Electrochimica Acta 176 (2015) 10–17

Computational Chemistry and Electrochemical Studies of AdsorptionBehavior of Organic Additives during Gold Deposition in Cyanide-freeElectrolytes

Xuefeng Rena, Ying Songa, Anmin Liua, Jie Zhanga, Guohui Yuana, Peixia Yanga,Jinqiu Zhanga, Maozhong Ana,*, Daniel Materab, Gang Wub,**a State Key Laboratory of Urban Water Resource and Environment, School of Chemical Engineering and Technology, Harbin Institute of Technology,Harbin 150001, ChinabDepartment of Chemical and Biological Engineering, University at Buffalo, The State University of New York, Buffalo, NY 14260, United States

A R T I C L E I N F O

Article history:Received 15 April 2015Received in revised form 1 June 2015Accepted 30 June 2015Available online 3 July 2015

Keywords:Density functional theorymolecular dynamic simulationsgold electroplatingadditivescathodic polarization

A B S T R A C T

In this work, a combined approach using density functional theory (DFT), molecular dynamicsimulations, and electrochemical experiments was developed to study organic additives used for acyanide-free gold electrodeposition process. When 5,5-dimethylhydantoin (DMH) is used as thecomplexing agent, organic additives with either imino groups or nitrogenous heterocyclic rings includingpolyethyleneimine (PEI), 2,20-bipyridine, and nicotinic acid (NA) were systematically investigated.Electronic properties and orbital information of these studied additives were determined by quantumchemical calculations based on DFT. Compared to the other additives, PEI was predicted as the mostpromising one for the introduced DMH based gold electrodeposition. The theoretical prediction wasfurther verified by using experimental results. In particular, the cathodic polarization of the goldelectrodeposition was significantly improved in the presence of PEI. Also, the resulting goldelectrodeposit shows obviously finer crystalline particles largely. In addition, molecular dynamicsimulations determined that PEI molecules possess optimized adsorption behavior on the gold surface,suggesting its effectiveness to be additive in the studied cyanide-free Au-electrodeposition electrolytes.

ã 2015 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Electrochimica Acta

journal homepa ge: www.elsev ier .com/locate /e lectacta

1. Introduction

Due to their excellent physical, chemical, and electrochemicalproperties, electrodeposited gold films are widely used fornumerous decorative and functional applications [1–3]. Nowadays,the majority of operational gold electroplating electrolytes are stillcyanide-based, capable of providing dense and uniform gold filmsat low cost. The cyanide-based gold electroplating electrolytes alsohave high current efficiency and good stability, regardless ofalkaline, neutral or acidic conditions. Unfortunately, cyanide is anextremely toxic chemical, yielding high risks to human health andthe environment, especially when operated at a relatively hightemperature [4]. To solve this problem, a number of attempts havebeen made in the past few decades to develop cyanide-free goldelectroplating electrolytes [5–10].

* Corresponding author. Tel/fax: +86 451 86418616.** Corresponding author.

E-mail addresses: mzan@hit.edu.cn (M. An), gangwu@buffalo.edu (G. Wu).

http://dx.doi.org/10.1016/j.electacta.2015.06.1470013-4686/ã 2015 Elsevier Ltd. All rights reserved.

Due to its ability to coordinate with many metal ions, includingAu3+ [11–13], 5,5-dimethylhydantoin (DMH), a derivative ofhydantoin, was selected herein for use as a complexing agent toreplace the toxic cyanide, The solubility of DMH in alkalinesolutions is better than that in neutral or acid solutions due to theionization reaction provided in equation (1):

DMH@DMH� þ Hþ: (1)

Thus, a stable cyanide-free gold electroplating electrolyte canbe utilized with DMH as the complexing agent. However,standalone DMH still cannot provide a gold electrodeposit ofsufficient quality, and proper additives are needed to improve theelectrodeposited Au films [11,14]. Unfortunately, investigatingpotential additives through experimental approach is bothexpensive and time-consuming. Therefore, it is highly desirableto create a more effective approach when designing andoptimizing additive structures for utilization in cyanide-free goldelectroplating.

