Surfactant Effects in Cu-Sn Alloy Deposition

D296 Journal of The Electrochemical Society, 159 (5) D296-D302 (2012)0013-4651/2012/159(5)/D296/7/$28.00 © The Electrochemical Society

Surfactant Effects in Cu-Sn Alloy DepositionArvydas Survila,∗,z Zenius Mockus, Stase Kanapeckaite,Dalia Brazinskiene, and Remigijus Juskenas

Center for Physical Sciences and Technology, Institute of Chemistry, Vilnius LT-01108, Lithuania

Adsorption behavior of different polyethers (PE) on copper and tin was studied in combination with kinetic investigations. Nopronounced effect of PE was detected in the Cu|Cu(II) system. In contrast, these substances show a high inhibition activity on the tinelectrode. Halides enhance adsorption ability on copper electrode but suppress it on tin; this effect intensifies in the sequence: Cl−< Br− < I−. It is supposed that specifically adsorbed halides can act as species bridging the copper substrate with Cu(I)-polyethercomplexes, whereas themodel of competitive adsorption is more suited for tin substrate. The inhibition degree of surfactants increaseswith the length of hydrocarbon chain. Underpotential deposition of Sn(II) is observed in the region of Cu(II) limiting current. Thecharacteristic current minimum arises in the region where the free Sn phase becomes thermodynamically stable. Cl− ions broaden therange of current densities where yellow bronze can be obtained, but Br− ions show the opposite effect. Iodides retard the reductionof Cu(II) therefore tin prevails in the coatings. The deposits obtained present a mixture of pure copper, α-CuSn (fcc) phase and hcpphase of α-CuSn.© 2012 The Electrochemical Society. [DOI: 10.1149/2.084205jes] All rights reserved.

Manuscript submitted October 18, 2011; revised manuscript received February 17, 2012. Published March 5, 2012. This was Paper1572 presented at the Montreal, QC, Canada, Meeting of the Society, May 1–6, 2011.

Copper-tin alloys are among the most beneficial coatings that areused for various purposes. They show good solderability, malleabil-ity, ductility and have good corrosion resistance. Besides, the coatingsknown as yellow bronze (containing 10–15 mass% of tin) confer ex-cellent decorative properties to the substrate. To obtain the desiredcharacteristics of bronze coatings, the control of Cu and Sn code-position is of fundamental importance. For this purpose, differentpolyethers (operating as surfactants) in combination with some otherorganic compounds (brighteners, stabilizators, etc) are widely em-ployed in modern plating industry.Recently much attention was focused on the role of polyethers

in deposition processes;1–26 however, for the most part these inves-tigations are concerned with single metals. Most concepts of copperelectrodeposition that were elaborated at the end of the XX century(see reviews14,15) have been acceptable up to now. It has been foundonce and again that polyethylene glycols (PEGs) show a rather weakadsorption in the absence of chloride ions. Moreover, the presenceof Cu(II) in the solution is also required to initiate the inhibitive ad-sorption of such surfactants.10 Then the formation of multiple sur-face layer, that represents a barrier to electron transfer, becomespossible.Among other factors which are responsible for PEG surface activ-

ity, the length of hydrocarbon chain seems to hold the lead. It has beenestablished26 that the binding strength between the adsorbed Cl− andPEG is proportional to the number of ether groups in PEG and thesmall PEG results in much more coverage defects than the large PEG.The effect of PEGmolecular mass on its adsorption rate and the prop-erties of adsorbates on platinum has been also considered.27 Besides,PEG has been found to affect the kinetics of palladium deposition aswell as both morphology and bulk properties of Pd deposits28: bulkdefectiveness increased with the size of PEG molecules. Investiga-tions of PEG role in electrodeposition of Zn-Cr alloys have shown29

that large PEG molecules enhance Cr(III) reduction.Our investigations30–42 performed at the Institute of Chemistry

(Vilnius, Lithuania) dealt mainly with such polyethers as laprol andsintanol (see Experimental) that are used in industrial baths. It hasbeen found that their electrochemical behavior is similar to that estab-lished for PEGs. A high surface activity on tin substrate and a weakadsorption on copper in the absence of halides are also typical of thesesurfactants.Despite certain progress in the understanding of partial processes,

the codeposition of copper and tin has been studied to a lesser degree.

∗ Electrochemical Society Active Member.z E-mail: [email protected]

Therefore, experimental research and theoretical modeling of Cu andSn codeposition, involving surface-active substances, is still a topicalpoint.The main purpose of this work was to reveal the main kinetic reg-

ularities of partial processes occurring in the Cu|Cu(II) and Sn|Sn(II)systems and to establish to what degree the observed regularities maybe applied for copper and tin codeposition. Special attention was paidto the adsorption behavior of surfactants on Cu and Sn electrodes, em-phasizing the role of their molecular mass (the length of hydrocarbonchain) in inhibition processes.Quantitative analysis of experimental data involved the transfor-

mation of voltammetric data into normalized Tafel plots and theconstruction of adequate equivalent circuits for the interpretation ofimpedance spectra. Both approaches made it possible to determinekinetic parameters of the cathodic processes and to obtain the adsorp-tion characteristics of different surfactants. Finally, the sequence ofpartial processes, the reasons for tin underpotential deposition (UPD),the elemental and phase composition of bronze coatings are discussedin this communication.

