z Electro, Physical & Theoretical Chemistry

The Influence of Ni(II) and Co(II) Adsorptions in theAnomalous Behavior of Co-Ni Alloys: Density FunctionalTheory and Experimental StudiesJorge Vazquez-Arenas+,*[a] Guadalupe Ramos-Sanchez+,*[a] Rene H. Lara,[b] Issis Romero-Ibarra,[c] M. Eng. Francisco Almazan,[a] and Luis Lartundo-Rojas[d]

The anomalous behavior arising during plating of Co-Ni alloyshas been extensively investigated, whence different qualitativeproposals have been suggested to describe it, although mostof them have been limited to capture underlying atomicinteractions between substrate and electroactive species. Thisstudy undertakes a different approach to account suchphenomenon based on density functional theory (DFT) calcu-lations supported on experimental data. Alloys formed exper-imentally consistently present an anomalous behavior, exceptat the most cathodic current. XRD, XPS and voltammetryconfirm the formation of solid solutions over alloy composi-tions from 40 to 90 wt% Co. SEM reveals that alloy morphologystrongly depends on applied current density, which likewise

affects Co content as output parameter. The effects of CoSO4

and NiSO4 (ion pairs) adsorptions on the anomalous behaviorare explained using DFT. More favorable adsorption freeenergies of CoSO4 are obtained on multiple alloy surfaces(different arrangements of 50-50 wt% Co-Ni) and pure metalsin comparison with NiSO4. Additionally, rich Co sites (substrate)enhance CoSO4 adsorption, while the specific solvation of thecation significantly contributes to the adsorption strength ofthe ion pairs, indicating that Co(II) reduction energeticallypossesses a definite advantage in alloy formation. Thesetheoretical findings provide strong evidence to explain theanomalous behavior of Co-Ni alloys through the NiSO4 andCoSO4 competitive adsorptions.

1. Introduction

Plating of Co–Ni alloys has been the motivation of multiplestudies due to their excellent magnetic properties, hardness,light weight, wear, versatility, abrasion and corrosion resistance,and particularly its prominent catalytic activity.[1] This method isattractive due to its low cost, flexibility (e.g., deposition assingle layer or multi-layer coatings on planar and non-planarsubstrates), efficiency, ease of high volume production, andcomposition modulation.[1d,2] As occurs with other iron-group

alloys (e.g. Fe–Ni, Fe–Co), the co-deposition is classified asanomalous since the mole fraction of the less noble component(i.e. Co) produced in the alloy becomes higher than the [Co2+

]/([Co2 +] + [Ni2+]) ratio in the solution.[3]

This behavior has spurred a considerable amount ofresearch, stemming different proposals to account for it: fasteradsorption rate of the less noble component (e.g. cobalthydroxide Co(OH)+

ads) in comparison with the more noblespecies (e.g. Ni(OH)+

ads),[4] enhanced surface coverage of the

less noble Me(I)ads intermediate,[5] formation of intermediateadsorbed species (i.e. NiCo(III)ads) which increase the cobaltcontent in the alloy and release Ni(II) species,[6] differences inelectronic structures and labilities of Co(II) and Ni(II) species onthe basis of ligand field theory,[7] and faster charge-transfer ofsingle Co(II) reduction in comparison with Ni(II) reduction.[8]

Additionally, it has been consistently reported that anomalousdeposition arises regardless of the electrolyte composition andthe presence of additives in the bath,[4b,7,9] although alloycomposition varies. Based on these findings, there are strongevidences to consider that the rate-controlling step describingthis anomaly needs to be evaluated at the atomistic scale, sinceit presumably relies on differences in electronic structures forCo and Ni atoms during reduction, rather than pH or solventeffects. Likewise, experimental and modelling investigationsdeveloped at the micro- and macroscopic scales have not beenable to comprehensively determine (i.e. isolating) and quantifysuch process. This has resulted from multiple limitations toevaluate kinetics of electrochemical reactions (e.g. surfaceconcentrations, inclusion of material properties), and the

[a] Dr. J. Vazquez-Arenas,+ Dr. G. Ramos-Sanchez,+ M. E. F. AlmazanDepartamento de Qu�micaUniversidad Aut�noma MetropolitanaIztapalapa, C.P. 09340 M�xico, D.F., M�xicoE-mail: jgva@xanum.uam.mx

jorge_gva@hotmail.comgramossa@conacyt.mx

[b] Dr. R. H. LaraFacultad de Ciencias Qu�micasUniversidad Ju�rez del Estado de Durango, Av. Veterinaria S/NCircuito Universitario, 34120, Durango, Dgo., M�xico

[c] Dr. I. Romero-IbarraUnidad Profesional Interdisciplinaria en Ingenier�a y Tecnolog�as Avanza-das- Instituto Polit�cnico Nacional. Av. IPN No. 2580Gustavo A. Madero, C.P. 07340, Ciudad de M�xico

[d] Dr. L. Lartundo-RojasInstituto Polit�cnico NacionalCentro de Nanociencias y Micro y Nanotecnolog�asUPALM, Zacatenco M�xico-D.F. 07738, M�xico

[+] CONACYT Research Fellow

Supporting information for this article is available on the WWW underhttp://dx.doi.org/10.1002/slct.201601957

Full PapersDOI: 10.1002/slct.201601957

1826ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

development of models which despite relying on first-princi-ples, incorporate semi-empirical parameters such as rateconstants which faintly integrate any conceivable physicochem-ical event related to activation barriers (e.g. adsorption, charge-transfer, desolvation, nucleation), in addition to multipleassumptions.