X. Ren et al. / Electrochimica Acta 176 (2015) 10–17 11

In the past decades, quantum chemical calculations andmolecular dynamic (MD) simulations have emerged as an effectivetechnique for the calculation [15–19] and simulation [20–23] ofsmall chemical systems to understand and predict their propertiesat the molecular level. Especially, both theoretical methods havebeen widely used to investigate the molecular configurations,electronic structures, reactivity of additives [24–26] and inhibitors[27–29] as well as the adsorption property of various organicmolecules on metal surfaces [30–34]. The highest occupiedmolecular orbital (HOMO) and lowest unoccupied molecularorbitals (LUMO) eigenvalues and the HOMO-LUMO gap (DE) arethe most commonly used parameters to interpret the interactionmechanisms occurring between organic additives and metalsurfaces at the molecular level [35–37]. MD simulations are alsowidely employed to study the adsorption interactions betweenorganics and metal surfaces [18,38–40], and can provide insightinto the understanding of additives with superior properties andinteraction energies to explain the influence of additives during theprocess of metal electrodeposition [41]. Particularly, high bindingenergy values mean that organic additive molecules will give highinhibition on the metal electrodeposition. Therefore, combiningquantum chemical calculations and MD simulations withexperimental investigation and experimental validation haveemerged as a highly efficient methodology to study additivesduring the electrodeposition process [42–46].

Regarding the gold electrodeposition, several polyamines ornitrogenous heterocyclic compounds have been demonstrated asgood additives [47–49]. However, their performance are still notsufficient to deposit high-quality Au films. Also, these explorationsto identify effective additives are only relied on time-consumingand high-cost electroplating experiments. In this work, wedeveloped a new approach to theoretically predicting additiveperformance, which was further verified by experimental studies.In particular, polyethyleneimine (PEI), 2,20-bipyridine, andnicotinic acid (NA) containing either imino groups or nitrogenousheterocyclic rings were selected to be the potential additives forthe DMH-based gold electroplating electrolytes. At first, quantumchemical calculations were employed to investigate the electronic

Fig. 1. Molecular structures of (a) DMH, (b

properties and orbital information of these studied additives.Molecular dynamic simulations were employed to study theadsorption behavior of PEI, 2,20-bipyridine, NA, and DMH on a goldsurface. Then, the adsorption of studied additives on cathodicmetal surfaces was predicted. Experimentally, cathodic polariza-tion curves, macroscopic images of gold deposits, and scanningelectron microscope (SEM) images were analyzed for furtherselection of additives. As a result, PEI, one polymer with manyimino groups, was demonstrated to be an excellent additive for thegold electroplating electrolyte using DMH as the complexing agent.

2. Theoretical methods and experimental details

2.1. Quantum chemical calculations and MD simulations

All quantum chemical calculations were carried out by densityfunctional theory (DFT) using the B3LYP functional method. In thegeometry optimization and molecular orbital characteristicanalysis, a 6-311G basis set was used for all atoms of the studiedsystems. All calculations on these systems under investigationwere performed using the Gaussian 09 program package at 318 Kwith water as solvent in the IEFPCM theoretical model.

MD simulations of the adsorption interactions between PEI,2,20-bipyridine, NA, or DMH molecules and Au (111) surface werecarried out in simulation boxes with periodic boundary conditionsusing Materials Studio (from Accelrus Inc). The MD simulation boxesconsisted of a gold surface (cleaved along the (111) crystal face), aliquid phase and a vacuum layer of 1 nm height. The liquid phase wascomposed of water molecules with a density of 1 g/cm3 and two PEImolecules, four 2,20-bipyridine, NA, or DMH molecules, respectively.NVT ensemble and COMPASS force field were employed, with a timestep of 1 fs and simulation time of 500 ps at 318 K. The interactionenergy Eorganic-surface between the organic and metal surface iscalculated as according to equation (2).