Experimental

Solutions were prepared using triply distilled water, CuSO4 · 5H2O(Mallinckrodt, USA, chlorides less than 0.005%), SnSO4 (Fluka orAldrich, chlorides less than 0.01%), H2SO4 and potassium halides(high purity, Reakhim, Russia). To maintain the stability of acid Sn(II)solutions, 1 mM of hydroquinone (HQ) was added. Special experi-ments and literature data43,44 showed that such a small amount of HQexerts no tangible effect on the results obtained.Different polyethers were added into the 0.01 M Cu(II) or/and

Sn(II) solutions. The molecular mass of ethylene glycol, its oligomersand polyethylene glycols with the general formula HO– (CH2–CH2–O)m–H covered the range from 62 (m = 1) to 40000 (m ≈ 909). Mono-di-, tri- and tetraethylene glycols were pure substances, the rest PEGswere fractionated mixtures with 5–10% deviation from the averagemolecular mass. The surfactant laprol 2402C is a bloc copolymerof polyethylene and polypropylene glycols. It may be symbolizedas X–O–X, where X represents the chain: –[CH2–CH(CH3)–O]10–[(C2H4–O)12–(CH2–CH(CH3)–O]2–H. Another polyether CnH2n+1–O–(C2H4–O–)mH, (n = 10–12,m = 8–10) (Orgsintez, Kazan, Russia)is known as sintanol DS-10. As distinct from the above PEGs, itcontains a comparatively long hydrocarbon radical, and owing to this,the molar fraction of C2H4O– groups averages only 0.62. Sintanolhas been widely used as an effective wetting agent and inhibitor forplating of tin and its alloys.

) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 131.211.208.19Downloaded on 2014-10-22 to IP

http://ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 159 (5) D296-D302 (2012) D297

Pure argon stream was passed through solutions for 0.5 h be-fore measurements. Argon atmosphere was also kept under solutionsthroughout the experiments.To prepare theworking electrodes, platinum substrateswere coated

with Cu or Sn in the solutions containing (g dm−3): a) CuSO4 · 5 H2O– 250, H2SO4 – 50; b) SnSO4 – 60, H2SO4 – 160, laprol 2402C– 1. The thickness of smooth coatings was 5–7 μm. The workingelectrodes were rinsed with water, immediately immersed into thesolution under investigation and kept in it for a controlled time τ.Besides, conventional three-electrode cells contained a copper, tinor bronze anode and an Ag|AgCl|KCl(sat) reference electrode. Elec-trode potentials E were converted to the standard hydrogen scale. Allexperiments were performed at 20◦C.A rotating disk electrode (RDE) was used in voltammetric (VA)

measurements that were performed at a potential sweep rate of5 mV s−1, using a potentiostat PI-50-1 (Russia). A Pt disk (1 cm2

surface area) fused into a glass holder served as a substrate for prepa-ration of working electrodes.A Pt wire with a surface area of 0.36 cm2 coated with copper

or tin was used in electrochemical impedance spectroscopy (EIS).Impedance spectra were obtained within a frequency f range from0.1 to 50 000 Hz using a Zahner Elektrik (Germany) IM6 impedancespectrum analyzer. Each record took about 5 min and was repeated3–4 times. Special computer programs were used for analysis ofimpedance data.A conventionalX-ray photoelectron spectroscopy (XPS) technique

was used to determine the elemental composition of coatings. Thesamples were prepared at a constant potential using RDE (1250 rpm).In some XPS experiments, an approximately 90-nm-thick surfacelayer was removed using the beam of Ar+ ions. A more detailed de-scription of the procedures applied is available in Refs. 38, 39, and 41.X-ray diffraction (XRD) measurements were performed with a

Bruker D8 Advance diffractometer equipped with a Gobel mirror as aprimary beam monochromator for Cu Kα radiation. The grazing inci-dence technique (X-ray incidence angle of 0.8◦) was used to preventan overlap of the brass support and deposit reflections in the XRDpattern. The detector scan mode was characterized by a step size of0.02◦ and a sampling time of 8–25 s. The phase composition was ob-tained from XRD patterns by calculating the net area of an XRD peakof every phase. Deconvolution of overlapping XRD peaks was per-formed using Cauchy–Lorentz function. Calculated intensities werenormalized to the most intensive one (111 peak of α-CuSn fcc phase)assuming it as 100.

Results and Discussion

Favorable conditions for codeposition of two metals are usuallyachieved when their equilibrium potentials are close. However, thisis not the case of Cu(II)-Sn(II) system, since the standard poten-tials of the Sn|Sn2+ and Cu|Cu2+ electrodes are equal to −0.136 and0.337 V respectively. However, a certain depolarization of Sn(II) re-duction is expected which is consistent with the energetic effects ofalloy formation. A shift in the equilibrium potential value can beobtained from the thermodynamic relationship

�E = −�G/nF, [1]

where�G is the partial Gibbs energy of themetal in the alloy, which isdirectly related to the change in activity of the metal. The so-called α-phase (supersaturated solid tin solution in copper) prevails in yellowbronze coatings, but the relevant thermodynamic data available inliterature refer, as a rule, to high temperatures. With the aim of roughestimation of depolarization effect,�G values, given in45 for 1000 K,were extrapolated to room temperature using the entropy of alloyformation46 �S= 20 J mol−1 K−1. This procedure yields�G≈ 25 kJmol−1 and�E ≈ 0.13 V. Similar use of thermodynamic data listed inRef. 43 results in �E ≈ 0.14 V. Eventually, the data given in Ref. 47for 25◦C allowed us to obtain the standard potential for the electrodereaction:

3 Sn2+ + 2 Cu3Sn+ 6 e = Cu6Sn5. [2]

Figure 1. Cathodic voltammograms obtained for PE-free Cu(II) solutions(lines) and in the presence of laprol (circles) or PEG-300 (triangles). Rotationvelocity of RDE (revolutions per minute) is indicated at the respective curve.Voltammograms for Cu(II)-free 0.6 M H2SO4 solutions were obtained in theabsence (line) and in the presence of laprol (circles).