Ab initio calculations based on quantum mechanics aremore suitable methods to describe atomic interactionsbetween substrate and electroactive species (i.e. structure-reactivity relationship), determining the anomalous behavior ofCo-Ni alloys. To this concern, an important issue delimiting thetheoretical calculations is to define the atomistic processpresenting a significant contribution to the energy of theelectrochemical reaction (i.e. overpotential). Under the premisethat multiple electrocatalytic processes (e.g. oxygen reductionreaction, ORR and oxygen evolution reaction, OER) stronglydepend on the adsorption energy of electroactive species,[10]

and electrodeposition studies of iron-group alloys have stoodout the role of adsorption for electroactive species,[4-5] a firstanalysis is conducted to assess the effect of NiSO4 and CoSO4

adsorption energies upon the anomalous behavior of Co-Nialloy in sulfate media; assuming that other reaction barriers,namely other contributions to activation overpotential aresmaller. To our current state of knowledge, none theoreticalapproach has been undertaken to analyze this anomaly atatomistic scale. Theoretical calculations using density functionaltheory (DFT) are supported on experimental studies carried outwith X-ray diffraction (XRD), scanning electron microscopy(SEM), and X-ray photoelectron spectroscopy (XPS). Alloycomposition, current efficiency and morphology are firstlyanalyzed as a function of current density, subsequently; thecrystal structure of each Co-Ni alloy is experimentally charac-terized to determine whether a solid solution or co-depositionwas involved during plating. The alloy structure with composi-tion around ~50 wt% of Ni (~50 wt% of Co) was then selectedto perform density functional theory (DFT) calculations aimingto account for the adsorption energetics of NiSO4 and CoSO4

species on substrate.

2. Results

2.1 Electrochemical plating

Table 1 presents the experimental conditions used to form thealloys with direct current (dc) techniques (S1-S6). Note that theapplied current densities (-0.005 to -0.1 A cm-2, or -5 to -100 mAcm-2) are within the typical range used to deposit alloyscontaining Iron-group metals. Stationary potentials at the endof electrolysis have been included in this table. As observed inthese experiments, the nickel content in the alloy increases ascurrent density becomes more cathodic, which has beenconsistently reported for other electrolytes.[8a] From an exper-imental standpoint, this behavior results from the faster Co(II)deposition which generates a mass-transfer control (i.e. Co(II)depletion on electrode) at lower current densities if comparedto nickel deposition. In general, it has been described that thecomposition of the more noble metal is augmented as the

current is made more negative in anomalous deposition, unlikewhat happens in regular plating.[3a] This phenomenon can beconfirmed if the cathodic partial current densities (ji) of Ni(II),Co(II) and H+ reductions are plotted as a function of currentdensity (Figure 1). These values can be readily calculated usingFaraday ’s law:

jNi ¼zNi mNi FA MNi t

ð5Þ

jCo ¼zCo mCo F

A MCo tð6Þ

jH2¼ j � jNi � jCo ð7Þ

where zi is the number of electrons transferred (2) to metal i(Ni, Co) during plating, mi is the mass of species i depositeddetermined from atomic absorption, F is the Faraday constant(96485.5 C mol-1), Mi is the molecular weight of metal i, t is theelectrolysis time and A is the geometric surface area ofelectrode. Note that jH2

(partial current density associated toHER, H+ and H2O reduction) is obtained from the difference

Table 1. Parameters used to perform the dc plating in electrolytescontaining 0.1 M NiSO4, 0.1 M CoSO4 and 1 M Na2SO4 at pH 3. Ni contentwas determined using atomic absorption as described in “Chemical and

textural analyses” (supporting information).

Experiment Applied currentdensity [A m-2]

Stationary Po-tentials [V vsSHE]

Ni con-tent[wt%]

Current ef-ficiency[%]

S1 -50 -0.63 13.14 80.00S2 -100 -0.70 20.81 89.21S3 -250 -0.94 37.68 85.74S4 -500 -1.04 48.82 71.43S5 -750 -1.18 49.87 65.30S6 -1000 -1.44 57.31 70.34

Figure 1. Partial current densities of Ni(II), Co(II) and H+ reduction calculatedfor the plating experiments as a function of dc current density.

