Electrodeposition of Cobalt Rich Zn-Co alloy Coatings from Citrate Bath

Zn-Co alloy coatings were produced on carbon steel, at room temperature, from citrate baths (0.100 mol L) containing [Zn] = [Co] = 0.05 mol L. Coatings with %m/m Co > %m/m Zn were observed under all deposition conditions, except for I = 10 A m. As the current density (I) was increased, a significant decrease (p<0.05) in %m/m Zn was noted, and the highest value of % m/m Co was observed at 80 A m. The Co-rich coatings presented refined microstructure, with small grain sizes. The smallest I corr of the studied substrate/coatings systems, 2.4 μA cm, was obtained at I = 10 A m. Under these conditions, the coatings contained ~ 30 % m/m Co and 70 % m/m Zn and only the γ-ZnCo phase was obtained. The I corr found in the present work is smaller than those usually found in the literature for Zn-Co coatings with small Co content.


Introduction
Zn coatings are generally used to protect steel components, acting as sacrifice coatings, and decreasing the substrate material corrosion 1 .Szczygiel et al. 2 and Chen and Sun 3 have shown that the co-deposition of zinc with other metals, such as nickel, iron and cobalt, enhances the anticorrosive properties of the zinc coatings.Additionally, the use of these metals to produce Zn-alloy coatings may be an ecological alternative to the high toxicity of Zn-Cd coatings 4 .
The electrodeposition processes of Zn-Ni, Zn-Fe, and Zn-Co alloys are considered anomalous, that is, the deposition of less noble metal ion Zn (II) is favored when compared to the other alloy metal ions 5 .This anomalous process is reported by many works in the literature 4,[6][7][8][9] and various explanations of the anomaly have been suggested, e.g.underpotential codeposition (UPD), kinetic behavior, the hydroxide suppression mechanism, and the influence of Co (III) 10 .However, it is always considered that certain experimental conditions, such as high values of current density (I), for example, can change the process from anomalous to normal and under these conditions the Zn-Co alloy coatings presented a cobalt content higher than that of zinc 11,12 .
The increase in cobalt content in the Zn-Co coating can be interesting for the anticorrosive performance of these layers.Kirilova et al. 12 have observed that coatings containing small cobalt content could still be considered as sacrifice coatings, while those with high amount of this metal were nobler than steel, acting as anticorrosive barrier.In addition, Lichušina et al. 10 have showed that Co-rich Zn-Co coatings (15-18 % m/m Co) presented good appearance and high corrosion resistance in NaCl solution.Lima-Neto et al. 13 have shown similar results for a Zn-Co alloy coating with 18 % m/m Co.However, there is no consensus in the literature concerning this topic.In most of the works, the anticorrosive performance of the alloy decreased with the elevation in % m/m Co 8,14 .
Generally, to fulfill the coatings requests for a better performance, the metallic-alloy deposition baths are always developed and/or modified to enhance the cathodic current efficiency, the amount of certain metal in the coating, or even to refine the grain sizes and the decorative characteristics of the coatings 8,12 .For this purpose, a complexant agent or a leveling additive can be added to an electrodeposition bath, causing several changes on the reduction mechanism of the ions in the solution and/or in the activation energy of the electrode surface [15][16][17] .For this reason, it is possible to find several works concerning the production of Zn-Co coatings from acidic or alkaline baths, containing additives and/or complexant agents 7,8,10,18 .Despite its well known toxicity and the rigorous maintenance control to avoid environmental problems 19,20 , the traditional cyanide-containing alkaline baths are sometimes used for electroplating of Zn and Zn alloys.Following the general trends for the electrodeposition of alloy coatings, however, most of the current researches found in the literature deals with nontoxic or environmentally friend complexant agents and additives 2,7,8,[21][22][23][24] .
Earlier, we have used a nontoxic citrate bath to produce Cu-Zn and Cu-Co alloy coatings on steel substrate 25,26 .In both cases, protective coatings with high anticorrosive performance were obtained, depending on the deposition parameters used.The quality of these coatings was considered satisfactory, even at high values of current density.Therefore, based on these results, the present work aims to study the production of Co-rich Zn-Co alloy coatings on a carbon steel substrate, from a citrate-base bath.Our goals are to verify the influence of citrate in the bath for Zn-Co deposition and evaluate the effects of current density (I) on the cathodic current efficiency (E f ), the metal contents in the coatings (% m/m Zn and % m/m Co), and on the coatings' morphology, texture and corrosion performance.This way, we intend to contribute to a better understanding about the Zn-Co reduction from this kind of bath, producing Co-rich coatings with a higher anticorrosive performance.