This quantity was found to be 0.18 V higher than that of Sn|Sn2+

electrode. The results obtained show that certain depolarization isquite probable. The rest difference between the equilibrium potentials(∼0.3 V) might be compensated through the proper use of ligands thatcould produce strongCu(II) complexes but display a sufficiently loweraffinity to Sn2+ ions (stability constant of Cu(II) complexes should behigher by∼10 orders of magnitude). In our opinion, a real solution ofthis problem lies ahead. From the above, it might be assumed that thekinetic control of electrochemical reduction is gaining in importance.The use of surfactants that selectively influence partial processes offerspossibilities for effective control of metal codeposition.

Surface activity of polyethers on copper and tin substrates.—Voltammetric and EIS data show that the adsorption ability ofpolyethers on copper and tin substrates is quite different. Voltammo-grams obtained for Cu(II) solutions that are carefully protected fromchloride impurities contain well-defined plateaus of limiting current(Figure 1). Background currents are minor up to −0.4 V, where thehydrogen evolution becomes tangible. No pronounced effect of PE onthe kinetics of Cu(II) reduction is observed. The results obtained arealso in agreement with EIS data: the PEs under investigation showa minor effect on impedance (Z) spectra both in the Cu(II)-free andCu(II)-containing solutions and the double layer capacitance (Cdl) re-mains actually on the same level, viz.∼ 90μF cm−2 at the open-circuitpotential Eoc = 0.25 V.30 Hence, it may be deduced that discrepanciesin literature data3,5, 8, 10, 14, 15 concerning Cu(II)-PE systems seem toarise from a poor control of the purity of solutions.RDE voltammograms of Sn(II) reduction obtained for the PE-free

solutions are nearly reversible. They contain the plateau of limitingcurrent (id) that obeys Levich equation with the diffusion coefficientD = 6.2×10−6 cm2 s−1. In contrast to copper system, just smallamounts of PE exhibit a significant effect on the rate of Sn(II) reductionover a wide range of potentials (Figure 2). At the same time, theSn|Sn(II) systems cease to obey the regularities of diffusive masstransport, when the concentration of surfactant is increased. The effectof the intensity of forced convection reduces progressively with thesurfactant concentration (c) and becomes negligible at sufficientlylow c (see inset in Figure 2). The results obtained show that the Sn(II)reduction acquires indications of adsorption-controlled regime.The inhibition activity of surfactants can be also revealed from

impedance data. Nyquist plots (dependencies of the imaginary part ofimpedance Z// on the real part Z/), obtained for PE-free solutions atopen-circuit potentials, are nothing else than lines that are observedover a nearly entire range of applied frequencies. In accordance with



D298 Journal of The Electrochemical Society, 159 (5) D296-D302 (2012)

Figure 2. Cathodic voltammograms obtained at 440 rpm for 0.01 M Sn(II)solutions containing laprol, the concentration of which (mg dm−3) is givenat the respective curve. The effect of the intensity of forced convection at−0.38 V is presented in the inset in Levich coordinates.

voltammetric data, Sn(II) reduction is mainly controlled by diffusivemass transport in this case.The addition of PE gives rise to a semicircle observed at high

frequencies and results in simultaneous decrease in double layer ca-pacitance. The latter effect is observed for both Sn(II)-containing andSn(II)-free solutions. For instance, Cdl decreases∼6 times when only1 mg dm−3 of laprol is added into Sn(II)-free solutions and falls upto ∼13 μF cm−2 at clap = 0.1 g dm−3 (open-circuit conditions, Eoc= −0.26 V).48 Somewhat higher Cdl values (∼17 μF cm−2) wereobtained for laprol-containing solutions at Eoc = −0.24 V. The sameorder of Cdl magnitude remains at sufficiently low (E > −0.29 V) orhigh (E < −0.4 V) cathodic polarizations31 (see Figure 2). However,considerably stronger inhibition effects emerge at the potentials ofnegative slope of voltammograms (−0.33 < E < −0.30 V). In thiscase, Cdl falls to 7–9 μF cm−2 and Z/ acquires highly negative valuesranging from −2.3 to −4.5 k� cm2. Similar behavior is also typicalof other polyethers.34,40

It should be noted that the semicircle part of Nyquist plot devel-ops with the exposure time τ (Figure 3). Such plots are typical ofthe processes the rate of which is controlled by charge transfer anddiffusive mass transport simultaneously. Impedance spectra obtainedfor Sn(II) solutions can be described with a frequency error less than2% by means of equivalent circuit the description code of which isR�([RctQd]Qdl). Here, elements in series are given in square brack-ets, and elements in parallel are enclosed in parentheses. The ohmicresistance of the solutions R� was found to be actually independent

Figure 3. Nyquist plots obtained at different exposure times τ of tin electrodein 0.01 M Sn(II) solutions containing 0.02 M of tetraethylene glycol (TEG).Open-circuit potential Eoc = −0.25 V is applied.