Full Papers

1827ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

between the total current density and the sum of jNi and jCo. Asobserved, jNi (-5.27 A m-2) contributions are lower than jCo

(-34.72 A m-2) at the lowest current density used in this study(-50 A m-2). Not surprisingly, this behavior results from the fastercobalt deposition generating the anomalous behavior of Co-Nialloy. Once that Co(II) concentration becomes rapidly depletedaround -500 A m-2 (jCo = 182.42 A m-2, jNi = 174.72 A m-2), thenickel content approaches very close cobalt composition in thealloy (48.82 wt% Ni), and the slope of jH2

turns out to besteeper as a result of the onset of H2O reduction (jH2 = 142.86).However, this increase of nickel composition with morecathodic current densities is not accurately the fading of theanomalous behavior (i.e. regular deposition) entailing anincrease in lability of Ni(II) reduction, but the depletion of Co(II)on electrode surface as a result of faster reduction and aneffective competition for electrons with H+ and H2O due toHER. Evidence of such behavior is observed in the lowestcurrent efficiencies obtained under dc conditions for samplesS5 and S6 in Table 1, consistently with the point that highcurrent efficiencies above 80 % are only obtained when Nicomposition is below 40 %wt. As a matter of fact, pulse platingand pulse reverse along with chloride media are the onlyeffective routes to suppress anomalous behavior at highcurrent efficiencies.[8c,11] Below, X-ray diffraction and XPS analysisare conducted to accurately determine alloy composition andstructure, while SEM is mainly utilized to study the morpho-logical features of the deposits. This information is then used toconstruct slab models (e.g. Co, Ni, Co-Ni) in order totheoretically analyze the importance of competitive adsorptionbetween NiSO4 and CoSO4 species on the anomalous behavior.

2.2 X-ray diffraction and X-ray photoelectron spectroscopy

Figure 2 shows X-ray diffraction patterns obtained for pure Niand Co metals, and Co-Ni alloys S1 (-50 A m-2), S5 (-750 A m-2)

and S6 (-1000 A m-2) (refer to Table 1). The Ni sample showsstrong diffraction peaks at 44.58 and 51.848 fitting to Ni PDFcard file 00–004-0850, while Co presents diffraction peaks at 2q

values of 41.588, 44.68, 47.38 and 62.738 (not shown) Co PDF file00–001-1278). The sample S1 is described by a noisy spectrumwhere four small peaks are barely distinguished. The peaks in~41.588, ~44.68 are associated to Co phase, while two peaks at~43.48 and ~50.68 are associated elemental fcc coppercrystalline phase, which has been used as substrate for thedeposits. These peaks were determined using the standardpowder diffraction card PDF, copper file 04–009-2090. Note thatthe Ni content is low for this alloy (13 wt%), in addition thepeaks at 44.58 (Ni) and 44.68 (Co) most likely overlap wherebythese effects could mask the presence of Ni, since no otherreflection appears for this metal in S1. Other explanationaccounting for the absence of peaks for Ni in terms of crystalstructure is the formation of a solid solution (alloy), which issupported by the lack of reflections at 47.38 (Co) and 51.848(Ni). In sample S5, three intense peaks are observed, those at~43.48 and ~50.68 are related to copper crystalline phase; whilethe peak at ~ 44.368 could be connected with the deposition ofNi or Co metals. Nevertheless, the absence of peaks at 47.38and 51.848 indicates a solid solution for S5. This sample is 49.9wt% of Ni and was indexed to PDF 04-004-8490 thatsatisfactorily fits to the ~0.5 nickel and ~0.5 cobalt alloycontents. On the other hand, four peaks were recorded for S6.Similar to S5 sample, peaks at ~43.48 and ~50.68 of belongs tocopper phase, while peaks at ~ 44.368 and ~51.68 (Ni) areassociated with metal deposition. Other than 44.368, nonesignal of Co was observed for S6, confirming a mixture of Coand Ni phases in a solid solution. Multiple studies concerningCo-Ni electrodeposition under different experimental condi-

tions (electrolyte, applied potential, bulk concentration) andcharacterized using XRD have also concluded that a solidsolution is formed over the whole composition range.[12]

Further chemical analysis were performed for samples S1,S4 and S6 using XPS to corroborate whether a co-deposition ora solid solution forms during plating of Co-Ni alloys around 50wt% Ni (~50 wt% Co) composition. To this concern, averageXPS survey spectra were used to quantify the elementalcomposition of the Co-Ni alloys. These data (wt %) confirm thatNi content has been adequately determined with atomicabsorption (Table 1) and X-ray diffraction: 14.9 Ni, 81.1 Co, 2.2 Oand 1.8 C (S1), and 51.0 Ni, 48.5 Co, 0.2 O and 0.3 C (S4). Notethe low oxygen content in these samples despite their currentefficiencies. Although alloy composition has been determinedusing atomic absorption, XRD and Energy Dispersive X-raySpectroscopy (see below), XPS analysis is a more powerfultechnique to determine alloy chemical species compositionsince it accounts for core level photoemission of electrons,which is not possible with the other methods. Additionally, XPScan detect different oxidation states of same element, atconcentrations in the range of 1000 to 5000 parts per million(ppm), approximately one to two orders of magnitude moresensitive than EDS or DRX techniques.

Additionally, and taking in consideration that a homoge-neous solid solution is characterized by the absence of

Figure 2. X-ray diffraction patterns obtained for pure Co and Ni, and alloys(S1, S5 and S6) deposited under the experimental conditions shown inTable 1.

