely, the nuclei density N0 and the nucleation rate A and also the constant of Hydrogen Evolution Reaction (HER) known as kPR. Each simulation has been repeated six times and has been tested for 6 different value inputs. Regardless of the initial hypothesis, the parameters of nucleation have the same order of magnitude.

$$J\left( t \right) = \left( {P_{1} + P_{4} t^{{ - 1/2}} } \right)~\left[ {1 - e^{{ - P_{2} \left( {t - \frac{{1 - e^{{ - P_{3} t}} }}{{P_{3} }}} \right)}} } \right],$$
(4)

where \(P_{1} = Z_{{PR}} Fk_{{PR}} \left( {\frac{{2CM}}{{\pi \rho }}} \right)\), \(P_{2} = N_{0} \pi kD\), \(P_{3} = A\) and \(P_{4} = \frac{{2FC\sqrt D }}{{\sqrt \pi }}\)  

In the first parameter \(P_{1}\), \(Z_{{PR}} F\) corresponds to the molar charge transferred during proton reduction (C mol−1), \(k_{{PR}}\) is the reaction constant of the parasitic reaction (mol cm−2 s−1) equivalent to the HER, \(z\) is the valence of the ion in solution, \(C\) is relative to concentration of the electroactive species (mol cm−3), \(M\) is the molar mass of the species (g mol−1) and \(\rho\) concerns the density of the species (g cm−3). In the second parameter \(P_{2}\), \({N}_{0}\) involves the nuclei density (cm−2) while \(k={\left(8\pi C/\rho \right)}^{1/2}\). The third parameter \(P_{3}\) is equivalent to \(A\) the nucleation rate (s−1). The latter parameter \(P_{4}\) depends on \(D\), the diffusion coefficient of the species (cm2 s-1). The chronoamperometric curve for the configuration AsM is fitted in Fig. 7 by the Palomar-Pardavé model to determine the nucleation parameters, such as the nuclei number (N0); the nucleation rate (A) and the constant reaction of the proton reduction (kPR). When A is greater than one, the nucleation tends to be instantaneous. In this work, z is equal 2, C is 1.03.10−3 mol cm−3 corresponding to the total of ZnII and NiII concentrations, we considered the values of M (64.33 g mol−1) and ρ (7.22 g cm−3) for an alloy at 16wt. % in nickel.

Fig. 7
figure 7

Fit of chronoamperometric curve by Palomar-Pardavé model on AsM substrate

Full size image

The nuclei number average in Fig. 8 remains in the same order of magnitude between 0.5 and 3.107 nuclei cm−2 no matter the surface preparation. On as-machined substrates (blue column), the cathodic activation involves a higher of active sites, in particular at 50 and 150 A.dm−2, although a strong standard deviation is visible on AsM-50, while on a sandblasted surface (green column), the number density remains constant for all surface preparations. This implies that the reaction sites are immediately activated with a constant number of nuclei and an identical size of seeds.

Fig. 8
figure 8

Density of nuclei determined by Palomar-Pardavé model for the different surface preparation

Full size image
Fig. 9
figure 9

Nucleation rate determined by Palomar-Pardavé model for the different surface preparation

Full size image

Cathodic activation appears to generate lot of active sites and a higher nucleation rate, which increases the active site density on as-machined or sandblasted substrates as illustrated in Fig. 9. In fact, the nucleation rate is superior to zero. In the cases of substrates not activated, the nucleation rate is equal to 26.9 s−1 and 40.0 s−1 (A > 1  corresponding to an instantaneous nucleation) for AsM and SB substrates respectively, while the nucleation rate increases to 1.94 × 1010 s−1 and 3.95 × 1010 s−1 (A > > 1 corresponding to an instantaneous nucleation) for AsM and SB activated substrates respectively by cathodic activation. These results show that the nucleation rate is widely increased by the cathodic activation and that zinc–nickel nucleation seems to be instantaneous no matter the studied surface preparation.

Fig. 10
figure 10

Constant reaction of proton reduction determined by Palomar-Pardavé model

Full size image

Furthermore, the constant kPR, in Fig. 10 related to the HER shows a weak increase in the same order of magnitude on substrates activated by cathodic activation with a weak or high roughness, and correlates with the sweep linear voltammetries in Fig. 2.

3.3 Characterization of materials

3.3.1 Morphology and crystal structure

The zinc–nickel alloys of thickness of around 10 μm electrodeposited for each configuration are composed of 17–18wt% of nickel. The alloy coatings from preparations AsM, AsM-150, SB and SB-150 are characterized by SEM and XRD investigations in order to observe the impact of cathodic activation and sandblasting on the morphology in Fig. 11 and the crystal structure in Fig. 12 respectively.

