Original Article. Recent Res. Devel. Electrochem., 9 (2013): 1-32 ISBN:

Transcription

1 Transworld Research Network 37/661 (2), Fort P.O. Trivandrum Kerala, India Original Article Recent Res. Devel. Electrochem., 9 (2013): 1-32 ISBN: Construction of electrocapillary curves for the thermodynamic study of the adsorption of PEG in presence of Cl - ions onto a polycrystalline Au electrode in a solution of KClO 4 Alia Méndez, G. Orozco, L. Ortiz-Frade, L.G. Arriaga and G. Trejo Laboratorio de Materiales Compuestos y Recubrimientos Funcionales. Centro de Investigación y Desarrollo Tecnológico en Electroquímica (CIDETEQ). Parque Tecnológico Sanfandila, Pedro Escobedo, A.P. 064, C.P , Querétaro, México Abstract. In this chapter, the adsorption of polyethylene glycol (MW 20,000) (PEG 20,000 ) in the presence of Cl - ions onto a polycrystalline gold electrode was studied using a quartz crystal microbalance, AC voltammetry, and chronocoulometry. This study provided new physical insights into the adsorption behavior of this polymer in the presence of chloride ions on a solid electrode in KClO 4. Specifically, three potential-based behaviors were observed: the adsorption of PEG 20,000 was observed to form a condensed phase within the -1.0 to -0.6V vs. SCE interval. The obtained G ads values allow for a hypothesis that adsorption occurs by means of hydrogen bridges. In the potential range from to -0.1 V vs. SCE, the addition of chloride does not lead to an increased adsorption of PEG 20,000, but rather to a simple superposition of both adsorption processes. The potential range Correspondence/Reprint request: Dr. G. Trejo, Laboratorio de Materiales Compuestos y Recubrimientos Funcionales. Centro de Investigación y Desarrollo Tecnológico en Electroquímica (CIDETEQ). Parque Tecnológico Sanfandila, Pedro Escobedo, A.P. 064, C.P , Querétaro, México. gtrejo@cideteq.mx

2 2 Alia Méndez et al. from -0.1 to 0.35 V vs. SCE is dominated by the adsorption of Cl - ions, and a displacement of adsorbed PEG 20,000 molecules associated with the adsorption of Cl - anions was observed. Based on the obtained G ads values, we propose that, in this potential range, the interaction of the PEG 20,000 molecules with the Au polycrystalline electrode surface is defined as weak chemisorption via the oxygen atoms. 1. Introduction Polyethoxylated compounds and Cl - ions are of great importance in the electroplating industry, where are principally used as additives in electrolytic baths. Additives have become indispensable components of electrolytic baths because their presence substantially improves the quality of the metal coatings obtained. The advantageous effects of additives on the morphology and physical properties of metal coatings include the following: reduction of the grain size of the metallic crystals, which aids in the production of smooth, adherent, and shiny coatings [1-4]; modification of the crystallographic orientation of the coatings; and increase in the corrosion resistance [5]. The beneficial effects of adding polyethylene glycol (PEG) and Cl - ions to an electrolytic bath have motivated a growing interest in the effects of these additives on the morphologies and physical properties of the deposits. In general, the effects of PEG and Cl - ions on the morphological characteristics of electrodeposited coatings principally stem from adsorption of the additives molecules on the electrode surface [6,7]. The role of PEG in controlling the coating quality is primarily associated with the adsorption of poly-molecules at the electrode surface; these traits require the presence of Cl - ions as another constituent of the plating bath. Various studies have suggested that, in the presence of Cl -, PEG adsorbed on an electrode surface forms a barrier that inhibits metal deposition and increases the overpotential for the discharge of metal ions, such as Zn(II) [8] or Cu(II) [9] ions. Furthermore, the degree of inhibition increases with increasing molecular weight of the polyethoxylated compound [6]. In the establishment of these relationships, special attention was paid to PEG adsorption. Recently, on the basis of cyclic voltammetry (CV) data, Safonova et al. [7] have proposed that, in an H 2 SO 4 electrolyte solution, PEG adsorption onto a Pt/Pt electrode is accompanied by electrooxidation of the PEG molecules (likely via a dehydrogenation mechanism) or by electroreduction of the PEG molecules (likely via a hydrogenation mechanism). Similar results have been reported by Smirnova and Kazakova [10] in a study of the adsorption of oligoethylene glycols on platinized platinum electrodes. Using capacitance measurements, Petri et al. [11] have established that PEG molecules adsorb

3 A study of PEG/Cl adsorption on an Au electrode 3 on an Au(111) electrode to form two-dimensional condensates. Using in situ spectroscopic ellipsometry, Walker et al. [12] have shown that PEG is only weakly adsorbed on a Ru electrode in the absence of Cl - ions. In addition, in situ electrochemical scanning tunneling microscopy (EC-STM) images of PEG (MW 1000) (PEG 1000 ) molecules adsorbed on a Fe(110) surface have revealed that PEG 1000 exhibits a well-ordered structure on this surface. On the basis of this finding, Kim et al. [13] have proposed a model for PEG adsorption in which PEG 1000 is adsorbed in a planar configuration via the oxygen atoms. Based on Raman spectroscopy data, Healy et al. [14,15] have suggested that during the copper electrodeposition processes, the type of PEG species adsorbed on an electrode surface depends on the applied potential. Specifically, they proposed that neutral PEG molecules are adsorbed at more negative potentials, where copper deposition occurs. At potentials close to the open-circuit potential, however, PEG adsorbs as a copper chloride complex with the polymer acting as a ligand, analogous to a crown ether. In addition, differential capacitance measurements in a PEG-containing H 2 SO 4 electrolyte solution exposed to a copper substrate revealed that PEG was weakly adsorbed onto the copper surface [16]. Using a quartz crystal microbalance (QCM) and electrochemical impedance spectroscopy, Kelly and West [17,18] have observed that the addition of PEG alone to an electrolytic bath had only a small effect on the electrode kinetics, and that Cl - alone promoted the copper deposition reaction. However, they found that when both PEG and Cl - were present in the solution, the PEG monolayer collapsed into spherical aggregates. Using ellipsometry, Bonou et al. [19] have shown that the adsorption of PEG on a platinum electrode surface depended on the applied potential and that PEG was not adsorbed at the open-circuit potential. The simultaneous presence of the additives Cl and PEG have specific effects on metal ion reduction. Nowadays it is know that those effects are related with their capacity to be absorbed onto theelectrode surface. A Thermodynamic studies of PEG adsorption in presence of chlorine anions, is necessary to the understanding of the mechanism by the PEG is adsorbed onto the metallic surface. This knowledge includes a quantitative characterization of the aspects as the formation of the inhibiting film when this species are present, blockage of adsorption sites on the electrode surface and displacement phenomena. In the present study, we used the thermodynamic method of Lipkowski et al. [20-22], who investigated the adsorption of organic molecules onto solid electrodes, to investigate the adsorption of PEG 20,000 and Cl - ions onto polycrystalline Au in a KClO 4 solution and to quantify thermodynamic parameters of the adsorption, such