Full Papers

1828ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

segregation effects and/or presence of different chemicalspecies, a comparative normalized XPS core level spectra ofNi2p an Co2p regions for the Ni, Co and Co-Ni alloys werecollected and are presented respectively, in Figs. 3a and 3b. Thespectra are normalized to the maximum intensity in order toappreciate possible differences in the position and shape of thespectra, and satellite features shown in this figure. Whencomparing the spectra of all Co-Ni alloys with their respectivemetals, all core level spectra are very similar in shape. Likewise,Ni2p characteristic satellite features associated with metallicspecies are present (Figure 3a), as well as Co2p loss feature and

Co LMM Auger are also distinguishable both in Co and Co-Nialloys (Figure 3b). Furthermore, a shift of ~ 0.1 eV between thepure Ni metal and the Co-Ni alloys core level spectra wasdetermined, which agrees with the 852.7 � 0.2 eV bindingenergy position proposed by Turner et al. for a Co-Ni alloy.[13]

On the other hand, a binding energy value of 778.5 eV wasobtained for the main core level Co2p peak of all Co-Ni alloys,which corresponds to a displacement of ~ 0.3 eV with respectto Co pure metal core level Co2p peak, 778.2 � 0.2 eV.[13] Theseenergy variations seemingly indicate a slight change in the Cocrystal lattice due to metallic Ni presence. Thus, confirming thata solid solution is obtained for the whole composition range ofalloys deposited in sulfate media, as previously reported in the

literature using other characterization techniques.[12,14] Furtherconfirmation of this behavior was corroborated by analyzingthe anodic branch of a cyclic voltammogram conducted in theelectrolyte bath (Figure S2, supporting information), whereconsistently a single wide peak was obtained during thebackward scan (alloy oxidation), carrying out the concomitantdissolution of both alloy components, previously electrodepos-ited.

2.3 Scanning electron microscopy

The microstructure of the Co–Ni alloys (S1-S6) was analyzedwith scanning electron microscopy to draw the influence ofplating parameters on this property. Figure 4 shows SEM

micrographs of S1 sample (Table 1) acquired with differentdetectors and magnifications. In Figure 4a, an image withoutimportant differences in contrast between the different grainsis observed using the LABe detector. Due the nature ofbackscattered electrons (high energies), LABE micrographshave great Z-sensitive contrast image. This result suggests ahomogeneous distribution of Ni and Co in the film, as indicatedfrom XRD results. Figure 4b corresponding to secondaryelectron images shows small grains regularly distributed withedges and contours well defined. As secondary electrons aremore sensitive to the surface, this image indicates a homoge-

Figure 3. Normalized core-level Ni2p XPS spectra collected for a) Ni puremetal, S1, S4 and S6 samples, and normalized core-level Co2p XPS spectrafor b) Co pure metal, S1, S4 and S6 samples. .

Figure 4. SEM images of Co–Ni alloys deposited under dc conditions using acurrent density of -50 A m-2 (S1), acquired with different detectors: (a) lowangle backscatter (LABe), (b) upper (SEI), (c) lower (LEI) and (d) higher scaleof LEI. Images (e) and (f) show a magnification of (a) and (b).

Full Papers

1829ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

neous distribution of grain clearly delimitated. Figures 4c and4d (LEI mode micrographs) show an uniform surface withoutdefects, holes and impurity incorporation during plating. Thisimportant result can be associated to the absence of H2Oreduction in HER (refer to current efficiency in Table 1). On theother hand, Figures 4e and 4f (LABe and SEI images, respec-tively) display magnifications of Figures 4a and 4b. In theseimages, grains with pyramidal-like edges and homogenous sizedistribution lower than 0.5 mm are observed. Once again, theabsence of contrast in the LABe image (Figure 4e) indicates ahomogenous distribution of Ni and Co in the film at roomtemperature. According to metal content, the alloy showeduniformity and homogeneity on surface composition at thiscurrent density.

Figure 5 shows SEM-LABe images of Co-Ni alloys formed atdifferent current densities: (a) -50 (S1), (b) -500 (S4), (c) -750 (S5)

and (d) -1000 A m-2 (S6). As observed, the crystal structure,particle shape and size of the deposited Co-Ni alloys consid-erably depend on current density, as a result of variations inalloy composition. Morphology of S1 sample was previouslydescribed as pyramidal-like edged particles, corresponding tocobalt-rich deposits (13 wt% Ni). When current densitybecomes more cathodic at -500 and -750 A m-2 (Figs. 5b and5c, respectively), a bimodal particle morphology is observed,consisting of semi-spherical grains with sizes lower than ~0.5mm and elongated particles of diameter equal ~1 mm.Particularly, the elongated particles (as dendrite-like) obtainedat -750 A m-2 exhibit a smooth surface. Large grains areobtained for these current densities, and edges are smoothedin comparison with the dc plating conducted at -50 A m-2 (S1).The appearance of pyramidal-shaped grains is not observedunder these conditions. At -1000 A m-2 (Figure 5d), the grainsare spherical, coarser, denser and more packed yielding a morecompact structure, although some holes appear all over itssurface and the elongated shape vanishes. Accordingly, it is

