Enhancement of the Catalytic Activity of Fe Phthalocyanine for the

Jun 27, 2012 - *E-mail: [email protected] (J.P.); [email protected] (F.M.); .... Herre S. J. van der Zant , Diana Dulić , Núria Aliaga-Alcalde , Mo...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/JPCC

Enhancement of the Catalytic Activity of Fe Phthalocyanine for the Reduction of O2 Anchored to Au(111) via Conjugated Self-Assembled Monolayers of Aromatic Thiols As Compared to Cu Phthalocyanine Ingrid Ponce, J. Francisco Silva, Ruben Oñate, Marcos Caroli Rezende, Maritza A. Paez, José H. Zagal, and Jorge Pavez* Facultad de Química y Biología, Departamento de Química de los Materiales, Universidad de Santiago de Chile, Casilla 40, Correo 33, Santiago 9170022, Chile

Fernando Mendizabal* and Sebastián Miranda-Rojas Departamento de Química, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile

Alvaro Muñoz-Castro and Ramiro Arratia-Pérez* Doctorado en Fisicoquimica Molecular, Relativistic Molecular Physics (ReMoPh) Group, Universidad Andres Bello, Av. Republica 275, Santiago, Chile S Supporting Information *

ABSTRACT: We have prepared self-assembled monolayers (SAMs) of 4-aminothiophenol (4-ATP) and 1-(4-mercaptophenyl)-2,6-diphenyl-4(4-pyridyl)pyridinium tetrafluoroborate (MDPP) functionalized with iron phthalocyanine (FePc) and copper phthalocyanine (CuPc) adsorbed on gold (111) electrodes. The catalytic activity of these SAMs/MPc was examined for the reduction of O2 in aqueous solutions and compared to that of bare gold and with gold coated directly with preadsorbed MPc molecules. Scanning tunneling microscopy (STM) studies confirm the functionalization of the 4-ATP by MPc. STM images reveal that iron phthalocyanine molecules are chemically anchored to 4-aminothiophenol organic monolayers, probably having an “umbrella” type orientation with regards to the surface. The electrocatalytic studies carried out with Au/4-ATP/FePc and Au/ MDPP/FePc electrodes show that the O2 reduction takes place by the transfer of 4-electron to give water in contrast to a 2-electron transfer process observed for the bare gold. The modified electrode obtained by simple adsorption of FePc directly to the Au(111) surface still promotes the 4-electron reduction process, but it shows a lower activity than the electrodes involving SAMs with FePc molecules positioned at the outmost portion of the selfassembled monolayers. The activity of the electrodes increases as follow: Au < Au/FePc < Au/4-ATP/FePc < Au/MDPP/FePc with the highest activity when FePc molecules are more separated from the Au surface. In contrast, the less active CuPc shows almost the same activity in all three configurations. Theoretical calculations suggest the importance of the back-bonding into the adduct formation, showing the relevance of the supporting gold surface on the electron-transfer process mediated by anchoring ligands.

1. INTRODUCTION Metallophthalocyanines have been extensively investigated as electrocatalysts for the O2 reduction (ORR).1−13 Together with other similar molecules, they belong to the “non-precious metal catalysts” class and are studied with the aim of replacing expensive Pt-based catalysts in the cathode of fuel cells. Such MN4 chelates, when present on electrode surfaces, catalyze the ORR, but they lack long-term stability in the aggressive environment of a fuel cell, and new heat-treated MN4 materials have been developed to solve this problem.8 O2 reduction in © 2012 American Chemical Society

aqueous media can undergo a 2-electron reduction to give peroxide, or 4-electron reduction to give water.2−4,7,10,11 However, the 4-electron reduction releases more energy than does the 2-e reduction process.2 In this context, Fe phthalocyanines catalyze the reduction of O2 directly to water via 4-electrons, with cleavage of the O−O bond in contrast to Received: February 2, 2012 Revised: June 27, 2012 Published: June 27, 2012 15329

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

Figure 1. (a) SAMs of 4-aminothiophenol (4-ATP) and 1-(4-mercaptophenyl)-2,6-diphenil-4-(4-pyridyl)pyridinium tetrafluoroborate (MDPP) on Au(111) functionalized with a phthalocyanine, where M = Fe, Cu.