Eorganic-surface ¼ Etotal � Eorganic � Esurface (2)

Where Etotal is the total energy of the organic together with themetal surface, Eorganic and Esurface are the total energy of the organic

) PEI, (c) 2,20-bipyridine, and (d) NA.

Fig. 2. Localization of the highest occupied molecular orbital (HOMO) of (a) DMH, (b) PEI, (c) 2,20-bipyridine, and (d) NA.

12 X. Ren et al. / Electrochimica Acta 176 (2015) 10–17

and the metal surface, severally. The binding energy, Ebinding,represents the negative value of the interaction energy.

2.2. Electrochemical measurements

The gold electroplating electrolyte was prepared by dissolving0.3 mol/L DMH and 0.36 mol/L K2CO3 together into deionized waterto coordinate with 0.025 mol/L HAuCl4 to form [Au(DMH)4]�. Afterstirring for more than 5 hours to make sure that the Au (III) was fullycoordinated, 30 mg/L studied additives were added into theelectrolytes. All the used reagents in this work were analyticalgrade. The pH of all the electrolytes was maintained at 10.

Fig. 3. Localization of the lowest unoccupied molecular orbital

Gold electrodeposits used to study the macroscopic images andsurface morphologies were prepared under galvanostaticcondition (8 mA/cm2) with mild agitation at 318 K in variouselectrolytes. Cathodic polarization measurements were performedin a typical three-electrode cell coupled with an electrochemicalworkstation at a scanning rate of 1 mV/s at 318 K. A gold rotatingdisk electrode (Au-RDE) with a diameter of 4 mm was employed asthe working electrode (WE) and at a rotation speed of 500 rpm. Asaturated calomel electrode (SCE) and platinum foil wereemployed as the reference electrode (RE) and the counter electrode(CE), respectively. All potentials (E) are referenced to the SCE.Scanning electron microscopy (SEM) images were studied by field

(LUMO) of (a) DMH, (b) PEI, (c) 2,20-bipyridine, and (d) NA.

Fig. 4. Schematic diagrams of frontier molecular orbitals of (a) DMH, (b) PEI,(c) 2,20-bipyridine, and (d) NA, and the calculated E (eV).

X. Ren et al. / Electrochimica Acta 176 (2015) 10–17 13

emission scanning electron microscopy (FE-SEM, Hitachi SU8000)at an 20 kV working voltage.

3. Results and discussion

3.1. Density functional theory

The molecular structures of DMH and studied additives,including polyethyleneimine (PEI), 2,20-bipyridine, and nicotinicacid (NA), are shown in Fig. 1. The oxygen or nitrogen atoms inthese organic molecules may result in strong adsorption on metalsurface, which affects Au electrodeposition behavior.

Quantum chemical calculations based on DFT principles wereemployed to study the electronic properties and orbitalinformation of the complexing agent DMH, and studied additives.According to the frontier molecular orbital theory, the highestoccupied molecular orbital (HOMO) is often associated withelectron donating ability of organic molecules. Fig. 2 exhibits thedistribution of the HOMO and electron cloud densities of DMH andstudied additives.

These crimson and green clouds with symmetrical structures inFig. 2 stand for the phases of orbital wave functions. In the analysisof quantum chemical calculations, if one color stands for thepositive phase, the other one is for the negative phase, asestablished on the study of bonding reaction and clouddistributions of organics. As displayed in Fig. 2, the nitrogen oroxygen atoms in these studied organics show significantcontribution to the HOMO. This is due to the electron donatingproperties of the nitrogen and oxygen atoms, associated with thestrong adsorption ability on metal surfaces.

The lowest unoccupied molecular orbital (LUMO) of these fourorganics was also distributed on the nitrogen and oxygen atoms(Fig. 3), which is similar to the distribution of HOMO in Fig. 2. Theseresults suggest that the nitrogen or oxygen atoms containing

Fig. 5. The MD simulation boxes and the simulation energy and temperature vs. time cu(b) top view), (c) and (d) the final equilibrium configuration ((c) front view, (d) top vie

regions are the likely reactive sites for the adsorption of organicson the metal surfaces. The density distributions of the frontiermolecular orbitals (HOMO and LUMO) of these additive moleculesshow that the nitrogen atoms are the preferential sites foradsorption on the metal surface.