Figure 4. Variations of the exchange current density i0 (ordinate to the left) andthe effective double-layer capacitance Cdl (ordinate to the right). Sn electrodewas kept for the time τ in the Sn(II) solution of indicated composition.

of TEG concentration and equal to 0.24 ± 0.01 � cm2. Diffusionand double-layer impedances are presented by the constant phase el-ements (CPE) Qd and Qdl, respectively. According to the impedancetheory, the complex conductivity of CPE (admittance Y) is given bythe relationship

Y = Y0( jω)n, [3]

where ω = 2πf, j = √−1 and exponent n characterizes the nature ofCPE. When n is equal to 1, 0.5, 0 or −1, CPE transforms into the ca-pacitance, Warburg impedance, resistance or inductance, respectively.Time-independent parameters of Qd were found to be as follows:

Y0 = 0.37 ± 0.02 �−1 cm−2 sn with n = 0.48 ± 0.02. The valueof Warburg coefficient calculated for the 0.01 M Sn(II) solution withD given above is equal to 0.384 �−1 cm−2 s0.5 and fairly coincideswith the experimental value. The exponent n obtained for Qdl variesbetween 0.88 and 0.91, this being indicative of a capacitive characterof the double layer impedance. According to the analysis performed,in contrast to R� and Qd, other elements (Rct and Qdl) vary withtime. Therefore, the exchange current density i0, obtained from thewell-known equation

i0 = RT/nF Rct, [4]

changes, as well. The data in Figure 4 show an obvious correlation be-tween the structure of the double layer and kinetics of Sn(II) reduction.Both i0 and the effective double-layer capacitance Cdl (it was obtainedusing special procedures47) vary in the same manner. The decreasein Cdl with time seems to arise from the progressive saturation of theinterface with an adsorbed TEG, causing an increasing inhibition oftin reduction. This manifests itself in the respective lowering of theexchange current density. A more than tenfold decrease in i0 is ob-served after two hours, as compared with the surfactant-free solution.It should be noted that the adsorption of PE with a high molecularmass, (laprol or sintanol) occurs significantly faster; a steady-state canbe attained in 5–10 minutes.The results obtained give grounds to suppose that the adsorption

of surfactant is the main factor responsible for the inhibition of ca-thodic process. From this point of view, voltammetric data might betreated invoking a simple kinetic model. According to it, the activeand passive sites with different current densities are formed on theelectrode surface. Then, the surface coverage θ can be obtained fromthe relation

i = iθ=0(1− θ)+ iθ=1θ, [5]

where iθ = 0 is the current density in the surfactant-free solution andiθ = 1 concerns a fully saturated adsorption layer. The latter value canbe obtained by linear extrapolation of values to 1/

√c → 0. On the




Figure 5. Adsorption isotherms obtained for tin electrode from impedance(triangles) and voltammetric (circles) data. Experimental data are fitted toFrumkin isotherm (solid lines) with the listed parameters.

other hand, according to the model of two parallel capacitors,49 aquite identical equation (with C instead of i) is valid for double-layercapacitance. Thus, both voltammetric and impedance data may beused for determination of surface coverage. Appropriate experimentaldata were fitted to Frumkin isotherm

Bc = θ

1− θexp(−2aθ), [6]

where B is the adsorption constant and a is a parameter accountingfor the interaction between adsorbed species (see Figure 5). In allcases, voltammetric data were obtained at ∼ −0.3 V; however, θvalues, obtained for sintanol at the open-circuit potential (−0.242 V)from capacitance data, are very close. In all probability, adsorptionparameters weakly depend on E in this region. It is interesting to notethat the repulsive interaction between adsorbed molecules is specificto laprol and PEG-1000 (a < 0), whereas the attractive forces prevail inthe adsorption layer formed by sintanol (a> 0). As has beenmentionedabove, the latter substance contains a comparatively long hydrocarbonchain. Note, that positive a values are characteristic of similar aliphaticcompounds.45 Large variations in B quantities arise, for the most part,from different molecular masses of surfactants.

Effect of halides.— Halides exert an essential influence on the ad-sorption ability of polyethers. It is particularly remarkable that thiseffect is quite different for copper and tin electrodes. Significant en-hancement of adsorption resulting in the inhibition of Cu(II) reductionis observed in the Cu|Cu(II) system, whereas diminished polymer ad-sorption was found to occur in the case of tin electrode. This canbe seen from the example shown in Figure 6. In the case of coppersystem, Nyquist plots take shape of arcs centered below the abscissaaxis. An addition of Cl− ions gives rise to a decrease in impedanceas compared to that obtained for halide-free solution (dashed linesin Figure 6). This result is in accordance with voltammetric data:30

chloride diminishes the polarization of Cu(II) reduction. However,the rest halides (Br− and I−) increase the impedance over an entirerange of the frequencies applied. An opposite effect can be seen inthe case of tin system (lower part of Figure 6). Both enhancement andsuppression effects intensify in the sequence: Cl− < Br− < I−.To analyze the consecutive charge transfer processes in the

Cu|Cu(II) system, the basic equivalent circuit R�([R1W1][R2W2]Qdl)was used. According to Ref. 50, the exchange current densities, i01and i02, of the respective processes Cu2+ + e → Cu+ and Cu+ + e→ Cu can be obtained from the equations

i01 + i02 = RT

F

(1

R1+ 1

R2

), [7]

Figure 6. Comparison of Nyquist plots obtained for the copper (upper part)and tin (lower part) electrodes in 0.01 M Cu(II) or Sn(II) solutions containinglaprol and 30 μM of different halides. Open-circuit conditions.