suggested that high deposit rates promote high defectdensities and impurity incorporation, which can cause thedecrease of alloy grain size. Although the nickel content in thealloy plays an important role to refine the grain size, theparasitic reaction (HER) also exerts some influence in themorphology of the alloy as described in Figure 5, where H2Oreduction becomes dominant and H+ reduction enters themass-transport control. The fact that S1 sample only displayedone particle shape (i.e. pyramidal-like), while S4 and S5exhibited bimodal particle, similar to S6 but with smalleragglomerates (inset of Figure 5d), is the consequence of theapplied current density in the deposit morphology. Under thepremise that the Co deposit rate is faster than the Ni rate(anomalous behavior), it could be assumed that the micro-structure of electrodeposited Co-Ni alloys depends on thecobalt content, since the inhibition of this metal enriches the Nicomposition. Then, the crystal structure, particle shape and sizeof the deposited Co-Ni alloys considerably rely on currentdensity, as result of variations in alloy composition. Multiplestudies agree with this dependence.[12c,14-15]

Energy Dispersive Spectroscopy (EDS) performed for S6sample indicated a Ni content of ~58 wt%, which agrees withthe Ni percentage (57.31 wt%) reported in Table 1. Likewise,EDS measurements conducted for the other samples showedgood concordance with the alloy compositions estimated withatomic absorption and XPS.

2.4 DFT analysis of NiSO4 and CoSO4 adsorptions

As above described, the Co-Ni alloys electrodeposited in thiswork in sulfate media at different current densities formed solidsolutions over the whole composition range, involving ananomalous behavior, and in agreement with multiple studiesconducted in the literature. However, none theoretical ap-proach using ab initio methods has been undertaken to analyzethis anomaly at atomistic scale based on these findings.Particularly, these methods are powerful to account forstructure–reactivity relationships based on electronic structure(ground state). To this concern, a first analysis is conducted inthe present study to assess the effect of NiSO4 and CoSO4

adsorption energies upon the anomalous behavior of Co-Nialloys (~ 50 wt% Co - 50 wt% Ni); assuming that other reactionbarriers, namely other contributions to activation overpotentialare smaller. This idea fundamentally stemmed due to the factthat multiple electrocatalytic processes (e.g. ORR, OER) stronglyrely on the adsorption energy of electroactive species, [10] andplating studies of iron-group alloys stand out the role ofadsorption for electroactive species. [4-5] Thus, multiple slabscontaining Co (50 wt%) and Ni (50 wt%), pure Ni and Co havebeen used as models to attempt describing this process. Thesestructures were selected instead of metallic Cu (although asimilar result is expected for it), since early stages of alloynucleation last a couple of seconds, whence all the anomalousdeposition process is virtually conducted on multiple layers of asolid solution, while less likely scenarios encompass enrichedNi(111) or Co(111) regions. An additional assumption was madeto consider the substrate as monocrystalline, due to the

Figure 5. SEM micrographs of Co–Ni alloys deposited at different currentdensities: (a) -50 (S1), (b) -500 (S4), (c) -750 (S5) and (d) -1000 A m-2 (S6).

Full Papers

1830ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

difficulties involved to construct a model for polycrystallinematerials.

The nature of metallic ions in solution has been the topic ofseveral theoretical and experimental studies aiming to deter-mine the orientation and solvation of cations and anions,however no definite conclusions have been reached. [16]

Although electrode polarization and dehydration/desolvationphemomena are expected to play very important roles on theelectrochemical reactions, an ion pair without solvent andhydrated shell is considered in this first approach, assumingthat the adsorptive interactions arising between the ion pairand the surface are more important. Additionally, the electro-chemical reactions are expected to follow similar trends sincethe electroactive species and the solvent present a commoncounterion. This simplification is based on experimentalevidence combined with macroscopic modelling assigninggreat importance to the adsorption of electroactive speciesduring plating of Co-Ni alloys,[1d,4a, b,5-6] and the recognition thatmultiple electrocatalytic processes (e.g. oxygen reductionreaction, ORR and oxygen evolution reaction, OER) stronglydepend on the adsorption barriers of electroactive species.[10a,c]

First-principle modelling to account for electrode polarization,dehydration/desolvation and solvent effects are beyond thescope of the present work, but they can be the motivation offuture studies using for instance available source code mod-ifications implementing an implicit solvation model.[17] On thisstudy, the optimized structure of NiSO4 and CoSO4 on vacuumwere utilized as starting geometry for the interaction with thesurface, both of them were allowed to interact with pureNi(111) and Co(001) slabs, and on different sites of Co-Ni alloysconstructed according to details provided in “Computationaland system details” (supporting information) and Figure 6.