what is observed with Co phthalocyanines.2−4,7,10 Some authors have suggested that FePc catalyzes peroxide decomposition,14−17 whereas other have suggested that FePc phthalocyanines form Fe−O−O−Fe bridges promoting the cleavage of the O−O bond.13 It is still not clear how these molecules adsorb on graphite, so the formation of Fe−O−O−Fe bridges is highly speculative.13 Nevertheless, recent studies showed that the Fe center of hemoglobin interacts with O2 adopting a sideways configuration (side-on binding mode); hence both oxygen atoms form a three-center bond with Fe,18 suggesting that this type of interaction lowers the activation energy for the 4-electron reduction process. On the other hand, many methods has been reported to obtain solid films of metal macrocycles as they catalyze many electrochemical reactions,1−4,7,10−13,19−22 direct adsorption,1−4,7 dip-coating,19,20 Langmuir−Blodgett,21,22 spin-coating,23 electropolymerization,24−28 thermoevaporation,29 and self-assembled monolayers (SAMs).18,30−41 When using SAMs, the macrocyclic complex can be immobilized on a gold substrate essentially by two methods: (i) by using thiol functionalities (−SH) incorporated on the periphery of the macrocyclic ligand that work as anchors to form the SAMs,and (ii) by performing a SAM of thiols and then binding to this the macrocyclic complexes via, for example, an amino group present on the other end of the thiol molecule.38,39 The second approach has some advantages because it avoids the timeconsuming preparation of macrocycle molecules with thiol functionalities on the ligand. There are several reports in the literature that use a preformed SAM of 4-aminothiophenol (4ATP) to immobilize Co porphyrins30,34 and Fe phthalocyanines.35−39 In the case of Co porphyrins, these molecules are organized parallel to the gold surface30 showing an “umbrella” configuration (see Figure 1). The SAMs containing Co porphyrins promote the reduction of O2 via the transfer of two-electrons, leading to hydrogen peroxide.18,28,34 Mauzeroll et al.34 used scanning electrochemical microscopy (SECM) on SAMs 4-ATP-CoTPP (CoTPP = Co(II)5,10,15,20-tetraphenyl21H,23H porphine) to locally characterize the production of peroxide during O2 reduction. From a theoretical point of view, density functional theory methods have been applied to study the O2-binding abilities of transition metal macrocyclic complexes (phthalocyanine, porphyrin, and heme).40−43 The usual redox centers in the phthalocyanine macrocycles are Fe and Co atoms. The binding of O2 to FePc has been found to be energetically favorable on

both end-on and side-on configurations.43−46 However, the end-on configuration (i.e., one O atom bound to Fe) is more stable than the side-on configuration. The activation barrier for the O−O bond cleavage for the side-on MN4-O2 configuration is lower than the corresponding with the end-on orientation. Related to this, we have developed theoretical models based on DFT calculations for the formation of SAMs on a gold substrate and a thiolate ligand as an “anchoring” fragment of the metallophthalocyanine, leading to an interesting charge donation from the 4-ATP or 4-MP toward both gold substrate and the phthalocyanine molecule, denoting an effective gold− MPc interaction mediated by the tilted anchor ligands.39 In this work, we will use this model for studying the oxygen reduction reaction and quantifying the effect of the gold electrode surface. We have examined the catalytic activity of Fe−phthalocyanine for O2 reduction adsorbed directly on Au(111) and also anchored on SAMs formed on Au(111), to control the configuration of these molecules on the electrode surface. We have also investigated the less active Cu−phthalocyanine as an attempt to understand the role of the anchoring molecules on the catalytic activity. We combine electrochemical measurements and DFT calculations to elucidate the mechanisms of the ORR on gold-supported Fe phthalocyanine in different spacial configurations.

2. EXPERIMENTAL SECTION 2.1. Electrochemical Measurements. Electrochemical experiments were conducted with nanopure water solutions purified with a MilliporeMilli-QBiocel Ultrapurewater system fed with distilled water. For electrochemical measurements, solutions were purged with ultra nitrogen or ultrapure O2 for 45 min prior to each measurement, depending on the experiment. 4-Aminothiophenol (4-ATP, 97%) and Fe phthalocyanine were obtained from Aldrich. 1-(4-Mercaptophenyl)-2,6-diphenyl-4-(4-pyridyl)pyridinium tetrafluoroborate (MDPP) was synthesized according to a method described in the literature (see the Supporting Information).47 Tetrahydrofuran (THF) was obtained from JT Baker and used as provided. All other reagents were of analytical grade and used without further purification. Gold slides were annealed using a H2 flame, to obtain preferential Au(111) orientation. Au/FePc and Au/SAMs/FePc electrodes were prepared using procedures reported by Ozoemena and Nyokong.33−37 To obtain Au/FePc electrodes, the Au(111) electrodes were placed for 12 h in a 1.0 mM FePc under a constant flow of pure nitrogen. To 15330