1

i01+ 1

i02= R1 + R2

σ1 + σ2

1

F√

D

(1

[Cu2+]+ 4

[Cu+]

), [8]

where σ1 and σ2 are the coefficients of Warburg impedance (ZW):

ZW = σ(1− j)/√

ω. [9]

It has been found, for example, that i01 values, obtained for 0.01 MCu(II) solution containing tetraethylene glycol (TEG) and 30μMBr−,decrease with cTEG from 40 to 10 μA cm−2, whereas i02 remains onthe same level equal to∼ 1 mA cm−2. Inequality i01 � i02 shows thatthe transfer of the first electron is the rate-determining step. Similardata have also shown that co-adsorption of halides and sintanol resultsin a more than tenfold reduction in the double layer capacitance. Inthe case of bromide-containing solutions,36 Cdl can decrease up to∼5 μF cm−2.In our opinion, interesting information can be obtained from

voltammetric data. To obtain kinetic parameters, the following kineticrelationship was used

i = 2i01

{exp

((2− αc1)F

RTη

)− [Cu2+]s[Cu2+]b

exp

(−αc1F

RTη

)}

[10]that is valid at i01 � i02. Here subscripts s and b denote the surface andbulk concentrations of Cu2+ ions. The former concentration dependson i according to the equation:

[Cu2+]s[Cu2+]b

= 1− i

id, [11]

The linear relationship

log−i

1− i/ id= log 2i01 − αc1F

2.303 RTη [12]

(specified as normalized Tafel plot, NTP) follows from Equations 10and 11 at a sufficiently high cathodic overvoltage.Voltammetric data, obtained for TEG-containing solutions, were

transformed according to Equation 12 (Figure 7). Linear parts of NTPare observed over a certain range of potentials, where the normalizedcurrent density decreases with TEG concentration. It follows fromNTP slopes that the cathodic charge transfer coefficient αc1 is equal




Figure 7. Normalized Tafel plots obtained for 0.01 M Cu(II) containing dif-ferent amounts of tetraethylene glycol and 30 μM of chloride (upper part,ordinate to the left) or bromide (lower part, ordinate to the right).

to 0.50 ± 0.03 and 0.40 ± 0.02 for chloride- and bromide-containingsolutions, respectively. The respective inhibiting action of TEG per-sists up to−0.03 and−0.1 V. Afterwards, the process accelerates andi values approach those typical of surfactant-free solutions.It is common knowledge that most organic compounds are ad-

sorbed near the zero charge potential (Ezc), where the surfacecharge density is not too high. Different Ezc of copper electrodehave been reported up to date. According to the results of the lastinvestigations,51,52 Ezc, determined for the bare, specially renewed Cusurface, fall between −0.6 and −0.7 V. More positive values (closeto 0 V), that were reported earlier by Yegorov and Novoselskij,53,54

are now attributed to the presence of adsorbed oxygen on copper. Itis highly probable, that the most Cu electrodes, discussed herein, fallinto the latter category.In this connection, some SERS experiments concerning the ad-

sorption of anions are worthy of notice. According to Ref. 55, Cl−

ions are co-adsorbed with sulfate in 2 M H2SO4 solutions at low over-voltages (at −0.3 < E < 0 V), while at high cathodic overvoltages,chloride is displaced by sulfate. Approximately the same range ofspecific (non-electrostatic) Cl− adsorption was established for Cu(II)-free perchlorate solutions.56 SERS was also used to investigate copperelectrode surfaces in the presence of Cl− and PEG.10 It was concludedthat PEG can be adsorbed in two different forms. One predominatesclose to the open-circuit potentials (0.18–0.28 V) and may be a cop-per chloride complex with the PEG as a ligand. The other (simpleneutral molecule) prevails at more negative potentials where chlorideis desorbed.It is evident from Figure 7 that the region of TEG inhibitive ad-

sorption clearly depends on the nature of halide.When Cl− is replacedby Br−, this region is shifted to more negative potentials by ∼60 mV.It follows from Cdl measurements54 that the Ezc of copper electrodeare equal to 0.09, 0.025,−0.01,−0.03 V in the presence of F−, ClO−

4 ,SO2−4 and Cl

− respectively; and even amore negative value is plausiblefor bromide-containing solutions.53 This keeping in mind, it is safeto assume that desorption of halides is the main factor resulting inthe destruction of the inhibitive adsorption layer. Then, the model10,25

might be accepted, according to which specifically adsorbed halidescan act as species bridging the copper substrate with Cu(I)-polyethercomplexes, formation of which is discussed elsewhere.14,15 The pos-sible structure of surface cluster is shown in Figure 8. Of course,this image needs special substantiation based on quantum-mechanicssimulations.The above considerations seem to be hardly acceptable for tin, and

the model of competitive adsorption is more suited in this case. If

Figure 8. Image of surface cluster. Two TEG molecules form complex withCu+ ion that is attached to copper surface via specifically adsorbed halide X−.