The adsorption free energy was calculated as the energydifference between interacting systems (adsorbate-slab), minusadsorbate (ion-pair) and slab (Ni, Co or Co-Ni), thus, a morenegative adsorption energy corresponds to a more feasibleprocess. Figure 7 presents the adsorption free energies of both

ion pairs on the different substrates. It is evident that higherenergies of CoSO4 on all analyzed alloy surfaces are obtained incomparison with NiSO4, indicating that in a competitivescenario CoSO4 will have a more favorable tendency over NiSO4

to adsorb on Co-Ni alloys. The interaction site strongly affectsthe adsorption process, since the CoSO4 adsorption is morefeasible when the site is richer in Co (i.e. surrounded by otherCo atoms). This dependence on the adsorption site is clearlyobserved for the Ni3 site, in which the difference of adsorptionfree energy among the two electroactive species is the lowest;in fact, the adsorption trend without ZPE and entropycorrections is reversed. On the contrary, the adsorption energydifference between NiSO4 and CoSO4 is the highest on Co3sites. These results highlight the importance of ZPE andvibrational entropy corrections, although the contribution tothe total adsorption free energy is quite small. When thedifference in adsorption energy is minor, the corrections canchange the energy trends, but only for one out of sixadsorption scenarios, the ZPE corrections were importantconcerning the adsorption free energy trends. The preferencefor Co adsorption remarkably increases as more Co rich sitesare formed, thus, perpetuating the Co enrichment as electrol-ysis is prolonged. Similar trends were found for NiSO4 andCoSO4 adsorptions on Ni(111) and Co(001) slabs (Figure 7), thus,indicating that anomalous behavior depends mainly on differ-ences in electronic structures for Co and Ni atoms, rather thanon effects of substrate. Additionally, this finding confirms thatrich local zones in any of these atoms will still favor theadsorption of the less noble metal, as suggested to occur onrecently deposited Ni.[15]

Figure 6. Methodology used to construct the models used to study theadsorption of CoSO4 and NiSO4 ion pairs on Ni, Co and Co-Ni alloy surfaces.Blue, gray, red and yellow spheres represent cobalt, nickel, oxygen and sulfuratoms. The CoSO4-1 and CoSO4-2 stands for the one and two-fold ion pairconfigurations. .

Figure 7. Adsorption free energies of NiSO4 and CoSO4 ion pairs on Co-Nialloys, pure Ni(111) and Co(001) corrected for zero-point energy andvibrational entropy. The inset shows the configuration of the ion pairinteracting with the metal substrate.

Full Papers

1831ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

To date, the importance of the local alloy structure onNiSO4 and CoSO4 adsorptions has been determined. However,the interaction arising between the cation and the counterion(i.e. the degree of cation solvation) might have an effect onelectrodeposition. In this preliminary study, the degree ofcation-counterion interaction can be changed by modifying thenumber of oxygen atoms of (SO4

-2) interacting with the cation.Therefore, in another set of simulations, the cation was allowedto interact with only one oxygen of SO4

-2, in this case, theadsorption trends are different to those previously discussedenabling a preferred adsorption of NiSO4 over CoSO4 on thealloy, except for Co rich sites (results not shown). It is worthmentioning that the one-fold interaction is not the most stableion-pair configuration; however, its use just serves for compar-ison purposes. These results indicate that presumably, most ofthe solvation sphere is conserved during charge transfer, asreported in single Co(II) reduction,[18] thus, providing strongevidence of the crucial effects of adsorption in electrodeposi-tion processes. In order to provide more convincing evidence,Bader charge[19] and charge accumulation/depletion featureswere performed on the systems presented in Figure 8, whereCoSO4 and NiSO4 are adsorbed in the two configurationsdepicted in Figure 6 on a NiCo2 site. Bader charges werecalculated on isolated systems and after adsorption, in order todetermine the partial reduction of the cation after this process.For the two-fold interaction, in which CoSO4 adsorption isfavorable, the partial reductions of Co and Ni after adsorptionare 0.51 and 0.31 (je- j /atom), respectively. On the other hand,Ni adsorption is more favorable for the one-fold interaction,

and partial reductions become 0.27 and 0.41 (je- j /atom) for Coand Ni, respectively. Furthermore, from a graphic point of view,the charge accumulation/depletion features after adsorption ofFigure 8 (blue/green areas), show mild changes. On the inter-face, charge depletion (green) is observed on the alloy surfaceatoms, and the charge depletion is higher on Co than in Ni ionpair (Figs. 8a and a’) for the two-fold ion pair configuration,while on the one-fold one, the charge depletion on the alloysurface is higher on Ni (Figure 8b’) than in Co (Figure 8b) ionpairs. Therefore, confirming that when charge donation to theadsorbed metal is higher, the adsorption free energy is alsohigher. On the SO4

-2 side, the electron accumulation/depletionfeatures are even more subtle; however, it can be observedthat Figs. 8b and 8b’ present more differences than Figs. 8aand 8a’, indicating that on the one fold ion pair configurationthe SO4

-2 group presents a higher effect on the charge donationcharacteristics. Thus, the situation in which the cation is able toaccept more charge strongly affects the degree of interactionwith the surface, and consequently its capacity to be readilyreduced. Hence, two components made up the increasedinteraction energy between the ion pair and the surface: thespecific nature of the surface, Co rich sites favor CoSO4

adsorption, and the capacity of the electroactive species toaccept charge either from the surface or the counter ion.