Voltammetric experiments
The voltammetric experiments were carried out in the electrolyte solutions described in Table 1.
It was used a three-electrode cell consisted of graphite as the working electrode (exposed area = 0.28 cm 2 ), a Pt spiral as the counter electrode, and the reference electrode was a saturated mercury (I) sulfate electrode (Hg/Hg 2 SO 4 ), SSE.The working electrode was always polished with emery paper (600 mesh), washed with deionized water and alcohol and finally dried, before being immersed in each experiment.The Pt spiral was immersed in 20 %v/v HNO 3 solution for 1 minute, immediately before being used in the experiments, to remove any oxide layer that could be present.The voltammetric experiments were performed without stirring in a potentiostat/galvanostat AUTOLAB PGSTAT 302 N from 0 to -1.6 V SSE , with a scanning rate of 1 mV s -1 .

Cathodic polarization curves
Cathodic polarization curves of the steel substrate were galvanostatically obtained in the solution 5 of Table 1, in the current density range from 0.1 to 120 A m -2 using a potentiostat/galvanostat developed for this purpose 25,26 .AISI 1020 carbon steel discs (area = 4.90 x 10 -4 m 2 ) were used as working electrodes.These samples were first polished with emery paper (100 to 600 mesh), washed with deionized water and alcohol and finally dried, before being immersed in the solutions.The counter electrode was platinum net, which was cleaned in a similar way earlier described for the voltammetric experiments.The reference electrode was the same saturated mercury (I) sulfate electrode (Hg/Hg 2 SO 4 ), SSE.The experiments were carried out at room temperature with a stirring speed of 300 rpm.

Electrodeposition experiments
The electrodeposition experiments were carried out in triplicate series, using the same potentiostat/galvanostat and the electrochemical cell earlier described in topic 2.2, as well as the same cleaning procedures for both the work and the counter electrodes used.Based on the polarization curves, four values of I (10 A m -2 , 20 A m -2 , 40 A m -2 and 80 A m -2 ) were selected to produce the Co-rich Zn-Co alloys on a steel substrate from the Bath 5 of Table 1 at the same stirring speed (300 rpm).Each electrodeposition time was calculated, based on the Faraday's law, to produce a 10 mg coating.The conditions for the electrodeposition experiments are shown in Table 2.
The produced layers were dissolved in 20 %v/v HNO 3, and the alloy composition was determined by flame atomic absorption spectrometry (FAAS), applying the conditions recommended by the instrument operation manual (AANALYST 300).These results and the theoretical mass obtained from Faraday's law were used to calculate the E f value for each deposition time 18,25,26 .
The content of each metal in the coatings was calculated as the % m/m Co and % m/m Zn, using Equations 1 and 2, respectively.
The cobalt content in the electrodeposition bath (Bath 5 of Table 1) was also calculated using Equation 1.The metallic contents of the electrodeposited coatings were compared with these results.
Based on the electrodeposition experiments, partial polarization curves were obtained by calculating the effective currents densities used for zinc and cobalt deposition (I Co and I Zn , respectively) from the element contents in the alloy coating and the current efficiency.In each case, the partial currents were associated with the corresponding potential response of the total applied current density.The effective

Morphological analysis
Scanning electron microscopy (SEM) was performed using a ZEISS EVO MA-10 microscope to evaluate the effect of the deposition parameters on the surface morphology of the coatings.The samples were cleaned with alcohol, dried and adapted to the stub with a conductive tape.The analyses were carried out in high vacuum and in the secondary electron mode.An X-rays energy dispersive spectrometer, EDS (EDS IFRX), coupled to the microscope, was used for elemental mapping at selected regions of the sample.

Microstructure characterization
X-rays diffractograms (XRD) of the coatings were obtained using a diffractometer (Rigaku Miniflex II) with a copper source (CuK = 1.5406Å), at 40 kV.The 2θ ranged from 10° to 120°, at a scanning rate of 0.050° s -1 .The more prominent diffractogram lines from the Zn-Co alloy deposit were fitted by a Gaussian equation (using Microcal Origin®, release 8.0) to obtain their 2θ values, which used to calculate the d (h k l) parameters of the diffraction lines.