4 4 Alia Méndez et al. as the surface tension, the relative Gibbs surface excess, and the Gibbs energy of adsorption. 2. Experimental The study of PEG 20,000 adsorption was performed using solutions S 0 (0.1 M HClO 4 ) + x M PEG 20,000 and S 0 with 12.3 mmkcl+ x M PEG 20,000, where x = 0.5, 1, 5, 10, 50, or 100. When PEG 20,000 concentrations outside this range were used, the experiments exhibited poor reproducibility. Specifically, the adsorption process was unstable at concentrations less than 0.5 M, and the viscosity of the solution increased substantially when concentrations in excess of 100 M were used. A stock solution that contained 5 mm PEG 20,000 was prepared by dissolving PEG 20,000 in 0.1 M KClO 4, and another stock solution that contained 1M KCl was prepared by dissolving KCl in water. The solutions were prepared immediately prior to each electrochemical experiment using ultra-pure water (18 M cm -1 ) and analytical-grade reagents (Sigma-Aldrich, in the highest purity available). Before each electrochemical experiment, the solution was deoxygenated for 30 min with ultra-pure nitrogen (Praxair), and the experiments were performed under a nitrogen atmosphere at C. The experiments were performed in a conventional three-electrode cell with a water jacket. A potentiostat/galvanostat (Autolab, model PGSTAT 30) and a QCM (Maxtek, model 710), controlled by independent computers, were used to simultaneously measure the electrochemical parameters and the frequency of the quartz crystal. An AT-cut quartz crystal of nominal frequency f 0 = 5 MHz, which was covered on both sides with an Au film (Maxtek, CA), was used as the working electrode (Au-QCM). The geometric area of the Au-QCM electrode was 1.37 cm 2. The real area of the electrode was 2.98 cm 2, as determined using the method described by Woods 23 ; the surface roughness factor, R f, was All results reported in this work refer to the real area. A saturated calomel electrode (SCE) and a graphite rod were used as the reference and counter electrodes, respectively. To minimize ir-drop effects, the reference-electrode and workingelectrode compartments were connected with a Luggin capillary. The QCM signal was recorded as f, where f = f-f initial, as a function of time and of electrode potential. The experimental frequency change can be expressed as [24,25] f C f m f f r... (Eq.1) where the first term on the right-hand side of Eq. (1) is the Sauerbrey term [26], which represents the total mass change at the electrode surface. Other

5 A study of PEG/Cl adsorption on an Au electrode 5 possible contributions to the frequency change include changes in the solution viscosity ( f ) [27] and the surface roughness ( f ) [28]. r Experimental studies have shown that surface roughness can drastically affect the resonance frequency. This effect was attributed to rougher surfaces trapping a larger quantity of solvent molecules in surface cavities [29-32]. In the present work, the use of polished Au-QCM electrodes (with a roughness of 1.2 nm, as measured by AFM) should minimize the effects of surface roughness, and the effects of viscosity variations are expected to be negligible. Prior to the measurements, the sensitivity factor (C f = Hz cm 2 ng -1 *R f ) of the quartz crystal was determined using the chronoamperometry calibration method described by Vatankhah et al. [33]. 3. Results and discussion 3.1. Characterization of the electrode surface Cyclic voltammetry was used for qualitative characterization of the electrode surface. Prior to each experiment, the Au-QCM electrode in 0.5 M H 2 SO 4 was activated by cycling the potential between the onset potentials of hydrogen and oxygen evolution at 100 mv s -1 until no changes were observed in the voltammogram ( 10 cycles). The Au-QCM electrode was then transferred to a cell containing the supporting electrolyte (S 0 solution (0.1 M KClO 4 )), where the Au-QCM electrode was again activated by cycling the potential between the onset potentials of hydrogen and oxygen evolution at 100 mv s -1 until no changes were observed in the voltammogram ( 4 cycles). The observed behavior was similar to that reported by Stolberg et al. [34] for an Au electrode in a KClO 4 solution. This behavior was interpreted as indicating an active, clean Au-QCM surface Voltammetric and QCM studies Adsorption procedure After a voltammogram typical of an Au-QCM polycrystalline electrode in a S 0 solution was obtained, and after the frequency of the microbalance was observed to remain stable ( f 0) for 160 s in the electrolytic S 0 solution (with constant agitation) at the open-circuit potential, E OCP (0.30 V vs. SCE), the PEG 20,000 stock solution was injected into the S 0 solution to the desired final concentration. The change in the resonance frequency of the quartz crystal was then recorded for 7 min to monitor the adsorption of the additive; agitation was continued throughout this 7 min period. Immediately