In addition to the localization of the molecular orbitals, EHOMO,ELUMO, and their differences (DE) are useful tools to characterize theelectronic properties and adsorption behavior of each additive.EHOMOand ELUMO, along withDE are important molecular electronicproperties that can dictate the behavior and adsorption propertiesof the studied organics. High value of EHOMO indicates a tendency ofthe molecule to donate electrons to appropriate acceptor molecules(with lower energy due to empty atomic orbitals). The energy of the

rves for PEI on Au (111) surface. (a) and (b): the initial configuration ((a) front view,w), (e) the simulation energy vs. time curves, and (f) temperature vs. time curves.

Fig. 6. Cathodic polarization curves of the Au-RDE at 500 rpm in the DMHbased gold electroplating electrolytes without and with 30 mg/L PEI, 30 mg/L2,20-bipyridine, and 30 mg/L NA additives.

14 X. Ren et al. / Electrochimica Acta 176 (2015) 10–17

lowest unoccupied molecular orbitals can indicate the moleculesability to accept electrons. The less negative EHOMO and the smallerDE are reflected in stronger chemisorption bond and greateradsorption stability [50]. A schematic depiction of these valuesbased on frontier molecular orbitals is presented in Fig. 4.

As presented in Fig. 4, DMH and NA show similar and relativelyhigh EHOMO values of �7.503 eV and �7.407 eV, respectively. Thissuggest more stabilized electrons are close to the core. EHOMO of PEIand 2,20-bipyridine are �5.410 eV and �6.855 eV, respectively,which is likely good adsorbents, due to their capability to donateelectrons to the metal surface and form rapid adsorption. Inprinciple, the adsorption capacity of the organics to the metalsurface increases with an increase of EHOMO or a decrease of ELUMO,this would yield a stronger organic adsorption on the metalsurface. Therefore, these results indicate that the electron donatingability of PEI and 2,20-bipyridine is higher than that of DMH and NA.Therefore, compared with DMH and NA, the unoccupied metalatom orbitals accept electrons more easily and form strongadsorption bonds with PEI and 2,20-bipyridine.

Oppositely, PEI has a smaller DE than that of DMH, NA, and2,20-bipyridine, suggesting a more effective adsorption of PEI onthe metal surface. As the schematic diagram displayed in Fig. 4,the DE decreases from DMH, NA, 2,20-bipyridine to PEI,indicating that electron transfer between molecular orbitalHOMO and LUMO is more favorable in the PEI molecule. Inprinciple, when charge transfer from organics takes place duringtheir adsorption on the metal surface, increased EHOMO facilitatesthe adsorption. Thus, adsorption stability is then largelyinfluenced by the charge transfer process throughout theadsorbed organic layer [50–52]. The effective adsorption ofPEI is due to stronger inhibition of gold electrodeposition on thecathode surface.

ELUMO of DMH, PEI, 2,20-bipyridine, and NA were �0.888 eV,�0.659 eV, �1.826 eV, and �2.192 eV, respectively. The differenceof ELUMO between DMH and PEI is the smallest, indicating thatthese two organics have a similar ability to form back-donationbonds when accepting electrons from the anti-bonding orbitalswithin cathodic metal atoms. In addition, the relatively low ELUMO

values of 2,20-bipyridine and NA imply that acceptance of electronsfrom the cathodic metal atoms is easier relative to DMH and PEI byforming an anti-bond with the cathodic metals. Thus, theadsorption of 2,20-bipyridine and NA on the Au may be less stablethan that of DMH and PEI.