Conclusions

Co-Ni alloys were electrodeposited in electrolytes containing0.1 M NiSO4, 0.1 M CoSO4 and 1 M Na2SO4 (pH = 3) using

Figure 8. Charge accumulation/depletion (blue/green) isosurfaces after CoSO4 and NiSO4 adsorption on a Co2Ni site. a and a’ depicts Co and Ni ion pairs in atwo-fold configuration, respectively while b and b’ depicts Co and Ni ion pairs in a one-fold configuration, respectively. The isosurface (0.011 e-/�3) wascalculated from the electron density of the relaxed structure minus the separated slab and ion pair. Blue, gray, yellow and red spheres represents cobalt, nickel,sulfur and oxygen atoms.

Full Papers

1832ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

copper substrates. These alloys displayed an anomalousbehavior as consistently reported in the literature, except at themost cathodic current dominated by the depletion of Co(II) onelectrode surface and an effective competition for electronswith H+ and H2O (HER) reduction. Thus, Ni content increased inthe alloy as current was made more negative. Chemicalanalyses of alloys conducted with atomic absorption, X-raydiffraction, X-ray photoelectron spectroscopy and energydispersive spectroscopy enabled to determine similar composi-tions. Particularly, XRD, XPS and cyclic voltammetry confirmedthe formation of solid solutions over alloy compositionsranging from 40 to 90 wt% Co. The microstructure of Co-Nialloys strongly relied on applied current density as it occurs forthe cobalt content (refer to Table 1), low current densities witha high Co content exhibited one particle shape (edgedpyramidal-like), while the decrease of Co in the alloy favored abimodal particle morphology consisting of semi-sphericalgrains and elongated particles, which were refined at the mostnegative current. Thus, high deposit rates promoted highdefect densities and impurity incorporation, but also decreasedthe alloy grain size.

Density functional theory calculations were carried out toaccount for the effects of CoSO4 and NiSO4 adsorptions on theanomalous behavior. Multiple slabs were used as substrates,considering a 50-50 wt% Co-Ni alloy (multiple arrangements),pure Ni(111) and Co(001). More favorable adsorption freeenergies of CoSO4 were obtained for all alloy surfaces and puremetals in comparison with NiSO4, suggesting that in acompetitive scenario as in alloy formation, Co atoms possess adefinite advantage to deposit. The interaction site (adsorbate-substrate) strongly affected this process, since rich Co sites (i.e.surrounded by other Co atoms) enhanced CoSO4 adsorption,which could account for the perpetual enrichment of Co aselectrolysis is prolonged. The specific solvation of the cationwas also found to contribute to the adsorption strength of theion pairs, thus, suggesting that most of their solvation spheresare conserved during charge transfer in an anomalous Co-Nialloy. In summary, two components remarkably influenced theadsorption free energy between the ion pair and the surface,the specific nature of the surface, and the capacity of the cationto accept charge either from the surface or to the counter ion.These theoretical findings provide strong evidence to considersignificant the NiSO4 and CoSO4 competitive adsorptions in theanomalous behavior of Co-Ni alloys (sulfate media), andsupport experimental proposals reported in the literature,concerning a faster adsorption rate and an enhanced surfacecoverage of the less noble component (e.g. Co(II) species).

Supporting Information Summary

Materials and methods (Plating of Co-Ni alloys, Chemical andtextural analyses, Computational and system details), andadditional data.

Acknowledgments

Financial support is greatly appreciated from CONACYT (GrantsNo. 2013-205416 and 2014-237343). GRS and JVA also acknowl-edge the Catedras-CONACYT program via project No. 1456“Diseno y construccion de sistemas sustentables de generacion yalmacenamiento de energıa”. Partial economic support ofProject SIP-IPN 20160169 is also recognized. Computationalresources from “Laboratorio de Supercomputo y Visualizacionen Paralelo (LSVP) at UAM-Iztapalapa is gratefully acknowl-edged.

Conflict of Interest

The authors declare no conflict of interest.

Keywords: adsorption · anomalous behavior · Co-Ni alloys ·Density Functional Theory · electrodeposition

[1] a) F. A. Lowenheim, Modern Electroplating, 3rd ed., John Wiley & Sons,New York, 1974; b) N. Fenineche, C. Coddet, A. Saida, Surf. Coat. Technol.1990, 41, 75–81; c) L. Brossard, C. Messier, J. App. Electrochem. 1993, 23,379–386; d) D. Landolt, Electrochim. Acta 1994, 39, 1075–1090; e) P. C.Andricacos, N. Robertson, IBM Journal of Research and Development1998, 42, 671–680; f) M. Duch, J. Esteve, E. Gomez, R. Perez-Castillejos, E.Valles, J. Micromech. Microeng. 2002, 12, 400–405; g) M. Duch, J. Esteve,E. G�mez, R. P�rez-Castillejos, E. Vall�s, J. Electrochem. Soc. 2002, 149,C201–C208; h) D. Kim, D.-Y. Park, B. Y. Yoo, P. T.A. Sumodjo, N. V. Myung,Electrochim. Acta 2003, 48, 819–830; i) B. Chi, J. Li, X. Yang, Y. Gong, N.Wang, Int. J. Hydrogen Energy 2005, 30, 29–34.