6 6 Alia Méndez et al. after the adsorption of PEG 20,000 for 7 min, the necessary quantity of KCl stock solution was added to reach a concentration of 12.3 mm of Cl - ions in the solution, and the adsorption of the Cl - ions was monitored for 7 min. The same procedure was performed for each of the PEG 20,000 concentrations considered in the experiment Voltammetric studies Immediately after the adsorption, we performed simultaneous voltammetric and QCM studies in the potential range 1.0 to 0.6 V vs. SCE. The potential scans were initiated in the negative direction from the opencircuit potential (E OCP ), with a potential scan rate (v) of 50 mv s -1. Figure 1 shows the voltammograms obtained. The profile obtained for the support electrolyte (solution S 0 ; curve a) presents the characteristic behavior of an Au electrode, free of impurities in a KClO 4 [34] solution. Similar behavior was observed in the presence of Cl - ions (solution S 1 ; curve b). In Figure 1, curve c corresponds to the cyclic voltammogram obtained in the presence of PEG 20000, in which the formation of the peaks P A and P C is observed and is associated with the adsorption-desorption processes of the species [35,36]. In contrast, the addition of Cl - ions to the KClO 4 /PEG system (solution S 3 ; curve d) did not Figure 1. Cyclic voltammograms recorded for solutions (a) S 0 (0.1 M KClO 4 ), (b) S 1 (S mmkcl), (c) S 2 (S 0 +5 M PEG ), and (d) S 3 (S 1 +5 M PEG ). Potential range, -1.0 to 0.6 V vs. SCE; scan rate, 0.05 V s -1.

7 A study of PEG/Cl adsorption on an Au electrode 7 significantly modify the shape of the voltammogram or the intensity of the current density in the P A and P C peaks. Therefore, the adsorption-desorption mechanism of PEG 20000, which corresponds to the processes P A and P C, is independent of the presence of Cl - ions in the dissolution QCM studies To verify the adsorption of PEG and Cl - ions in the potential range 1.0 to 0.6 V vs. SCE based on the frequency changes recorded by QCM, we constructed graphs of d m dt -1 vs. E, known as massograms [37] (Fig. 2); d m dt -1 is the rate of mass change (mass flux) that occurs at the electrode surface. For mass changes that are associated with charge-transfer processes at the electrode surface (faradaic processes), d m dt -1 is directly proportional to the current density; hence, a massogram is analogous to a voltammogram. In a massogram, a negative mass flux corresponds to a mass loss (desorption), and a positive mass flux corresponds to a mass gain (adsorption). When PEG or the PEG /Cl - mixture were present in the dissolution (Fig. 2; curves c and d, respectively), a positive mass flux peak Figure 2. Series of massograms for an Au-QCM electrode obtained simultaneously with the cyclic voltammograms in Fig. 1 for solutions (a) S 0 (0.1 M KClO 4 ), (b) S 1 (S mmkcl), (c) S 2 (S 0 +5 M PEG ), and (d) S 3 (S 1 +5 M PEG ).

8 8 Alia Méndez et al. (mass gain) at approximately -0.8 V vs. SCE was observed during the potential sweep in the negative direction (P C(ads) peak). This potential value is similar to that obtained for the reduction process, P C, observed by cyclic voltammetry (Fig. 1). It is important to note that the mass flux density in the P C(ads) peak is independent of the presence of Cl - ions. This result implies that the adsorption velocity of the polymer is not affected by the presence of Cl - ions at the interface. In addition, during the potential sweep in the positive direction, a negative mass flux peak (mass loss, P A(des) peak) was observed at approximately V vs. SCE, which decreased in intensity in the presence of Cl - ions. Therefore, the velocity of PEG desorption from the electrode surface decreases in the presence of Cl - ions. In addition, the potential value (= V vs. SCE) was similar to that obtained for the P A peak during the voltammetric study. The adsorption of PEG in the presence of Cl - ions was further characterized by AC voltammetry, and the results obtained are described in the following section AC voltammetry and QCM measurements Immediately after the adsorption (see 3.2.1), a potential value E (= 1.18 V vs. SCE) was imposed on the electrode. Based on this potential as a starting point, a potential sweep was initiated in a positive direction. Figure 3 shows the typical C diff. vs. E curves obtained after the adsorption of PEG 20,000 and Cl - ions over the surface of a polycrystalline Au-QCM electrode based on the following solutions: S 0 only (0.1 M KClO 4 ) (curve a), S 1 (S mm KCl) (curve b), S 2 (S M PEG 20,000 ) (curve c), and S 3 (S M PEG 20, mm KCl) (curve d). Four potential regions are clearly observed: region I, from to V vs. SCE; region II, from to V vs. SCE; region III, from to V vs. SCE; and region IV, from to 0.7 V vs. SCE. In region I, all curves coincide, which indicates that both PEG 20,000 and Cl - ions are completely desorbed from the electrode surface in this potential range. The C diff. vs. E curve that was obtained solely in the presence of Cl - ions (curve b) represents the typical behavior for the adsorption of Cl - ions over an Au(221) [38] surface. Specifically, two capacitive peaks associated with the adsorption of Cl - ions over an Au surface are observed. The first peak is in the potential interval from to V vs. SCE (corresponding to region III); the second peak is observed in the interval from -0.1 to 0.7 V vs. SCE (region IV) and is larger than the first peak. Adsorption is, in general, admitted to be localized on solid surfaces; therefore, the peaks in the C diff vs. E curves correspond to the adsorption of the Cl - ions onto active sites with a different energy level. The most