Based on above DFT calculation, PEI and 2,20-bipyridine caneasily form adsorption bonds on the metal surface due to theirrelatively high EHOMO values along with strong electron donatingcapability. 2,20-bipyridine, with a relatively lower EHOMO than PEIas well as smaller DE than DMH and NA, could effectivelyadsorb on the metal surfaces. But the adsorption stability of2,20-bipyridine on the metal surface might be decreased due to itsanti-bonding ability. PEI could adsorb on the metal surfaces with

Fig. 7. Optical images of gold electrodeposits obtained from the electrolytes (a) in the a(d) 30 mg/L NA.

the most stability because of its smallest DE. In addition,considering its larger molecular structure, the adsorbed PEI maybe the most efficient additive to inhibit gold deposit on thecathodic metal surfaces. Hence, among all of studied additives, PEIis theoretically predicted as the most promising one for theintroduced DMH based gold electrodeposition. Next, electrochem-ical measurements and gold electroplating were carried out tofurther verify the prediction.

3.2. The interaction between additives and metal surface

According to above theoretical discussions, the addition of PEIshould enhance cathodic polarization of gold electrodepositionand improve the quality of gold electrodeposit. In order to furtherunderstand the role of additives during electrodeposition at themolecular level, MD simulation was employed to provide insightfulinformation. Therefore, MD simulations were performed to studythe adsorption behaviors of PEI, 2,20-bipyridine, NA, and DMH onAu (111) surface.

For the MD simulation boxes, initial and final equilibriumconfigurations, the simulation energy vs. time, and simulationtemperature vs. time fluctuation curves are given in Fig. 5, Fig. S1,Fig. S2, and Fig. S3, respectively for PEI, 2,20-bipyridine, NA, andDMH molecules on an Au (111) surface.

At the beginning of the simulation, two PEI molecules arerandomly positioned near the Au (111) surface at the initialconfiguration, as shown in Fig. 5 (a). When the simulation issuccessfully completed, the simulation energy vs. time and

bsence of additive and containing (b) 30 mg/L PEI, (c) 30 mg/L 2,20-bipyridine, and

Fig. 8. SEM images of gold electrodeposits obtained from the electrolytes in the absence of additive (a, b), and containing 30 mg/L PEI: (c, d), 30 mg/L 2,20-bipyridine: (e, f),30 mg/L NA: (g, h).

X. Ren et al. / Electrochimica Acta 176 (2015) 10–17 15

16 X. Ren et al. / Electrochimica Acta 176 (2015) 10–17

simulation temperature vs. time curves (Fig. 5 (e) (f)) indicate thatthe simulation systems are already at equilibrium. At theequilibrium state, the imino groups of the PEI molecules arevirtually parallel to the and Au (111) surfaces, this can be clearlyobserved from the front and vertical views of the simulation boxesas shown in Fig. 5 (c) and (d), respectively. As displayed in Fig. S1,Fig. S2, and Fig. S3, similar results can be obtained for the case of202-bipyridine, NA, and DMH on Au (111) surface. With nitrogenousheterocyclic rings, 202-bipyridine, NA, and DMH all have goodadsorption and achieve parallel absorption on Au (111) surface. TheEbinding of PEI, 202-bipyridine, NA, and DMH are 311.74 kJ/mol,337.98 kJ/mol, 259.33 kJ/mol, and 214.13 kJ/mol, respectively. Thisindicates that PEI and 202-bipyridine have larger binding energythan NA and DMH on Au (111) surface, which are strong enoughand reflect sufficient adsorption abilities of molecules on goldsurface. Furthermore, with the lower ELUMO values of �1.826 eV,the stronger anti-bonding ability of 202-bipyridine lead to areduction of adsorption stability of 202-bipyridine on the metalsurfaces. Hence, PEI reveals the best adsorption behavior on goldsurface. The adsorption behaviors of PEI on the metals show afurther explanation of the inhibition effect of PEI on the goldcathodic electrodeposition measured by the cathodic polarizationcurves.

3.3. Cathodic polarization measurement

Cathodic polarization curves were created to further study theeffects of various additives on gold electrodeposition in order tofurther validate the above theoretical predication. Fig. 6 displayscathodic polarization curves of the Au-RDE at 500 rpm in the DMHbased gold electroplating electrolytes with or without variousadditives.