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

O2−MPc−L were fully optimized in all calculations. In addition, we have used two small models, the first without spacer ligands (MPc−Au26) and the second without Au26 cluster and spacer ligands (MPc). The purpose of these models is to compare the effect of Au26 cluster and spacer ligands in the original models. The PBE (Perdew−Burke−Ernzerhof) nonlocal exchangecorrelation functional was employed in all of the calculations,50 by using the Turbomole 5.9 program.51 We used the PBE functional to be consistent with our previous publication and also for its improved description of long-range interactions.39 For heavy elements Au, Fe, and Cu, the Stuttgart small-core pseudorelativistic effective core potentials (ECPs) were used: 19 valence-electrons (VE) for Au, 16 for Fe, and 19 for Cu.52 The orbitals associated with the ECP are all Gaussian type 31*G or 31**G; in addition, two f-type polarization functions were added for Au (αf = 0.20, 1.19). The C, N, O, and S atoms were also treated with pseudopotentials, using a double-ζ basis set and adding one d-type polarization function.53 For hydrogen, a valence-double-ζ basis set with one p-polarization function was used.54 The spin unrestricted method was used for all open-shell systems. With the purpose of gaining more insight on the ORR process, we carried out calculations on the first step of ORR, which involves the adsorption of O2 on the metal center of the MPc molecules, employing the MPc−L−Au26 and MPc−L models. The adsorption energy (ΔEad) is defined as follows:

avoid the formation of FePc precipitates or microcrystals on the gold surface, the electrodes were rinsed with the pure solvent to eliminate any excess of phthalocyanine molecules. Au(111)/4ATP/FePc and Au(111)/MDPP/FePc were prepared as follows: after annealing, the Au slides were placed in an ethanol 50 μM solution containing 4-ATP or MDPP. The time required to obtain the SAMs with any of the thiols used was 24 h. After this period, the electrodes were further rinsed with ethanol for 20 min and dried using a nitrogen flow. The electrodes modified with SAMs were characterized using cyclic voltammetry and STM. The Au(111)/SAMs electrodes were further modified with FePc by incubating the preformed Au(111)/SAMs for 4 h in a 1 mM solution of FePc in THF. After that, the Au samples were placed for 5 min in THF under a smooth flow of pure N2. After all of these procedures, the electrodes were used immediately for the different electrochemical testings. Surface characterization was conducted using STM with a ECM-2 microscope (Veeco, U.S.) employing commercial cut Pt−Ir tips (Veeco probes, 0.25 mm in diameter, storage in O2-free atmosphere previous to use). An off-line plane-fit and low-pass filter was applied to the STM images. The cyclic voltammetry experiments were performed with a Bio-Analytical Systems, BasI-Epsilon electrochemical workstation, using a conventional three-electrode electrochemical cell. The working electrode was a thin vapor deposited Au film deposited on glass (12 × 12 mm slides purchased from Arrandee, Germany). The reference electrode was Ag/AgClsat, and a platinum wire of 5 cm2 geometrical area served as the counter electrode. Different tests of the electrocatalytic activity for O2 were performed in freshly prepared 0.1 M NaOH solutions. All experiments were performed at 25 °C. 2.2. Theoretical Calculations. The interaction of the metallophthalocyanine−L, MPc (Fe, Cu; L = 4-aminothiophenol (4-ATP) and 1-(4-mercaptophenyl)-2,6-diphenyl4-(4-pyridyl)pyridinium (MDPP)), moiety with the Au(111) surface was achieved by using an Au26 cluster as previously discussed.39 The MPc−L−Au26 clusters are shown in Figure 2. Note that the 4-ATP and MDPP molecules are acting as a thiolate. The cluster is composed by three layers containing 14, 8, and 4 gold atoms, respectively.48,49 The positions of the Au atoms were fixed on the basis of the bulk structure (Au−Au distances fixed at 2.871 Å). The geometries of the MPc−L and

ΔEad = EO2 ‐ MPc − (EO2 + EMPc)

(1)

where EO2‑MPc is the energy of the optimized structure of O2 adsorbed on MPc−L−Au26 and MPc−L, while EO2 and EMPc are the energies of the isolated O2, MPc−L−Au26, and MPc−L fragments, respectively. The following two steps consist of the reduction of the oxygen molecule and the formation of a HO2intermediate, respectively: ΔE2 = E−O2 ‐ MPc − EO2 ‐ MPc

(2)

ΔE3 = E HO2 ‐ MPc − (E−O2 ‐ MPc + E H +)

(3)

The counterpoise correction has been employed to avoid the basis-set superposition errors (BSSEs) in the calculated interaction energies defined in each step described above. In addition, the Morokuma−Ziegler partitioning scheme was employed55 as implemented in the ADF code56 by using the optimized geometries obtained from the Turbomole calculations. According to this scheme, the interaction energy is partitioned as follows: ΔEint = ΔEpauli + ΔVelstat + ΔEorb, where the ΔVelstat term accounts for the stabilizing electrostatic interaction, and ΔEorb stands for the stabilizing covalent character of the fragment interaction. Triple-zeta STOs plus polarization function were employed as a basis set, where the scalar relativistic effects were taken into account through a twocomponent zero-order regular approximation (ZORA) Hamiltonian in conjunction with the PBE functional (TZP-ZORA/ PBE) in the ADF calculation.