9 A study of PEG/Cl adsorption on an Au electrode 9 accessible adsorption sites be it because of their energy or their steric orientation correspond to the most cathodic peak. Conversely, the hardest-to-access adsorption sites correspond to the most anodic peak [38]. When only PEG 20,000 is present (curve c), a pseudo-capacitive peak appears in the interval from -0.9 to -0.6 V vs. SCE (region II). This peak is characteristic of the adsorption desorption processes of neutral organic macromolecules [39]. However, according to Miller and Grahame [40-42], this peak is caused by some segments of the adsorbed macromolecules that were not desorbed during the desorption process. In addition, in the potential range from to V vs. SCE (region III), the capacitance values reach a plateau, which suggests that two-dimensional condensation occurred [11]. The curve with PEG 20,000 and Cl - ions (curve d) shows a pseudocapacitive peak in the interval from -0.9 to -0.6 V vs. SCE (region II) and corresponds to the desorption adsorption processes of PEG 20,000. The intensity of this peak is less than that of the peak observed solely in the presence of PEG 20,000 (curve c); this behavior is associated with a decrease in the velocity of the desorption adsorption of PEG 20,000 when Cl - ions are present. Likewise, the value of the peak potential is independent of the presence of Cl - ions in the solution, which indicates that the activity of the supporting electrolyte is not affected by the presence of Cl - ions. In region III, in the interval from to V vs. SCE, the capacitance was essentially the same relative to that obtained from the S 2 solution, which suggests that two-dimensional condensation occurred. At less-negative potentials, within the interval from to -0.1 V vs. SCE, capacitance increases due to the adsorption of Cl - ions; however, it is still less than that obtained in the sole presence of Cl - ions (curve b). This behavior is associated with the co-adsorption of both PEG 20,000 and Cl - ions, as well as with the formation of the condensed film. This formation of a condensed film can be observed as a characteristic decrease in the double-layer capacitance, which leads to the so-called capacity pit. Molecules with alkyl-ether oxygen atoms are known to adsorb onto metal surfaces and form two-dimensional condensates [11]. In region IV, the pseudo-capacitive peak obtained in the presence of PEG 20,000 and Cl - exhibits a greater intensity than that obtained in the sole presence of Cl -, which indicates that a greater quantity of Cl - ions are adsorbed in this potential region in the presence of PEG 20,000. The same behavior was observed for the different investigated concentrations of PEG 20,000. The results shown in Fig. 3 suggest that the adsorption of both PEG and Cl - ions occurs on active adsorption sites of different energies.

10 10 Alia Méndez et al. Figure 3. Differential capacitance (C diff. ) vs. potential (E) curves obtained for the polycrystalline Au-QCM electrode obtained in solutions (a) S 0 (0.1 M KClO 4 ), (b) S 1 (S mmkcl), (c) S 2 (S 0 +5 M PEG ), and (d) S 3 (S 1 +5 M PEG ). Scan rate, V vs. SCE; AC modulation frequency, 25 Hz. Figure 4 presents plots of m vs. E corresponding to the C diff. vs. E curves shown in Fig. 3. The m values were calculated using Equation 1 and the measurements of the frequency change ( f). The plots of m vs. E exhibit the following characteristics: a) In the interval from to V vs. SCE, the values of superficial mass change are null for all of the considered systems, indicating that in this potential interval, there are no adsorbed species on the Au-QCM electrode surface. This result is in accord with that obtained from the C diff. vs. E curves (see Fig. 3). b) A loss of superficial mass was observed at more positive potentials in the presence of ClO 4 - ions (solution S 0 ; curve a) or in the presence of Cl - ions (solution S 1 (S mm KCl); curve b). This effect is attributed to changes in the interface structure due to the presence of the anions [43]. c) The m vs. E graph obtained from solution S 2 (S µm PEG ) (curve c) exhibited a constant increase in mass change (adsorption) in the interval from -0.9 to 0.40 V vs. SCE. At potentials more positive than 0.40 V vs. SCE, the amount of mass decreased,

11 A study of PEG/Cl adsorption on an Au electrode 11 indicating a superficial desorption associated with the oxidation of the adsorbed PEG species [44,45]. d) In the potential range to V vs. SCE, the curve obtained from solution S 3 (S M PEG ) (curve d) presents a similar increase in mass to that obtained from the S 2 solution (curve c). At potentials higher than V vs. SCE, the adsorbed mass of PEG and Cl - ions was lower than that obtained in the presence of PEG alone (solution S 2 ) in the same potential interval. The results presented in the C diff. vs. E graph (Figure 3) indicate that this behavior is associated with the co-adsorption of both PEG and Cl - ions and with a decrease in the number of sites available for the adsorption of PEG due to the adsorption of Cl - ions. In addition, it is important to note that a decrease in the superficial mass was observed at E = -0.8 V vs. SCE, which is indicative of a desorption process. This potential value (E = -0.8 V vs. SCE) is similar to that at which the peaks P A and P A(des) occur in the voltammograms and massograms and to the potential value of the pseudo-capacitive peak observed in the C diff. vs. E curves. Therefore, we propose that this behavior corresponds to the partial desorption of PEG from the Au-QCM electrode surface m / ng cm a b c d E / V vs SCE Figure 4. Mass change ( m) as a function of potential, obtained simultaneously with the differential capacitance (C diff. ) vs. potential (E) curves in Fig. 3 for solutions (a) S 0 (0.1 M KClO 4 ), (b) S 1 (S mm KCl), (c) S 2 (S 0 +5 M PEG ), and (d) S 3 (S 1 +5 M PEG ).