The polarization curves measured with the electrolytes in theabsence of additive and with NA are nearly overlapped with onsetpotentials around �0.450 V, suggesting insignificant role of NAduring the electrodeposition. Meanwhile, the onset potential ofcathodic deposition in the presence of 2,20-bipyridine is slightlynegative around �0.528 V showing very limited enhancement ofpolarization during the gold electrodeposition using DMH as acomplex agent. Because, the complexing agent mitigate theiranti-bonding ability on the metal surfaces, which leads to aunstable adsorption of 2,20-bipyridine and NA on the metal surface.Furthermore, addition of PEI into the electrolyte significantly shiftsonset potential to negative direction up to �0.656 V. Such a largeshift (206 mV) of the onset potential indicates a significantelectrochemical polarization for Au electrodeposition in thesolutions on the Au-RDE with PEI as the additive. [Au(DMH)4]�

ions diffuse to the cathode surface and accept electrons from thecathode, and then reduce to metallic Au on the cathode surface.More negative cathodic polarization means that it was moredifficult to reduce [Au(DMH)4]� to Au, thereby yielding largeroverpotential and smaller crystal grains in the gold electrodeposit.In other words, the electrodeposition process is significantlyinhibited with the addition of PEI, confirming the prediction of thequantum chemical calculations.

3.4. Morphology of gold electrodeposits

The effects of these three additives, PEI, 2,20-bipyridine, and NA,on the gold electrodeposits were further studied by directlyobserving the surface morphology of the obtained goldelectrodeposits by using SEM images. Visually, as shown in Fig. 7,the appearance of the gold electrodeposits obtained from theelectrolytes in the absence of additive and containing 2,20-bipyridineandNAwere all dark brown. Oppositely, the electrolytewith additionof PEI is able to lead to very bright golden surface.

Furthermore, SEM images shown in Fig. 8 were employed togain insight into the grain sizes of the gold electrodepositsobtained from various electrolytes. It can be seen that the goldelectrodeposits obtained from the blank electrolyte and theelectrolytes in the presence of 2,20-bipyridine or NA have similarrough and loose morphologies, showing large crystal grains. Asexpected, the gold electrodeposit obtained from the electrolytewith PEI is smooth and compact with small crystal grains (Fig. 8cand d). These results clearly provide an indication that the surfacequality of the gold electrodeposit could not be changed by theadditions of 2,20-bipyridine and NA, but can be significantlyimproved in the presence of PEI.

4. Conclusion

In summary, we report an effective approach to predicting thebehavior of organic additives during the gold electrodepositionprocess, by combining quantum chemical calculations andmolecular dynamic simulation with electrochemical experimentverification. Among studied organic additives, PEI and2,20-bipyridine were easier to form adsorption bonds on metalsurfaces due to their relatively higher EHOMO. However, adsorptionstability of 2,20-bipyridine might be decreased due to theanti-bonding ability. Oppositely, the smallest DE = 4.752 eV ofPEI results in the strongest stability of PEI adsorption on metalsurface. MD simulations further revealed that PEI has strongadsorption ability on Au (111) surface, which was better than2,20-bipyridine and NA. Thus, combined two types of theoreticalcalculations, PEI was predicted as the most promising additive forthe introduced DMH based gold electrodeposition.

Furthermore, the effects of these selected additives on real goldelectroplating were further experimentally investigated withcathodic polarization curves, optical images, and SEM images.When using PEI, the cathodic polarization of the goldelectrodeposition was significantly increased, thereby leading toa significant overpotential. This is highly desirable forelectrodplating yielding a golden bright, smooth, and compactgold electrodeposit with small grain sizes. However, the addition of2,20-bipyridine and NA almost had no effect on the cathodicpolarization and could not improve the macro- andmicromorphology of gold electrodeposits. Both the experimentaland theoretical results agree well with each other in this regardand confirm that PEI is an effective additive for the stable andenvironmentally friendly gold electroplating electrolyte. This workalso can extend to other metal electrodeposition, providing a newmethod to predict the adsorption behavior of organic additives.