Corrosion experiments
Zn-Co coatings were produced on steel substrate at chosen conditions, based on the chemical, microstructural, and morphological results earlier obtained.Polarization curves were performed in these selected coating/substrate systems employing a potentiostat/galvanostat AUTOLAB PGSTAT 302N.The working electrode (that is, each coating/ substrate system produced) was immersed in a 0.5 mol L -1 NaCl solution (pH = 7.06) and a linear voltametry was carried out from -2 V SCE to 1 V SCE , with a scanning rate of 1 mV s -1 , at room temperature.The counter electrode was a platinum net, while the reference electrode was a saturated calomel electrode (SCE).The corrosion current density values (I corr ) and the corrosion potential (E corr ) were obtained by Tafel extrapolation.

Cathodic linear voltametric curves
The voltametric experiments aimed to study the effect of the citrate ligand on the deposition process of zinc and cobalt ions.Citrate is present in the baths as a ligand, which may influence on the alloy deposition process.Additionally, the differences among the deposition processes of the Zn-Co alloy from a citrate bath and those from the separated metals ions in similar baths were also evaluated.
Figure 1 shows the differences in the cathodic scans among the solutions containing one of the ions (Co (II) or Zn (II)), with or without citrate.During the scans towards negative potentials, the cathodic current of the processes obtained in baths without citrate (Solutions 1 and 2 of Table 1) increased sharply once cobalt or zinc nucleation began (Figure 1a).For Solution 1, containing only Co (II), the cathodic current density reached a maximum value at approximately -1.211 V.The peak formation was then followed by a slow decay to a steady state.This behavior corresponds to a diffusion controlled process.On the other hand, the deposition peak was not very well resolved for Solution 2 (containing only Zn (II)), which suggests the existence of other processes in addition to mass control.The analysis of the isolated curves shows that the deposition of cobalt starts at more positive potentials, confirming that Co (II) behaves as the noblest ion.
The stability constants for Zn-Citrate complex and Co-citrate complex 27 are, respectively, K f Zn = 4.45 X 10 7 and K f Co = 6.76 x 10 4 , suggesting that the reduction of both metallic ions would be more difficult in a bath containing sodium citrate.However, the potential where a significant variation in the cathodic current density was noted (which was considered the starting deposition potential, E sd ), was  1.
shifted to more positive values when sodium citrate was added to the baths (Solutions 3 and 4), as shown in both Figure 1b and Table 3.This result may indicate that in the conditions of the present experiments, the reduction of Co-Citrate or Zn-Citrate complexes could be energetically more favorable than their respective aquo complexes.Similar to what has been observed for the metallic aquo complexes, the potentials for the reduction of Co-Citrate complexes are considerably more positive than those for the Zn-Citrate ones at the same conditions.This result was expected based on their respective stability constant values and suggests that if a bath containing both ions and the citrate ligand was used, the alloy deposition process would be a normal one and Co (II) ions would reduce preferentially.
Table 3 also shows that the cathodic peak potentials (E cp ) are at more negative values than the E sd ones.Although the E sd cannot be considered the E eq of a M n+ /M reduction pair in a foreign substrate, it is possible to suppose that this difference can be related to the fact that E cp may involve not only the thermodynamic equilibrium potential but also the potential required for the deposition of metals onto a foreign substrate 28 .The differences between E cp and E sd for the studied deposition processes are also presented in Table 3 and it is possible to observe that the total potential required for the electrodeposition of the metallic citrate complexes are more negative than that obtained for the metallic aquo complexes.The differences observed for each studied solution indicate that the steps involved in the electrocrystallization process of both zinc and cobalt ions on graphite surface require different energies depending on the coordination sphere of the ionic species.Therefore, based on the values shown in Table 3, although the reduction of the metallic ions from their citrate complexes started at more positive potential, the overall energies needed for deposit cobalt or zinc from a Co-Citrate or Zn-Citrate complexes are higher than those needed to produce these metallic layers from their respective aquo complexes.
Figure 1b also presents the cathodic linear voltametric curves of the graphite substrate in the solution 5 of Table 1, aiming to compare the isolated processes of metal reduction from a citrate complex with the reduction process when both metallic ions are in the same citrate bath.It is important to point out that the simultaneous discharge of different ions on the cathode is not a simple process, and it can be influenced by several factors, such as the substrate surface, the current density or potential applied, as well as the changes in the structure and activity of each metal cation in the double layer 29,30 .
It is possible to observe that the E sd value for the alloy reduction occurred at potentials more negative than those of zinc and cobalt reduction, separately.A small shift in the applied potential to more negative values caused a rapid increase in the current density and the alloy curve cross both cobalt and zinc curves.This result may indicate that, at the conditions of these experiments, the content of the metal in the coating and the kind of deposition process (normal or anomalous), would depend on the applied potential or current density, which is in agreement with the results of Roventi et al. 11 and Kirilova et al. 12 .
Rashwan et al. 7 showed that, for the steel polarization in a Zn-Co electrolyte containing glycine 1 mol L -1 , the curve of the alloy lied between those of isolated metals, indicating that glycine electrolyte was able to brought the reduction potentials of the metallic ions together and the nobler metal (cobalt) deposited at less negative potentials, while the less noble metal (zinc) co-deposited at more negative potentials 7,31,32 .Therefore, normal co-deposition would take place, independent of the current density or potential applied.However, these authors have also shown that the expected behavior was not verified indeed when the alloys were produced and anomalous deposition was always observed.These results show the limitation of using only these experiments to verify the deposition process of metallic alloy coatings.