12 12 Alia Méndez et al Thermodynamic study We used the chronocoulometric method proposed by Lipkowski et al. [20-22] to describe the adsorption of PEG 20,000 in the presence of Cl - ions onto polycrystalline Au-QCM in a KClO 4 solution and to quantify the thermodynamic parameters of the adsorption, such as the surface tension and film pressure. The electrocapillary equation for the Au-QCM electrode in equilibrium with the electrolyte containing PEG 20,000 and Cl - ions can be written as: dγ QdE (Γ Cl(ads) Γ Cl )dμ KCl (Γ PEG Γ )dμ 20000(ads) PEG20000 (sol) PEG20000 (Eq.2) Where is the Gibbs excess, is the interfacial tension, and Q is the total charge density. According to Eq. (2), the relative Gibbs surface excess of adsorbed PEG 20,000 ( ) can be determined by differentiation of with respect PEG ( ads ) to PEG : PEG20000 ( ads) PEG20000 ( sol) PEG20000 E, T, P, KCl (Eq.3) where PEG ( sol ) represents the Gibbs excess of PEG 20,000 present in the diffuse part of the double layer. Because KClO 4 is present in excess, the following assumptions can be made: (a) The Gibbs excess of the PEG 20,000 present in the diffuse part of the double layer can be considered to be negligible, and the total Gibbs excess can be considered to be equal to PEG ( ads ). (b) The activity coefficients of PEG 20,000 do not change with the concentration of PEG 20,000 ; hence, d RTd c. PEG ln PEG When the concentration of KCl is kept constant, the chemical potential of the Cl - ions does not change with the PEG 20,000 concentration. Equation (3) shows that the Gibbs excess of PEG 20,000 adsorbed in the presence of Cl - ions can be determined if the change in the interfacial tension due to a change in the bulk PEG 20,000 concentration can be measured. The relative interfacial tension was determined from a measurement of the total

13 A study of PEG/Cl adsorption on an Au electrode 13 charge density using the back-integration procedure proposed by Lipkowski et al. [20-22,46,47] Determination of the potential of zero charge (PZC) To examine the PEG adsorption in more detail and to determine the potential of zero charge (PZC) of the electrode, AC voltametry measurements were performed. Fig. 5 shows the C diff. versus potential (E) curves obtained for the polycrystalline Au-QCM electrode in the presence of 0.1,0.01 and M KClO 4. The capacitance curve obtained in the presence of 0.01 M KClO 4 displays a diffuse layer minimum at approximately V vs. SCE, which correspond to the PZC. This value is in excellent agreement with those reported in the literature for polycrystalline gold electrodes [20,47]. Figure 5. Differential capacitance (C diff. ) vs potential (E) curves obtained for the polycrystalline Au-QCM electrode in the presence of: a) 0.1, b) 0.01 and c) MKClO QCM measurements Chronocoulometry measurements were performed to determine the surface charge density on the metal between two potential values: E i (initial potential) and E f (final potential). The values of E i and E f for the potential

14 14 Alia Méndez et al. step experiments were chosen by considering the C diff. vs. E curves (see Fig. 3). The value of E i corresponded to the potential at which the adsorption of both PEG 20,000 and/or Cl - ions takes place and varied from -0.9 to 0.35 V vs. SCE. The E f (-1.18 V vs. SCE) corresponded to the potential at which all PEG 20,000 and Cl - ions are desorbed from the electrode surface. To verify that the adsorption of PEG 20,000 and Cl - ions onto the electrode surface at each E i value had reached a steady state, simultaneous measurements of the mass change on the electrode surface were performed using QCM. The study of PEG 20,000 adsorption in the presence of Cl - ions on the polycrystalline Au-QCM electrode surface was performed according to the following procedure: After the voltammogram typical of an Au-QCM polycrystalline electrode in a S 0 solution (see sec. 3.1) was obtained, an adsorption potential value E i was imposed on the electrode. After the frequency of the microbalance had remained stable ( f 0) for 160 s in the electrolytic S 0 solution (with constant agitation) at E i, the PEG 20,000 stock solution was injected into the S 0 solution to the desired final concentration (i.e., 0.5, 1, 5, 10, 50 or 100 M PEG 20,000 ). The change in the resonance frequency of the quartz crystal was then recorded for 7 min to monitor the adsorption of PEG 20,000 ; agitation was continued throughout this 7-min period. Immediately afterwards, the necessary amount of KCl was added to obtain a 12.3 mm concentration of Cl - ions in the solution, and the adsorption of Cl - ions was monitored for 7 min. Figure 6 shows typical m vs. t ( m was calculated using Eq. (1)) plots obtained for all of the evaluated concentrations and at three potential values of adsorption (E i ): (Fig. 6a), (Fig. 6b), and 0.29 V vs. SCE (Fig. 6c). These potentials correspond to regions II, III, and IV in the AC impedance graphs (Fig. 3). The behavior exhibited in the graphs evinces the following characteristics: Immediately after the addition of the PEG 20,000, the mass on the electrode surface increases rapidly and, after 7 min, reaches a stationary state; this behavior is characteristic of PEG 20,000 adsorption on an Au-QCM [35] surface. However, when the Cl - ions are added, the adsorption behavior depends on the imposed potential. Specifically, when the imposed adsorption potential was V vs. SCE (Fig. 6a), the addition of the Cl - ions to the solution caused a decrease in the adsorbed mass. This behavior is associated with the alteration of the structure of the double layer due to the presence of the Cl - ions. This behavior is followed by the observed competition between the formation of the condensed phase of adsorbed PEG 20,000 and the adsorption of Cl - ions. Given the negative polarity of the surface in this region of potentials, Cl - ions are rejected by the diffuse double

15 A study of PEG/Cl adsorption on an Au electrode 15 layer, and the formation of the condensed phase of PEG 20,000 is re-established. Consequently, the value of the adsorbed mass returns to the value obtained before the addition of the Cl - ions. In addition, Figure 6a shows the curve obtained when PEG 20,000 is absent, where the observed loss of mass in the electrode s surface is associated with the simultaneous movement of the adsorbed ClO - 4 ions and the adsorption of Cl - ions. When the adsorption potential was V vs. SCE (Fig. 6b), which corresponds to region III of the capacitance curves (Fig. 3), a loss of mass was observed immediately after the addition of the Cl - ions, followed by the presence of a stationary state. This result indicates that the adsorptive equilibrium is reached. For concentrations within the interval of 1 to 100 M of PEG 20,000, the amount desorbed by the presence of Cl - ions in the interphase is similar, even when the sub-monolayer of PEG 20,000 is formed. This result indicates the co-adsorption of PEG 20,000 and Cl - ions on the Au surface, given that the adsorption of these two types is selective to the sites available with a particular energy level [48-49]. Therefore, the loss of mass observed corresponds to the movement of the ClO - 4 ions from the electrode s surface by means of the adsorption of the Cl - ions. Likewise, after 14 min of adsorption, the adsorbed mass decreased considerably for the lowest concentration, which indicates that an important portion of the adsorbed PEG 20,000 was displaced by Cl - ions at this concentration.