Acknowledgements

The authors are grateful for financial support from the State KeyLaboratory of Urban Water Resource and Environment (HarbinInstitute of Technology) (2015DX09). G.W. gratefully acknowledgesfinancial support from the New York State Center of Excellence inMaterials Informatics and the Startup funds from the University atBuffalo (SUNY).

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.electacta.2015.06.147.

References

[1] J. Wu, H.J. Bai, X.B. Zhang, J.J. Xu, H.Y. Chen, Langmuir 26 (2010) 1191–1198.[2] A. He, Q. Liu, D. Ivey, J. Mater. Sci. Mater. Electron. 20 (2009) 543–550.

X. Ren et al. / Electrochimica Acta 176 (2015) 10–17 17

[3] J. Haber, P. Nowak, P. Zurek, Langmuir 19 (2003) 196–199.[4] T.Z. Sadyrbaeva, Sep. Purif. Technol. 86 (2012) 262–265.[5] S. Sobri, S. Roy, D. Aranyi, P.M. Nagy, K. Papp, E. Kalman, Surf. Interface Anal. 40

(2008) 834–843.[6] G. Baltrunas, A. Valiunien _e, E. Vienožinskis, T. Jankauskas, Ž. Margarian, J. Appl.

Electrochem. 38 (2008) 1519–1526.[7] L.M.A. Monzon, F. Byrne, J.M.D. Coey, J. Electroanal. Chem. 657 (2011) 54–60.[8] B. Bozzini, E. Tondo, A. Bund, A. Ispas, C. Mele, J. Electroanal. Chem. 651 (2011)

1–11.[9] S. Dimitrijevic, M. Rajcic-Vujasinovic, V. Trujic, Int. J. Electrochem. Sci. 8 (2013)

6620–6646.[10] E. Vanecht, K. Binnemans, J.W. Seo, L. Stappers, J. Fransaer, Phys. Chem. Chem.

Phys. 13 (2011) 13565–13571.[11] K. Oyaizu, Y. Ohtani, A. Shiozawa, K. Sugawara, T. Saito, M. Yuasa, Inorg. Chem.

44 (2005) 6915–6917.[12] A. Liu, X. Ren, M. An, J. Zhang, P. Yang, B. Wang, Y. Zhu, C. Wang, Sci. Rep. 4

(2014) 3837.[13] G.Z. Pavlovich, R.G. Luthy, Water Res. 22 (1988) 327–336.[14] X. Yang, M. An, Y. Zhang, L. Zhang, Electrochim. Acta 58 (2011) 516–522.[15] H.J. Kulik, N. Luehr, I.S. Ufimtsev, T.J. Martinez, J. Phys. Chem. B 116 (2012)

12501–12509.[16] P. Söderhjelm, F. Aquilante, U. Ryde, J. Phys. Chem. B 113 (2009) 11085–11094.[17] C.C. Wang, S.S. Siao, J.C. Jiang, Langmuir 27 (2011) 14253–14259.[18] H. Shokry, J. Mol. Struct. 1060 (2014) 80–87.[19] C. Murray, S.D. Elliott, ACS Appl. Mater. Interfaces 5 (2013) 3704–3715.[20] B.C. Stephenson, A. Goldsipe, D. Blankschtein, J. Phys. Chem. B 112 (2008)

2357–2371.[21] V. Tsui, D.A. Case, J. Am. Chem. Soc. 122 (2000) 2489–2498.[22] Z. Shen, H. Sun, X.Y. Liu, W.T. Liu, M. Tang, Langmuir 29 (2013) 11300–11309.[23] C. Zhang, Y. Wen, X. Xue, ACS Appl. Mater. Interfaces 6 (2014) 12235–12244.[24] Z.P. Liu, P. Hu, J. Am. Chem. Soc. 123 (2001) 12596–12604.[25] T.M.T. Huynh, F. Weiss, N.T.M. Hai, W. Reckien, T. Bredow, A. Fluegel, M. Arnold,

D. Mayer, H. Keller, P. Broekmann, Electrochim. Acta 89 (2013) 537–548.[26] M.J. Swadley, T. Li, J. Chem. Theory Comput. 3 (2007) 505–513.[27] A. Kokalj, S. Peljhan, M. Finšgar, I. Milošev, J. Am. Chem. Soc. 132 (2010)

16657–16668.