Cathodic galvanostatic polarization curves
Figure 2 presents the cathodic polarization curve of the steel substrate in Bath 5, earlier described in Table 1.The main goal of this experiment was to select the values of I to be used in the electrodeposition of the alloy, which were: 10 A m -2 , 20 A m -2 , 40 A m -2 and 80 A m -2 .
At smaller current densities, the zinc deposition in a Zn-Co alloy is generally preferential, as a result of the zinc   hydroxide oscillation 1,3 .Therefore, the increase in current density may favor the cobalt deposition.Similar results were shown by Karahan et al. 1 and Roventi et al. 11 in the deposition of Zn-Co alloys from a bath containing sodium citrate.

Electrodeposition experiments
As the aim of this work was to produce Co-rich coatings, six values of I were selected, based on from the polarization curve of Figure 2, to perform the electrodeposition of the Zn-Co alloy on steel substrate, from Bath 5 of Table 1.However, the coatings electrodeposited at I values smaller than 10 A m -2 were not adherent to the steel substrate.Therefore, only four values of I (10 A m -2 , 20 A m -2 , 40 A m -2 and 80 A m -2 ) were considered in this work.The colors of the produced coatings varied from light to dark gray and from opaque to bright, as a function of I.
The E f values were always low, ranging from 23 + 3 to 42 + 1 %.The maximum E f value was obtained for I = 40 A m -2 , as shown in Figure 3. Since E f is related to the content of metals deposited, these results indicated that other parallel reactions might have been competing with the main reactions (the metallic ions reduction), decreasing the values of E f .The most probable parallel reaction is the hydrogen evolution reaction (HER), which may have consumed part of the applied I causing the low E f values.Both Gharahcheshmeh and Sohi 8 and Kirilova et al. 12 have reported the dependence of % m/m Co on I, even though none of these authors have obtained normal deposition.In fact, it is usually found in the literature reports concerning the anomalous nature of Zn-Co alloy electrodeposition process, both in the absence and in the presence of complexant agents and/or additives 2,6,7,33,34 .It is important to note, however, that most of these works used The presence of a complexant agent may also interfere on the alloy deposition process if one of the ions in solution is more intensively complexed than the other.Citrate is a tri-carboxylate anion and it usually forms stable metallic complexes, which are shifted to smaller wavelength length, when analyzed by spectrophotometry 36 .Since the stability constant for Zn-Citrate complex is almost three orders of magnitude higher than that for the Co-citrate complex 27 , it is expected that most of the Zn (II) ions were preferentially as Zn-citrate complexes in the solution, while most of the Co (II) ions could be reduced directly from the aquo complexes.Similar results were found for Survila et al. 37 for Sn-Co alloys in citrate baths at similar conditions.Szczygieł et al.Even though the complexant concentration in Bath 5 was enough to complex all the Zn (II), the condition I = 10 A m -2 still stimulated zinc deposition.Therefore, it  is important to point out that both the presence of a ligand and its concentration, as well as the I value used must be considered when studying the conditions to obtain normal or anomalous Zn-Co deposition process, or even Co-rich coatings.
At pH value of solution 5 (Table 1), the hydrogen evolution reaction (HER) is the main parallel reaction in the process.Since H + ions from water were reduced, the interface metal/solution was enriched of OH -, which could have enhanced the zinc deposition as hydroxide, originating the zinc hydroxide oscillation process and, consequently, the observed anomalous deposition.Even though this kind of deposition process was observed at 10 A m -2 , the coatings were Co-rich ones, presenting approximately 30 % m/m Co.Therefore, another highlight that could explain the high amount of Co obtained, even at 10 A m -2 , and the normal deposition process in Zn-Co citrate baths at higher I values would be the buffering effect of citrate that could have contributed to decrease the formation of zinc hydroxide, which is considered the responsible for the oscillatory zinc hydroxide process 8 .It means that the zinc deposition was hindered, while cobalt deposition was enhanced in the presence of citrate.Similar results were obtained by Trejo et al. 18 using benzidilacetone in the bath.However, new experiments concerning the local pH measurements on the electrode/solution interface are still needed to reach a final conclusion about this topic.
Using E f and the contents of the elements in the alloy (% m/m Co and % m/m Zn) data, it was possible to calculate the partial current densities of each one of the metals (I Zn , I Co ) and of the HER process (I H ). It is possible to observe, in Figure 5, that the transition from anomalous to normal deposition occurred with the increasing I, as earlier seen in Figure 4.Moreover, a strong decrease for the I Zn curve can be observed at high values of I.These results confirm that the increase in I promoted the deposition of cobalt in the alloy.
It is interesting to note that, comparing to the I Zn and I Co curves, the I H curve is always the closest one to the total I curve.This result justifies the low values of E f presented in Figure 3, and it means that most of applied I was used by the HER process.Even though, it is important to remember that both metals could have also be deposited as hydroxides under these conditions, and that the I H curve was obtained by the difference between the values of I Zn and I Co and may not be entirely conclusive.