[2] M. Datta, D. Landolt, Electrochim. Acta 2000, 45, 2535–2558.[3] a) A. Brenner, Electrodeposition of Alloys, Principles and Practice, Vol. 1,

Academic Press, New York, 1963; b) H. Dahms, I. M. Croll, J. Electrochem.Soc. 1965, 112, 771–775.

[4] a) S. Hessami, C. W. Tobias, J. Electrochem. Soc. 1989, 136, 3611–3616;b) K. Y. Sasaki, J. B. Talbot, J. Electrochem. Soc. 2000, 147, 189–197;c) W. C. Grande, J. B. Talbot, J. Electrochem. Soc. 1993, 140, 675; d) Y.-P.Lin, J. R. Selman, J. Electrochem. Soc. 1993, 140, 1299; e) A. Bai, C.-C. Hu,Electrochim. Acta 2002, 47, 3447–3456.

[5] M. Matlosz, J. Electrochem. Soc. 1993, 140, 2272–2279.[6] N. Zech, E. J. Podlaha, D. Landolt, J. Electrochem. Soc. 1999, 146, 2892–

2900.[7] D. Golodnitsky, N. V. Gudin, G. A. Voluanyuk, J. Electrochem. Soc. 2000,

147, 4156–4163.[8] a) J. Vazquez-Arenas, L. Altamirano-Garcia, T. Treeratanaphitak, M.

Pritzker, R. Luna-S�nchez, R. Cabrera-Sierra, Electrochim. Acta 2012, 65,234–243; b) J. Vazquez-Arenas, M. Pritzker, Electrochim. Acta 2012, 66,139–150; c) J. Vazquez-Arenas, M. Pritzker, J. Solid State Electrochem.2013, 17, 419–433.

[9] a) L. Altamirano-Garcia, J. Vazquez-Arenas, M. Pritzker, R. Luna-S�nchez,R. Cabrera-Sierra, J. Solid State Electrochem. 2015, 19, 423–433; b) W. E.Hansal, B. Tury, M. Halmdienst, M. L. Varsanyi, W. Kautek, ElectrochimicaActa 2006, 52, 1145–1151.

[10] a) J. K. Nørskov, J. Rossmeisl, A. Logdottir, L. Lindqvist, J. R. Kitchin, T.Bligaard, H. J�nsson, J. Phys. Chem. B 2004, 108, 17886–17892; b) A. B.Anderson, ECS Trans. 2010, 28, 1–17; c) A. B. Anderson, Phys. Chem.Chem. Phys. 2012, 14, 1330–1338.

[11] aJ. Vazquez-Arenas, T. Treeratanaphitak, M. Pritzker, Electrochim. Acta2012, 62, 63–72; bA. Bai, C.-C. Hu, Electrochim. Acta 2005, 50, 1335–1345.

[12] a) N. Myung, K. Nobe, J. Electrochem. Soc. 2001, 148, C136–C144; b) M.Duch, J. Esteve, E. G�mez, R. P�rez-Castillejos, E. Vall�s, J. Electrochem.Soc. 2002, 149, C201–C208; c) D. Golodnitsky, Y. Rosenberg, A. Ulus,Electrochim. Acta 2002, 47, 2707–2714.

[13] N. H. Turner, A. M. Single, Surface and interface analysis 1990, 15, 215–222.

Full Papers

1833ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

[14] V. D. Jovic, B. M. Jovic, M. G. Pavlovic, Electrochim. Acta 2006, 51, 5468–5477.

[15] E. Gomez, J. Ramirez, E. Valles, J. App. Electrochem. 1998, 28 71–79.[16] a) H. Liu, C. Fang, Y. Fang, Y. Zhou, H. Ge, F. Zhu, P. Sun, J. Miao, J. Mol.

Model. 2016, 22, 1–9; b) C. Fang, X. Lu, W. Buijs, Z. Fan, F. E. G. G�ner,M. A. van Huis, G.-J. Witkamp, T. J. Vlugt, Chem. Eng. Sci. 2015, 121, 77–86; c) V. Vchirawongkwin, B. M. Rode, I. Persson, The Journal of PhysicalChemistry B 2007, 111, 4150–4155.

[17] https://github.com/henniggroup/VASPsol.

[18] L. H. Mendoza-Huizar, M. Palomar-Pardav�, J. Robles, Electrochim. Acta2001, 46, 2749–2755.

[19] G. Henkelman, A. Arnaldsson, H. J�nsson, Computational MaterialsScience 2006, 36, 354–360.

Submitted: December 12, 2016

Accepted: January 27, 2017

Full Papers

1834ChemistrySelect 2017, 2, 1826 – 1834 � 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具