16 16 Alia Méndez et al. Figure 6. Mass change ( m) as a function of time after the addition of Cl - ions (12.3 mm) and various concentrations of PEG 20,000 to S 0 solutions while maintaining a constant electrode potential: a) V vs. SCE, b) V vs. SCE, c) 0.3 V vs. SCE.

17 A study of PEG/Cl adsorption on an Au electrode 17 When the adsorption potential was 0.29 V vs. SCE (region IV) (Fig. 6c), the mass quantity of PEG 20,000 desorbed by the addition of Cl - ions to the solution is greater than that obtained from a solution without PEG 20,000 (S 1 solution). This result indicates that the interaction between the substrate and the Cl - ions at these potentials is sufficiently strong to provoke a displacement of the PEG 20,000 molecules adsorbed on the surface. Later, an increase in the adsorbed mass was observed; this increase is associated with the adsorption of Cl - ions onto the surface of the Au-QCM electrode Chronocoulometry measurements Immediately after the 14-min adsorption period, agitation was stopped, and the potential was stepped down to E f (-1.18 V vs. SCE) the potential at which both PEG 20,000 and Cl - are desorbed from the electrode for 200 ms to avoid the charge contribution from hydrogen evolution. The time window in which the transients were recorded was the same for the supporting electrolyte as for solutions that contained PEG 20,000 and Cl - ions. Figure 7a shows several current time transients obtained for a polycrystalline Au- QCM electrode in a S M PEG 20, mmkcl solution at various adsorption electrode potentials (E i ). The transients exhibit a form typical of the charging of a capacitor in a simple series RC circuit. The current decay through the step down (i.e., E i to E f ) was integrated to determine the relative charge density ( M(E i )) that results from the adsorbed PEG 20,000. Figure 7b shows the chronocoulometric curves ( M vs. t) obtained from the data in Fig. 7a. The curves display an initial fast-rising section that corresponds to the charging of the double layer, followed by a quasi-plateau in which the charge varies linearly with time when slowing. The linear portion of the transient was extrapolated to zero on the time axis. In this way, the relative charge densities M(E i ) were determined. This method ensured that the extrapolated charge (relative charge densities M(E i )) was equal to the charge difference between the electrode charge densities at E f and a given value of E i. M E i M E i M E f 1.18V vs. SCE (Eq. 4) Using the value of the point of zero charge (pzc), which was determined independently on the basis of the differential capacitance curves (see 3.4.1), the absolute charge densities ( M ) were calculated for each potential value using the following equation:

18 18 Alia Méndez et al. Figure 7. (a) Current density time transients recorded for S M PEG 20, mm KCl solution over the potential interval -0.9 to 0.3 V vs. SCE. (b) Charge density time transients obtained by integration of the current density time transients shown in (a).

19 A study of PEG/Cl adsorption on an Au electrode 19 M ( Ei ) M ( Ei ) M ( E pzc) (Eq. 5) Figure 8 shows a family of absolute charge density curves for several PEG 20,000 concentrations that were determined from the previously described chronocoulometric experiments. The charge densities were measured by waiting at the potential where PEG 20,000 and Cl - ions are adsorbed for a period of time sufficient for the adsorption equilibrium to be established (14 min). Therefore, the curves represent the state of adsorption equilibrium. The absolute charge density plot measured in the presence of Cl - and several PEG 20,000 concentrations is composed of several sections. At potentials greater than zero, the co-adsorption of PEG 20,000 and Cl - ions apparently causes a positive charge to flow to the metal side of the interface; this charge is smaller than that observed during the adsorption of only Cl - ions. At potentials more negative than -0.1 V vs. SCE, the charge flow on the electrode is negative and increases according to the increase in the PEG 20,000 concentration in the solution. In the interval from -0.1 to 0.20 V vs. SCE, a change in the slope of the charge-density potential curves occurs. This change may be related to a transition between the adsorbed phases at different potential ranges. For potential values greater than -0.4 V vs. SCE, the curve obtained at the lowest PEG 20,000 concentration (0.5 M PEG 20,000 ) exhibits a behavior similar to that of the electrolyte that contained only Cl - ions. This result suggests that, at these potentials, the sub-monolayer of PEG 20,000 formed prior to the addition of Cl - ions has been displaced by the Cl - anions, which agrees with the result obtained by QCM at this same concentration (see Fig. 6b). Additionally, the behavior of the curve obtained from the KClO 4 + KCl solution demonstrates that the adsorption of the Cl - ions begins at potentials greater than -0.6 V vs. SCE, which agrees with the results obtained using the AC voltammetry technique (see Fig. 3) Electrocapillary curves The surface tension ( (E)) at each potential can be calculated by integration of the absolute charge density with respect to the electrode potential using the following equation: Ei ( E) M de ( E f ) E f (Eq. 6)

20 20 Alia Méndez et al. Figure 8. Absolute charge density as a function of electrode potential, as determined by chronocoulometry, for solution S 1 (0.1 M KClO mmkcl) with different PEG 20,000 concentrations. The lower integration constant, ( E f ), is not known, but its value is independent of the presence of the PEG 20,000 molecules and Cl - ions because no adsorption occurs at ( 1.18 V vs.sce). Figure 9 shows the relative E f electrocapillary curves obtained from the S mmkcl + x M PEG 20,000 solutions (x = 0.0, 0.1, 1, 5, 10, 50, or 100). An examination of Fig. 9 reveals that the surface tension decreases considerably as the concentration of PEG 20,000 increases an effect characteristic of surfactant additives (surface active agents) used in metal electrodeposition [50-53]. Likewise, the addition of Cl - to the solution containing variable concentrations of PEG 20,000 shifts the potential of zero charge (Epzc) (electrocapillary maximum) toward negative potentials and points to a clear domination of Cl - ions in the formation of the adsorption film Film pressure ( ) The film pressure,, as a function of the electrode potential can now be calculated from the values of the surface tension using the follow equation:

21 A study of PEG/Cl adsorption on an Au electrode 21 Figure 9. Electrocapillary curves for PEG 20,000 and Cl - ions adsorption onto a Au-QCM polycrystalline electrode in contact with different solutions. ( i 0 i i E ) ( E ) ( E ) (Eq. 7) Where 0( Ei ) is the surface tension obtained for the electrolyte support (KClO 4 ) at different potential adsorption values. The calculated values of film pressures obtained from the solutions of composition 12.3 mm KCl + x M PEG 20,000 are plotted as a function of the electrode potential in Fig. 10a. These curves do not show the typical bell shape usually observed for the adsorption of neutral organic molecules, such as those observed in the case of PEG 20,000 [35]. Hence, the adsorption of PEG 20,000 in the presence of Cl - ions is more typical of anion adsorption than of adsorption of a neutral organic molecule. To determine the influence of Cl - ions on the PEG 20,000 adsorption process based on the adsorption potential applied to the electrode, a comparison was performed between the behavior of the vs. E i curves obtained from the solutions with the PEG 20,000 + KCl mixture and those obtained from solutions that contained only the PEG 20,000 surfactant (Fig. 10b). The PEG 20,000 adsorption at the polycrystalline gold electrode from chlorine-free solutions has been described in a previous article [35].

22 22 Alia Méndez et al. Figure 10b shows a different behavior based on the applied potential within the potential interval of -1.0 to -0.6 V vs. SCE. In this region, the values of obtained for both solutions are independent of the presence of Cl - ions in the solution. This result indicates that, in this interval potential, only PEG 20,000 molecules are adsorbed over the surface of the Au electrode. These results coincide with those obtained using AC impedance (Fig. 3) and QCM (Fig. 4) in this potential region. In the interval from -0.6 to V vs. SCE, the values of obtained in the absence and presence of Cl - ions follow a similar trend. A more detailed inspection of these values revealed that the pressure values of the film obtained for the mixture are similar to those obtained from the expression mix PEG20000 Cl. This additive tendency corroborates the data obtained by QCM in this potential region (Fig. 4) and thus confirms the co-adsorption of the PEG molecules and Cl - ions in this region (region III, Fig. 3). The curves exhibit a change in their incline relative to the values obtained in absence of the anion in the interval from to 0.35 V vs. SCE. This result points to a clear domination of Cl - ions in the formation of the film in this region of potential (region IV, Fig. 3).

23 A study of PEG/Cl adsorption on an Au electrode 23 Figure 10. a) Film pressure curves ( ) determined with respect to the electrode potential in solution S 1 (0.1 M KClO mmkcl) with various PEG 20,000 concentrations. b) Comparison of vs. E curves associated with PEG 20,000 adsorption based on solutions with different concentrations in the absence and presence of Cl - ions Adsorption isotherm Film pressure data were used to calculate the relative Gibbs surface excess of adsorbed PEG 20,000 ( PEG ( ads ) ) by differentiation of the film pressure ( ) vs. the logarithm of the bulk concentration curves at each adsorption potential (E i ) from the following equation: RT C PEG ( ads) ln PEG (Eq. 8) T, P, E i, KCl The relative Gibbs surface excess is plotted as a function of potential in Fig. 11. Clearly, the behavior depends on the PEG 20,000 concentration and the electrode potential. Specifically, the superficial excess increases slightly within the interval from to -0.6 V vs. SCE for PEG 20,000 concentrations less than or equal to 10 M. For higher concentrations, the surface excess increases monotonically. The decrease in surface excess is observed in the

24 24 Alia Méndez et al. Figure 11. Curves of relative Gibbs surface excess as a function of electrode potential, associated with the adsorption of PEG 20,000 in the presence of Cl - ions. interval from -0.1 to 0.3 V vs. SCE. This behavior is associated with the desorption of PEG 20,000 from the surface of the electrode, which is caused by the displacement of the adsorbed Cl - ions. In studies on the adsorption of a poly (acrylic acid) sodium salt at the air water interface using tensiometry and X-ray reflectivity, Millet at al. [54] have reported that the surface excess predicted using the Gibbs equation (Eq. (8)) only measured the concentration of chain segments located very close to the interface, which constituted the trains of the adsorbed layer. Hence, based on the theoretical analysis that corresponds to the optimization of the molecular geometry and to the calculation of the highest occupied molecular orbital (HOMO) using the density functional theory (DFT) (Fig. 12), we propose that the PEG 20,000 molecules should bind to the surface via the hydrogen atoms at potentials more negative than -0.1 V vs. SCE (Fig. 13a). Therefore, the relative Gibbs surface excess due to the adsorption of the polymer at the Au-QCM polycrystalline electrode surface is associated with the hydrogen atoms of the adsorbed monomer units. At potentials greater than -0.1 V vs. SCE, the relative Gibbs surface excess due to the adsorption of the polymer at the Au-QCM polycrystalline electrode surface is associated with the oxygen atoms [55-57] of the adsorbed monomer units (Fig. 13b).

25 A study of PEG/Cl adsorption on an Au electrode 25 Figure 12. Molecular geometry optimization and HOMO obtanined for a representative moity in a PEG molecule. Oxygen atoms are represented in red.