Characterization of selected coatings
Based on the results obtained for the metal contents (% m/m Co and % m/m Zn) the coatings produced from experiments 1 (30.25 % m/m Co and 69.75 % % m/m Zn) and 4 (92.38 % m/m Co and 7.32 % m/m Zn) of Table 2, were selected to be electrochemically, morphologically, and microstructurally characterized.These experiments represent conditions in which Co-rich coatings containing % m/m Co < % m/m Zn and % m/m Co > % m/m Zn, respectively, were produced.The distribution of the elements in the coatings is shown by the EDS analysis (Figure 6) and corroborates the results previously presented in Figure 4.

Morphological and microstructural evaluation
Figure 7 shows the surface morphology of the Zn-Co coatings produced from the experiments 1 and 4 (Table 2).
From the E f results, it is possible to suppose that thin coatings were produced for both I values studied.Comparing the surface morphologies of the coatings obtained from experiments 1 (Figure 7a) and 4 (Figure 7b), it can be noted that an increase in I, and consequently, in % m/m Co as shown in Figure 4, resulted in a decrease of the grain size.At lower I values, heterogeneous surfaces with several clumps and irregular grains were obtained, while fine grains and higher surface homogeneity was observed from the samples produced from experiment 4. Similar results have been found in literature for Zn-Co and Zn-Ni coatings 9,38 .These results show that the variable I actually interfered in both the composition and the morphology of the deposited layers.
The diffractograms of the coatings produced under the selected conditions (Figure 8a) presented relatively sharp diffraction lines.However, an amplification of the diffratograms (Figure 8b) show broad and low intensity peaks, which can be explained by the low thickness of the coatings and the small particle diameters (26 nm and 22 nm, for experiments 1 and 4, respectively).In fact, several authors have showed that XRD analysis is not always able to identify the deposited Zn-Co or Zn-Ni alloys phases because of the presence of a large amount of crystallites too small to be resolved by this technique 9,14,[38][39][40] .
It is important to point out that most of the electrolytically obtained alloys, produced from different baths, usually consist of fine crystals, non-uniform in composition and characterized by a considerable distortion of the crystal lattice, originated during the formation of the nonequilibrium phases at the cathode, which occurs at high overpotential values.As a result, it is normally difficult to analyze these alloy deposits by X-rays examination and, in some cases, phases present in a given deposit can be only partially revealed 29,39 .In the present work, the coatings were obtained at high current density values and, consequently, at high overpotentials.Therefore, it is possible to relate any lattice distortion to the production of non-equilibrium phases.Lattice distortion can be responsible for non-homogeneous microscopic residual stress and can significantly contribute to the coating microhardness.Moreover, all coatings were produced at room temperature and the thermal mismatch could also contribute to macroscopic residual stress as well as to the microhardness values of the electrodeposited layers [39][40][41] .In some cases, non cataloged crystalline phases were obtained 9,14,41 .
Table 4 presents the experimental values obtained for 2θ and d of the main lines observed for both experiments.Since some phases observed in the electrodeposited Zn-Co alloys were not identified by the available XRD pattern databases 42 , it was also used the literature references 10,43,44 to identify the experimental lines obtained from both experiments.