Ezc in the Sn|Sn(II) system is close to that established for perchloratesolutions57 (−0,4−−0,43V), joint adsorption of both PEG and halideis possible over a wide potential region where Sn(II) reduction occurs(see Figure 2).Reasons for such different adsorption properties of Cu and Sn

electrodes are not clear yet. Different hydrophility of copper and tinsubstrates might be among the factors responsible for the adsorptionbehavior of polyethers. Due to a weaker adsorption of water on thetin surface, more favorable conditions are created for direct contactbetween PEG and metal, and the competitive adsorption of PE andhalides seems to occur. On the other hand, chloride-suppressed PEGadsorption was also observed58 on the Pt electrode, and this effectincreased with chloride concentration.

Effect of length of the hydrocarbon chain.— As mentioned above,the molecular mass of the surfactants under investigation varied in awide range, beginning with ethylene glycol (EG) and its oligomers.In the case of copper electrode, a certain amount of halides wasadded in order to stimulate adsorption processes. Kinetic parametersof Cu(II) reduction were obtained from impedance (see above) andvoltammetric data. It was established that experimental NTP, obtainedfor EG-free solutions or in the presence of mono-, di-, and triethyleneglycols, actually coincide and may be approximated by one line in awide region of cathodic overvoltages. Hence, the inhibition propertiesof EG oligomers reveal themselves only in the presence of halides, ifthe number of ether oxygen atoms m ≥ 3, i.e. beginning with TEG.In this case only the transfer of the first electron onto Cu2+ ions isaffected. It should be noted that somewhat similar effects were foundto occur on the Pt electrode.27 According to this article, adsorptionproperties become distinct for PEGs with a molecular mass below600.The main conclusion, resulting from the experiments performed,

consists in the following: the inhibition degree of polyethers increasesconsiderably with their molecular mass. To compare the exchangecurrent densities, which fall drastically with the length of hydrocarbonchain, the molar concentration of each surfactant was multiplied bythe number of unit chains (–CH2–CH2–O–) in it. In this way the“concentration of unit chains”, cuc, can be obtained. These data arepresented in logarithmic coordinates in Figure 9. It can be seen thatthe inhibition activity of unit chain is different for each polyether andincreases with the polymerization degree.Similar effects were also observed in the case of tin electrode. The

inhibition activity of PE was evaluated from the impedance data. Itwas found that the shape of linear Nyquist plots does not actuallychange on addition of the short EG oligomers listed above. Close i0values (84 ± 4 mA cm−2) were detected for these solutions. OnlyTEG showed an increased surface activity, giving rise to a smallsemicircle at high frequencies with i0 = 46 mA cm−2. The inhibition




Figure 9. Variations of exchange current densities i01 obtained for 0.01 MCu(II) solutions containing 30 μM Br– and indicated PEGs. Their amount istransformed into “concentration of unit chains” (see text).

activity of surfactants was estimated in a similar way: the i0 valueswere normalized in respect to “1 mole of ethereal oxygen”. Accordingto the results obtained, the effect of molecules with 10–20 etherealbonds is ∼2000 times higher as compared with that of TEG. Thus,the length of hydrocarbon chain is also a prime factor responsible forthe inhibition activity of polyethers on tin electrode.

Codeposition of copper and tin.— Most of regularities that aretypical of partial processes persist in the case of codeposition ofCu andSn. Typical cathodic voltammograms are shown in Figure 10. Severalregions corresponding to different electrochemical processes may bedistinguished. Cu(II) reduction starts at its equilibrium potential equalto 0.24 V. It follows fromNernst equation that similar characteristic ofSn|Sn2+ electrode is−0.24V; therefore the phase consisting of free tinshould be stable at more negative potentials. Nevertheless, reductionof Sn(II) starts at ∼0 V. Thus, codeposition of Cu and Sn starts withcertain depolarization and with formation of specific Cu–Sn phases(see below).To understand the peculiarities of voltammograms, some regulari-

ties regarding the partial processes of Cu(II) and Sn(II) reduction areexpedient to invoke. As has been noted above, most PEs do not show

Figure 10. Effect of PEG-6000 on voltammograms of copper and tin code-position. The concentrations of PEG (in mg dm−3) are given at the respectivecurves. Cu-coated RDE, 1250 rpm.

a noticeable inhibition activity on copper substrate in the absenceof halides. Such behavior is also characteristic of other surfactantsin mixed Cu(II) and Sn(II) solutions. However, when the Sn|Sn2+

equilibrium potential is approached, the specific voltammetric mini-mum develops. The effect of surfactants becomes detectable at ratherlow concentrations (some mg dm−3). Its depth depends both on thePE molecular mass and concentration, as well as on the intensity offorced convection. This minimum deepens and the effect of forcedconvection weakens when the molecular mass of surfactant grows.It was established that the formation of free tin phase followed by

strong inhibitive adsorption is possible in this region. It is reasonablethat this feature is absent in Sn(II)-free solutions. The minimum underdiscussion develops gradually on addition of small amounts of Sn(II),and its depth may be used as a measure of Sn(II) concentration inthe presence of Sn(IV).59 Besides, anodic currents of Sn dissolutionare observed in cyclic voltammograms at E > −0.24 V, when thereverse potential scan is applied. An increase in current density arisingat higher cathodic polarizations (E < −0.4 V) is concomitant withhydrogen evolution.The role of halides that show quite opposite effects on copper and