26 26 Alia Méndez et al. Figure 13. a) Representation of the adsorption of a PEG molecule when the electrode is negatively charged. b) Representation of the adsorption of PEG molecule when the electrode is positively charged. For the highest PEG 20,000 concentration studied (100 M), the maximum value of the relative Gibbs surface excess was mol oxygenadsorbed cm -2 at E = -0.1 V vs. SCE, which corresponds to molecules of adsorbed PEG 20,000 when the fact that each molecule of PEG ( ads ) assumes the planar zigzag conformation is taken into account [58]. In addition, because PEG 20,000 adsorbs as a flat monolayer of rod-like molecules, the closest packing arrangement would be for the PEG 20,000 rods to lie next to one another; hence, the area occupied by each rod can be approximated as nlw, where n (= 455) is the number of repeat units, L is the length of a monomer unit ( cm), and W is the width of a monomer unit ( cm) [16,17]. Thus, the area occupied by one rod is 60 nm 2, and the highest possible total area occupied by the adsorbed

27 A study of PEG/Cl adsorption on an Au electrode 27 PEG 20,000 molecules is 0.34 cm 2, which corresponds to 11.73% of the real area of the electrode. When the concentrations of PEG 20,000 injected into the solution were 10 M and 0.5 M, the percentage of the highest possible real area of the electrode covered by adsorbed PEG 20,000 molecules at E = -0.1 V vs. SCE decreased to 6.73% and 3.47%, respectively. Figure 14 shows the values obtained for the Gibbs energy of adsorption, G ads, using the Henry adsorption isotherm: max c RT (Eq. 9) Where is the surface pressure, max is the limiting surface concentration, c is the bulk PEG 20,000 concentration, and is the equilibrium constant evaluated from the initial slopes of plots of the surface pressure versus the bulk concentration. The equilibrium constant is related to the Gibbs energy of adsorption by the following equation: G ads RT ln (Eq. 10) Figure 14. Gibbs energy of adsorption for PEG 20,000 onto polycrystalline gold in solution S 0 (0.1 M KClO mmkcl) with various PEG 20,000 concentrations, calculated at 298 K from the Henry adsorption isotherm.

28 28 Alia Méndez et al. A max value of mol oxygen adsorbed cm -2 at E = -0.1 V vs. SCE was employed to calculate using the highest bulk concentration (100 M PEG 20,000 ). As shown in Fig. 14, the absolute value of G ads increases from approximately 30 kj mol -1 to 39 kj mol -1 with increasing potential in the interval of -0.9 to -0.3 V vs. SCE. Based on the values of G ads obtained in this interval of potential, we propose that the PEG 20,000 molecules are adsorbed to the polycrystalline Au-QCM substrate through physisorption via hydrogen bridges. In the same way, at potential values greater than -0.1 V vs. SCE, the PEG 20,000 molecules are adsorbed via the oxygen atoms, and the values of G ads ( 39 kj mol -1 ) correspond to a weak chemisorption. 4. Depiction of PEG and Cl - ions adsorption during Cu deposition In accordance to the information presented above, it is possible to propose a representation for the adsorption of PEG and Cl - ions during Cu deposition. This proposal is in agreement with other models that based on kinetic assumptions explain the additive function to obtain the required bottom-up fill in interconnect metallization [59, 60]. Since copper deposition occurs near V vs. SCE [11], which is within the studied potential interval of region II, the information here obtained can be used to give a better understanding of the synergistic effect produced between PEG and chloride ions during Cu deposition. From AC voltammetry, it was shown that from -0.6 to 0.6 V vs. SCE, both PEG and Cl - ions are adsorbed in an independent way and into different energetic active sites. PEG macromolecules are adsorbed into the more exposed sites, such as convex and flat sites, meanwhile Cl - ions adsorb into sites of more difficult access like the bottom of the feature. On the other hand, a superficial saturation was achieved for the two highest PEG concentrations studied (100 and 50 µm), in the presence of 12.3 mm of KCl. This is demonstrated in the electrocapillary curves where the surface tension for these two concentrations is practically the same. Using Quartz Crystal Microbalance it was proved that even when the PEG film has been previously formed, the Cl - ions can be adsorbing on specific active sites of the surface, suggesting a filtration phenomenon by the anions. Finally, through AFM measurements it was possible to calculate the average size of the features present on the Au-QCM electrode (155 nm wide x 64 nm deep trench). Thus, taking into account that PEG molecule area is 60 nm 2, a scheme illustrating the adsorption of PEG and Cl - ions is suggested in Fig. 15.

29 A study of PEG/Cl adsorption on an Au electrode 29 Figure 15. Representation of the adsorption of PEG molecules and Cl - ions onto an Au-QCM electrode. As depicted in Fig. 15, Cl - ions adsorb on surface sites inside the features, meanwhile PEG molecules adsorb into protruding surfaces. It is widely reported that Cl - ions promoting metallic deposition and suppressors such as PEG form a protective film on upward superficial parts inhibiting metallic discharge. Therefore, it is expected that current density at the bottom is high while the current density at the top surface is suppressed, leading to a perfect bottom up filling that results in high-grade commercial level coatings. 5. Conclusions The present study examined the adsorption of PEG 20,000 in the presence of Cl - ions through an analysis of quantitative data obtained from a quartz crystal microbalance, differential capacitance, and chronocoulometry. A thermodynamic analysis of the charge-density data yielded plots of the surface tension, relative pressure of the film, relative Gibbs surface excess, and adsorption energy as a function of the applied potential. The results obtained provided basic thermodynamic data for the adsorption of PEG 20,000 in the presence of Cl - ions. The results obtained using AC voltammetry and QCM showed that, in the interval from -1.0 to -0.6 V vs. SCE, the presence of Cl - ions reduces the rate of desorption/adsorption of PEG 20,000. At the same time, in the interval from -0.1 to 0.35 V vs. SCE, the presence of both PEG 20,000 and Cl - ions increases the adsorption of Cl - ions. The results from the study based on chronocoulometry showed that, in the interval from -1.0 to -0.6 V vs. SCE, only the adsorption of PEG 20,000 occurs, thus forming a condensed film in two dimensions. As such, the