There most difference between the diffractograms of the coatings produced from experiments 1 and 4 is the intensity of the diffraction lines.The low intensities observed for experiment 4 agrees with the small grain size verified for this experiment (Figure 7b).The more intense diffraction lines (Figure 8a) are seen in the coating produced from experiment 1 and are correspondent with the Fe (110) line at 2θ = 44.60°(datasheet no.06-0696) and the γ-Zn 21 Co 5 (330) line at 2θ = 43.14°[ 10,43].Small diffraction lines, corresponding to Zn phases (datasheet no.04-0831) were also verified.The γ-Zn 21 Co 5 (330) line was smaller and broader in the experiment 4, confirming that a coating with % m/m Co >> % m/m Zn and small grains was produced.
The presence of γ-Zn 21 Co 5 was reported in coatings produced at high current densities, when the deposition potential becomes more negative 3 .Rashwan et al. 7 observed that, at lower current densities, the coating consisted of a cubic structure of γ-Zn 21 Co 5 , while in higher values of current density, the coating were composed of both a cubic phase of γ-Zn 21 Co 5 and another phase γ 2 -Zn 13 Co with monoclinic structure.Deposits with % m/m Co above 3 % m/m show a diffraction peak that is associated with the formation of the γphases 45 .In addition, these phases have been earlier observed in Co-rich coatings containing from 10 % m/m to 15 %m/m 10,43,46 .In the present work, the γ 2 -Zn 13 Co phase was not verified.These results indicate that deposition conditions that enhance cobalt deposition may produce γ-ZnCo phases phases However, this γ-phase diffraction line was significantly decreased in the coating containing almost 92 % m/m Co.A magnification of the diffractograms presented in Figure 8a ranging from 2θ = 40° to 2θ = 47° (Figure 8b), shows   .This result suggests that a Co segregated phase was preferable obtained under the conditions of experiment 4.

Corrosion experiments
The I corr , corrosion rate (C.R.), and E corr of the selected carbon steel/alloy coating systems in a 0.5 mol L -1 NaCl solution are presented in Table 5.It is possible to observe that there are significant differences among the coatings produced from experiments 1 and 4. The E corr of the coating produced in experiment 4 was shifted to more negative values, when compared to that produced at the conditions of experiment 1 (% m/m Co < % m/m Zn).It was expected that the increase in the Co codeposition in a Zn-Co alloy changes the coating protection process from a sacrificial coating to a protective one 3 .However, the E corr values of both studied coating/substrate systems were more negative than that of naked steel in the same medium (-0.583V SCE ), showing that all the coatings produced behave as sacrificial coating.It could be explained by the small thickness of the coatings and for the Zn-rich phases observed by XRD analysis, due to the segregation of the excess of cobalt.
Table 5 also shows that the smallest values of I corr and C.R. were obtained for the coatings where % m/m Co < % m/m Zn (experiment 1).It is known that the presence of cobalt in the Zn-Co coating increases the dissolution of zinc in a solution of NaCl, causing the reaction of the metal ion with Cl -to form zinc hydroxide chloride (ZHC) according to the following reaction 8 , ensuring higher protection ability for the Zn-Co coatings: 5Zn 2+ + 8OH -+ 2Cl -+ H 2 O → Zn 5 (OH) 8 Cl 2 .H 2 O Lichušina et al. 10 have shown that the increase in % m/m Co (15 -18 % m/m Co) produced Zn-Co alloy coatings with a corrosion resistance about threefold higher than the coatings with low cobalt content (1-3 % m/m Co).Bajat et al. 47 have found the smallest I corr (4.0 µA cm -2 ) for a coating containing approximately 18 % m/m Co, which is comparable to the results obtained for the coating containing ~ 30 % m/m Co (2.4 µA cm -2 , experiment 1).In addition, it is known that the γ -Zn 21 Co 5 alloy phase is characteristic of a high corrosion resistance coating and the absence of additional phases is favorable to the alloy stability 46 .In the present work, although both coatings presented the γ -Zn 21 Co 5 , the presence of a segregated Co-phase and the low thickness of the coatings may have contributed to the worse performance of the coating produced from experiment 4.