tin electrodes is also important. Chlorides are often used in platingbaths for improvement of anodic dissolution of copper. This compo-nent increases to some extent the cathodic polarization and widens thecurrent range of yellow bronze deposition. However, the tin contentin the coatings increases in this case. By contrast, bromides narrowthis range and lower the tin content in the coatings. They can act asbrighteners in the presence of laprol, but a similar effect was not ob-served in the case of sintanol. Iodides inhibit copper electroplating sostrongly that the yellow bronze deposition becomes unfeasible. Forinstance, the threshold iodide concentration is only 0.5 μM.As mentioned above, UPD of tin results from the formation of

specific copper-rich phases. Their composition depends both on thesolution composition and electrolysis conditions. A specific feature ofelectrochemically deposited bronze coatings is the presence of variousphases that are thermodynamically stable only at high temperatures.In addition, our recent investigations38,39,42 demonstrated that Cu-Sn coatings obtained in the presence of some polyethers contain theintermediate hexagonal hcp phase of α-CuSn. When heated at 350 ◦Cthis phase transforms into ζ−Cu10Sn3. This unusual phase has neverbeen detected in the cast alloys. We associate its formation with theunderpotential deposition of tin on copper.Representative XRD patterns of CuSn deposits obtained in the

solution containing PEG-40000 are shown in Figure 11. It is evidentthat the phase composition strongly depends on potential: with anincrease in negative potential the quantity of the α-CuSn fcc phase(supersaturated solid solution of tin in copper) decreases while the

Figure 11. XRD patterns of Cu–Sn coatings deposited at indicated potentialsin the solutions containing 0.01 M Cu(II), 0.01 M Sn(II), 1 M H2SO4 and0.1 g dm−3 PEG-40000.




Figure 12. Phase composition of bronze coatings obtainedat −0.1 V in the solutions containing different polyethyleneglycols. XRD data are normalized with respect to XRD peak111 of α-CuSn fcc phase.

quantity of hcp increases. The hcp phase is also a solid solution of tinin copper, however, with a hexagonal structure.38

An example of phase composition is shown in Figure 12. The de-posits obtained in the PEG-free solution present a mixture of purecopper, α-CuSn phase (supersaturated solid solution of tin in copper)and hcp phase. The content of pure Cu decreases with PEG molecularmass (except PEG-40000), but the α-CuSn phase prevails in all cases.The tendency was observed according to which the luster character-istics of coatings improve when the content of hcp phase decreases.For instance, addition of a brightener (benzaldehyde) into sintanol-containing solutions reduces this phase by half.

Conclusions

1. Voltammetric and EIS data show that the adsorption behavior ofpolyethers (PE) on copper and tin substrates is quite different. Nopronounced effect of PE was detected in the Cu|Cu(II) system.In contrast, these substances show a high inhibition activity onthe tin electrode.

2. Halides enhance adsorption ability on the copper electrode butsuppress it on tin; this effect intensifies in the sequence: Cl−

< Br− < I−. The inhibition degree of surfactants increases withthe length of hydrocarbon chain.

3. Underpotential deposition of Sn(II) is observed in the regionof Cu(II) limiting current. The characteristic current minimumarises in the region where the free Sn phase becomes thermody-namically stable.

4. Cl− ions broaden the range of current densities where yellowbronze can be obtained, but Br− ions show the opposite effect.Iodides retard the reduction of Cu(II) and tin prevails in thecoatings. The deposits obtained present amixture of pure copper,α-CuSn fcc and hcp phases.

References

1. S. Meibuhr, E. Yeager, A Kozawa, and F. Hovorka, J. Electrchem. Soc., 110, 190(1963).

2. Y. E. Gerenrot, P. I. Dimarskaya, and A. P. Eichiss, Zashchita Metallov, 3, 465(1967).

3. M. R. H. Hill and G. T. Rogers, J. Electroanal. Chem., 86, 179 (1978).4. S. H. Glarum and J. H. Marshall, J. Electrochem. Soc., 130, 1088 (1983).5. M. Yokoi, S. Konishi, and T. Hayashi, Denki Kaganu, 51, 460 (1983).6. G. I. Medvedev and N. A. Tkachenko, Zashchita Metallov, 20, 484 (1984).7. G. I. Medvedev and O. N. Trubnikova, Elektrokhimiya, 20, 846 (1984).8. J. D. Reid and A. P. David, Plat. Surf. Finish., 74, 66 (1987).9. T. Pearson and J. K. Dennis, J. Appl. Electrochem., 20, 196 (1990).10. J. P.Healy,D. Pletcher, andM.Goodenough, J. Electroanal. Chem., 338, 155 (1992).11. J. J. Kelly and A. C. West, J. Electrochem. Soc., 145, 3472, (1992).12. O. Galdikiene and Z. Mockus, J. Appl. Electrochem., 24, 1009 (1994).13. T. C. Franklin, Plat. Surf. Finish., 81, 62 (1994).14. D. Stoychev and C. Tsvetanov, J. Appl. Electrochem., 26, 741 (1996).15. D. Stoychev and C. Tsvetanov, Trans. IMF, 76, 73 (1998).16. G. A. Noyanova, L. V. Kosmodamianskaya, and K. M. Tyutina, Zashchita Metallov,

35, 617 (1999).17. V. D. Jovic and B. M. Jovic, J. Serb. Chem. Soc., 66, 935 (2001).18. G. I. Medvedev and N. A. Makrushin, Zh. Prikl. Khim., 75, 1260 (2002).