Fig. 11
figure 11

SEM images of surface morphology of ZnNi alloy coatings and schematic representation

Full size image

As-machined substrates generate a dense and compact alloy coating without roughness while the sandblast implies a dense and coarse surface morphology with different sizes of nodules. Indeed, the microgeometry analysis facilitated the determination of an average roughness of 0.1 and 2.1 μm of surface alloys coatings which grew on as-machined and sandblasted substrate respectively. The profile of coatings remains similar to the substrate profile.

The identification of the phase of the deposits was obtained from X-ray diffraction reflection plotted as a function of 2θ. All coatings are composed by one crystallized phase identified as a γ-phase zinc-nickel alloy. This cubic γ-phase corresponds to the Zn21Ni5 alloy composition having a wt% of nickel coherent with our experimental value. The cubic γ-phase have a space group \(I\bar{4}3m~\left( {217} \right)\) according to the JCPDS 00-006-0653 with an average lattice parameter of 8.92 Å.

Fig. 12
figure 12

XRD patterns of ZnNi alloy coatings

Full size image

By way of comparison, the crystallite size of zinc-nickel was determined by using Scherrer equation according to the surface preparation. The crystallite size was determined from the (600) plane orientation avoiding the substrate interference. The crystallite size lies in the range 74.9–79.8 nm for alloy coatings obtained on AsM and AsM-150 substrates respectively. For the zinc–nickel alloy coatings on SB and SB-150, the size is decreasing to 52.5 nm and 46.8 nm respectively. The texture coefficient was calculated for four main planes in zinc-nickel coating; (330), (600), (552) and (811). Figure 13 shows the texture coefficient and indicates the preferential orientation. The as-machined substrates generate a preferential orientation in the direction (600) for zinc–nickel alloy, although the preferential orientation is less marked on AsM. The sandblasted substrates show a preferential orientation in (330) direction, for zinc-nickel alloy on SB and SB-150 substrates.

Fig. 13
figure 13

Texture coefficient as a function of crystallographic planes of ZnNi coatings

Full size image

3.3.2 Adhesion of zinc–nickel on substrate

It shows that as-machined substrates activated (AsM-150) or not activated (AsM), red and black curves respectively, imply a high variation of the friction coefficient µ from 10 cycles, related to a poor adhesion of zinc–nickel alloy coatings prepared in these conditions displayed in Fig. 14. These strong variations reflect the delamination of zinc-nickel and different contacts between the spherical indenter and the zinc-nickel or the austenitic steel material. While on sandblasted substrates, the roughness involves a better adhesion. Indeed, on SB (blue curve), the friction coefficient remains stable with a strong decrease after 950 cycles relative to the beginning of the friction between the spherical indenter and the steel substrate. The cathodic activation of sandblasted substrates, SB-150 (green curve), improves the adhesion of zinc-nickel. The friction coefficient evolves in the same manner, but the beginning of the friction between the indenter and the steel substrate starts from 1650 cycles.

Fig. 14
figure 14

Evolution of friction coefficient of ZnNi alloy coatings

Full size image

The roughness created by sandblasting and the cathodic activation after sandblasting significantly improve the adhesion of zinc–nickel. In fact, Fig. 15 displays the scratch and shows that on as-machined substrates (AsM and AsM-150), the alloy coating is very friable and brittle as soon as mechanical stress is applied, resulting in a poor adhesion. On SB substrates, however, the zinc-nickel coatings do not break around the scratch of the ball than in AsM substrates.

Fig. 15
figure 15

Optical microscope images of scratch

Full size image

4 Conclusion

In this study, the nucleation type of zinc–nickel alloy from acidic industrial bath was investigated according to the effects of surface preparation such as polishing, activating and sandblasting by means of sweep linear voltammetries and chronoamperometries curves. Voltammetric curves show that the surface preparation changes the HER production potential but does not alter the electrodeposition potential of zinc-nickel alloy. Based on the chronoamperometric curves, the nucleation is three-dimensional instantaneous at − 1.15 V vs. Ag/AgCl and at − 1.30 V vs. Ag/AgCl whatever the surface preparation. The cathodic activation at 150 A dm−2 involves a better reactivity of the surface without changing the density of nuclei. However, the sandblasting of the surface, for the creation of high roughness of the substrate, involves a real change of the characteristics and the features of zinc–nickel, although the alloy composition remains identical. In particular, the morphology barely becomes rough without changing the crystallized phase. The roughness involves a preferential growth on (330) plane with a small crystallite size and a better adhesion performance particularly when the sandblasted substrate is activated by acid cathodic activation at 150 A dm−2, while the as-machined substrates offer a preferential orientation in (600) direction with a larger crystallite size and a poorer adhesion.