[28] A. Kokalj, S. Peljhan, Langmuir 26 (2010) 14582–14593.[29] N. Kova9cevi�c, A. Kokalj, J. Phys. Chem. C 115 (2011) 24189–24197.[30] G. Hong, H. Heinz, R.R. Naik, B.L. Farmer, R. Pachter, ACS Appl. Mater. Interfaces

1 (2009) 388–392.[31] J. Wen, J. Ma, Langmuir 26 (2010) 5595–5602.[32] Z. Rinkevicius, X. Li, J.A.R. Sandberg, K.V. Mikkelsen, H. Ågren, J. Chem. Theory

Comput. 10 (2014) 989–1003.[33] S. Peljhan, A. Kokalj, J. Phys. Chem. C 113 (2009) 14363–14376.[34] M. Pagliai, S. Caporali, M. Muniz-Miranda, G. Pratesi, V. Schettino, J. Phys.

Chem. Lett. 3 (2012) 242–245.[35] L. Feng, H. Yang, F. Wang, Electrochim. Acta 58 (2011) 427–436.[36] S.W. Xia, M. Qiu, L.M. Yu, F.G. Liu, H.Z. Zhao, Corros. Sci. 50 (2008) 2021–2029.[37] P.B. Raja, A.K. Qureshi, A. Abdul Rahim, H. Osman, K. Awang, Corros. Sci. 69

(2013) 292–301.[38] B.-Y. Liu, Z. Liu, G.-C. Han, Y.-H. Li, Thin Solid Films 519 (2011) 7836–7844.[39] T. Panczyk, T.P. Warzocha, P. Szabelski, W. Rudzinski, Langmuir 24 (2008)

8719–8725.[40] S. Monti, C. Li, V. Carravetta, J. Phys. Chem. C 117 (2013) 5221–5228.[41] C. Wang, M. An, P. Yang, J. Zhang, Electrochem. Commun. 18 (2012) 104–107.[42] C. Wang, J. Zhang, P. Yang, M. An, Electrochim. Acta 92 (2013) 356–364.[43] V. Jaiswal, R.B. Rastogi, J.L. Maurya, P. Singh, A.K. Tewari, RSC Adv. 4 (2014)

13438–13445.[44] B. Wrzosek, J. Cukras, J. Bukowska, J. Raman Spectrosc. 43 (2012) 1010–1017.[45] H. Orita, T. Kubo, T. Matsushima, A. Kokalj, J. Phys. Chem. C 114 (2010)

21444–21449.[46] D. Loffreda, Y. Jugnet, F. Delbecq, J.C. Bertolini, P. Sautet, J. Phys. Chem. B 108

(2004) 9085–9093.[47] H. Honma, K. Hagiwara, J. Electrochem. Soc. 142 (1995) 81–87.[48] H. Honma, Y. Kagaya, J. Electrochem. Soc. 140 (1993) L135–L137.[49] T. Osaka, M. Kato, J. Sato, K. Yoshizawa, T. Homma, Y. Okinaka, O. Yoshioka, J.

Electrochem. Soc. 148 (2001) C659–C662.[50] S. El Issami, L. Bazzi, M. Mihit, B. Hammouti, S. Kertit, E.A. Addi, R. Salghi, Pigm.

Resin. Technol. 36 (2007) 161–168.[51] K. Barouni, L. Bazzi, R. Salghi, M. Mihit, B. Hammouti, A. Albourine, S. El Issami,

Mater. Lett. 62 (2008) 3325–3327.[52] M.K. Awad, J. Electroanal. Chem. 567 (2004) 219–225.