19. L. Bonou, M. Eyraud, R. Denoyel, and Y. Massiani, Electrochim. Acta, 47, 4139(2002).

20. B.-H. Wu, C.-C. Wan, and Y.-Y. Yang, J. Appl. Electrochem., 33, 823 (2003).21. S. D. Beatie and J. R. Dahn, J. Electrochem. Soc., 150, A894 (2003).22. G. I. Medvedev, N. A. Makrushin, and O. V. Ivanova, Zh. Prikl. Khim., 77, 1120

(2004).23. G. I. Medvedev and N. A. Makrushin, Zh. Prikl. Khim., 77, 1799 (2004).24. A. Finke, P. Poizot, C. Guery, and J.-M. Tarascon, J. Electrochem. Soc., 152, A2364

(2005).25. K. R. Hebert, S. Adhikari, and J. E. Houser, J. Electrochem. Soc., 152, C324 (2005).26. W.-P. Dow, M.-Y. Yen, W.-B. Lin, and S.-W. Ho, J. Electrochem. Soc., 152, C769

(2005).27. T. Ya. Safonova, N. V. Smirnova, and O. A. Petrii, Russ. J. Electrochem., 42, 995

(2006).28. T. Ya. Safonova, D. R. Khairullin, G. A. Tsirlina, O. A. Petrii, and S. Yu. Vassiliev,

Electrochim. Acta, 50, 4752 (2005).29. T. Akiyama, S. Kobayashi, J. Ki, T. Ohgai, and H. Fukushima, J. Appl. Electrochem.,

30, 817 (2000).30. A. Survila, Z. Mockus, S. Kanapeckaite, and M. Samuleviciene, Polish J. Chem., 76,

983 (2002).31. A. Survila, Z. Mockus, and S. Kanapeckaite, Trans. IMF, 80, 85 (2002).32. A. Survila, Z. Mockus, and S. Kanapeckaite, J. Electroanal. Chem., 552, 97 (2003).33. A. Survila, Z. Mockus, S. Kanapeckaite, and V. Jasulaitiene, Russ. J. Electrochem.,

40, 855 (2004).34. A. Survila, Z. Mockus, S. Kanapeckaite, and M. Samuleviciene, Electrochim. Acta,

50, 2879 (2005).35. D. Brazinskiene and A. Survila, Russ. J. Electrochem., 41, 979 (2005).36. A. Survila, Z. Mockus, S. Kanapeckaite, and M. Samuleviciene, Trans. IMF, 84, 94

(2006).37. Z. Mockus, S. Kanapeckaite, V. Jasulaitiene, and A. Survila, Protection of Metals,

42, 485 (2006).38. R. Juskenas, Z. Mockus, S. Kanapeckaite, G. Stalnionis, and A. Survila, Electrochim

Acta, 52, 928 (2006).39. A. Survila, Z.Mockus, S. Kanapeckaite, V. Jasulaitiene, and R. Juskenas,Electrochim

Acta, 52, 3067 (2007).40. A. Survila and D. Brazinskiene, J. Solid State Electrochem., 11, 65 (2007).41. A. Survila, Z. Mockus, S. Kanapeckaite, V. Jasulaitiene, and R. Juskenas, J. Appl.

Electrochem., 39, 2021 (2009).42. A. Survila, D. Brazinskiene, S. Kanapeckaite, Z. Mockus, and V. Jasulaitiene, J. Solid

State Electrochem., 14, 507 (2010).43. N. M. Martyak and R. Seefeldt, Electrochim. Acta, 49, 4303 (2004).44. C. T. J. Low and F. C. Walsh, Electrochim. Acta, 53, 5280 (2008).45. C. B. Alcock and K. T. Jacob, Acta Metall., 22, 539 (1974)46. N. Saunders and A. P. Mlodownik, Bull. Alloy Phase Diagrams, 11, 278 (1990).47. G. J. Brug, A. L. G. van den Eeden, M. Sluyters-Rehbach, and J. H. Sluyters, J.

Electroanal. Chem., 176, 275 (1984).48. A. Survila, Z. Mockus, S. Kanapeckaite, and M. Samuleviciene, Chemistry (Vilnius),

12, 210 (2001).49. B. B. Damaskin, O. A. Petrii, and V. V. Batrakov, Adsorption of Organic Compounds

on Electrodes, Plenum Press, New York (1971).50. A. Survila and V. Baliukiene, Chemistry (Vilnius), 12, 195 (2001).51. G. J. Clark, T. N. Andersen, R. S. Valentine, and H. Eyring, J. Electrochem. Soc.,

121, 618 (1974).52. A. Lukomska and J. Sobkowski, J. Electroanal. Chem., 567, 95 (2004).53. L. Ya. Yegorov and I. M. Novoselskij, Elektrokhimiya, 7, 988 (1971).54. I. M. Novoselskij, N. I. Konevskikh, and L. Ya. Yegorov, Elektrokhimiya, 8, 1480

(1972).55. G. M. Brown and G. A. Hope, J. Electroanal. Chem., 405, 211 (1996).56. G. Niaura and A. Malinauskas, Chem. Phys. Letters, 207, 455 (1993).57. T. N. Andersen, J. L. Andersen, and H. Eyring, J. Phys. Chem, 73, 3562 (1969).58. E. Bahena, P. F. Mendez, Y. Meas, R. Ortega, L. Salgado, and G. Trejo, Electrochim.

Acta, 49, 989 (2004).59. G. Rozovskis, Z. Mockus, V. Pautieniene, and A. Survila, Electrochem. Comm., 4,

76 (2002).



Documents

Surfactant Effects in Cu-Sn Alloy Deposition