Conclusions
Both I and the presence of sodium citrate as a ligand influenced on the process of deposition of the alloy, affording adherent Zn-Co alloy coatings with different chemical compositions.The E f values were low in all cases.The I H values corroborated these findings by showing that the reduction reaction of H + was always priority over the deposition of metals.In addition, the action of citrate as a ligand may have caused an increase in the total energy to reduce both metallic ions.
I influenced the deposition process and affected significantly (p<0.05) the composition of the alloy.Low values of I stimulated anomalous deposition and the production of coatings with higher % m/m Zn while an increase of I promoted a normal type deposition and coatings with higher % m/m Co.
Although the coatings containing % m/m Co > % m/m Zn showed a homogeneous morphology with the smallest grain size, the best anticorrosive performance was obtained by the coating containing approximately 30 % m/m Co.It means that, beside the morphological characteristics, the presence of the γ -Zn 21 Co 5 alloy phase and no segregated phases must have contributed to the performance of the coating produced from experiment 1.

Figure 1 .
Figure 1.Cathodic scans of the solutions containing one of the ions (Co (II) or Zn (II)), without (a) or with (b) citrate.Figure 1b also shows the cathodic linear voltametric curve of the graphite substrate in solution 5 of Table1.

Figure 2 .
Figure 2. Cathodic galvanostatic polarization curves of the steel substrate in Bath 5 of Table1.

Figure 4
presents the results obtained for % m/m Zn and % m/m Co, for each value of I studied.The % m/m Co in solution was also calculated (47.43 % m/m) and it was considered as the base value of % m/m Co 5 .It can be noted that, except for I = 10 A m -2 , normal deposition processes occurred in the experiments performed with Bath 5. Most of the values of % m/m Co are always higher than that of the base value of % m/m Co, and also higher than those of % m/m Zn, showing that the deposition of the nobler metal was favored in the studied conditions and that Co-rich coatings were produced.There was also a significant trend (p<0.05) of increasing % m/m Co and decreasing % m/m Zn with I. Therefore, the deposition of cobalt was enhanced by this parameter.
2 , have also studied sodium citrate as a complexant agent for Zn-Co alloy deposition in a bath using the same metallic ion relationship as it was used in the present research ([Co 2+ ] = [Zn 2+ ]).Nonetheless, they observed anomalous deposition for Zn-Co alloys at high I values.It is important to point out that, differently from our work in which [Cit 3-] = 2 [Zn 2+ ], the amount of ligand used by these authors ([Zn 2+ ] = [Cit 3-]) was probably insufficient to complex all Zn (II) ions in the solution.

Figure 3 .
Figure 3. Average values of cathodic efficiency (E f ) of the deposited alloy on steel substrate.

Figure 4 .
Figure 4. Average values of zinc and cobalt contents in the coatings (% m/m Zn and % m/m Co) produced at different values of I.

Figure 5 .
Figure 5. Partial current densities of each one of the deposited metals (I Zn , I Co ) and of the HER process (I H ).

Figure 6 .
Figure 6.EDS analysis showing the distribution of the elements in the coatings selected from Table 2. Experiment 1: (a) Co and (b) Zn; Experiment 4: (c) Co and (d) Zn.

Figure 7 .
Figure 7. Surface morphology evaluation of the Zn-Co coatings produced from the selected experiments shown in Table 2. (a) Experiment 1; (b) Experiment 4.

Table 1 .
Chemical composition and pH values of the studied solutions.

Table 2 .
Experimental conditions to produce Zn-Co coatings from citrate bath.

I Co and I Zn .
.

Table 3 .
Voltametric parameters obtained from linear voltametric experiments of Co 2+ or Zn 2+ solutions, containing or not sodium citrate 0.10 mol L -1 , on graphite electrode.
*E sd -starting deposition potential; E cp -cathodic peak potential.

Table 4 .
d (h k l) values for the Zn-Co coatings produced from the selected experiments of Table2.

Table 5 .
Results of I corr , E corr , C.R., and R p of the selected carbon steel/alloy coating systems in a 0.5 mol L -1 NaCl solution.