3. RESULTS AND DISCUSSION Before examining the electrocatalytic properties of the different molecular configurations, the Au(111) surface was characterized by using cyclic voltammetry and STM experiments, to establish reproducible standard starting conditions to obtain similar bare Au electrodes before further modification. Figure 1 illustrates the different hypothetical idealized configurations of a

Figure 2. Structure of the complexes studied theoretically, MPc−L− Au26 (M = Fe, Cu; L = 4-ATP, MDPP). 15331

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

SAMs.58 The electrical charge associated with the monolayer of 4-ATP (66 μC cm−2) is larger than the charge related to MDPP (44 μC cm−2). This difference in coverage and the rather broad desorption peak can be attributed to the larger size of the MDPP molecules that should be assembled less close-packed on the Au(111) surface, giving rise to a more open monolayer and a lower coverage than 4-ATP and poorly defined film structure. This is in agreement with the STM images and surface profile analysis. The surface morphology of 4-ATP and MDPP SAMs was analyzed by ex situ STM (see Figure 4b and e). The STM images show the typical morphology of thiolate SAMs covering the Au(111) surface. The STM images essentially reveal that the SAMs morphologies have significant differences for the two thiol molecules used, and they are in agreement with the voltammetry measurements (Figure 4a and d) discussed above. Because the images of the bulkier MDPP molecules would appear more separated from each other than those of 4-ATP, in pyridinium SAMs the surface coating is less continuous and has low homogeneity, as shown in the inset of Figure 4e. The cross-section height profiles of SAMs (through the yellow line on STM images) shown in Figure 4c and f provide valuable information to confirm that the self-assembled MDPP molecules are less close-packed on the Au(111) surface than the corresponding 4-ATP SAMs. The MDPP SAMs exhibit typical pits (dark regions on figure 4e) with a depth of about 2.5 Å, equal to a monolayer step. Even considering that the STM images have rather low resolution, a quasi-molecular ordered pattern of 4-ATP adsorbed molecules can be observed, as illustrated in the inset of Figure 4b. This is in agreement with the single desorption wave for 4-ATP. The cross-section height profiles of 4-ATP SAMs (Figure 4c) show that the distance between adjacent lines stacking of 4-ATP molecules is ca. 5.1− 5.2 Å along the yellow line on the STM image, which is consistent with a (√3 × √3)R30° packing structure on Au(111).59a,b Figure 5a shows the STM image of the Au(111) surface modified with a Fe−phthalocyanine anchored via a 4-ATP SAMs. On the image, bright spots stand out showing some regularity in size and homogeneously distributed structures on the SAMs surface. These local bright regions, with diameters between 2 and 3 nm and about 0.5 nm height, are associated with the presence of the FePc molecules anchored via a 4-ATP layer to the surface. The sectional analysis of this surface, shown in Figure 5b and c, passing just over some of these bright regions (labeled 1−5 on red lines of Figure 5a), reveals subnanometric height structures with a regular and common pattern, very similar to those reported on previous STM-UHV reports on FePc layers adsorbed directly on Au(111).38b,58,59c,d Despite the fact that our STM measurements were collected in air and considering the important effect of air exposure on the FePc layers, our STM results together with the electrochemical performance for O2 reduction of this modified surface (discussed below) clearly show that the FePc molecules are present and effectively anchored on the 4-ATP SAMs as illustrated in the inset of Figure 5a. Figure 6 compares the potentiodynamic response of the Au(111) electrode in the presence of oxygen before and after modification using different molecular configurations containing SAMs with and without FePc. The intensity of the currents for all systems varied linearly with the potential scan rate, which corroborates that the ORR is under mass-transport control, that is, controlled by the diffusion of O2 from the bulk of the electrolyte to the electrode surface. The dashed line shows the

Au(111) gold surface modified by FePc directly adsorbed on the surface (Figure 1a), modified with a FePc phthalocyanine anchored via a 4-ATP (Figure 1b) and by a MDPP (Figure 1c). It is important to point out that these are only schematic representations of SAMs arrangements on the Au(111) surface, as these might present some degree of distortion from an ideal perpendicular orientation as the Au−S bond angle with respect to the plane of the gold surface can deviate from 90°33,38,39,57 as illustrated in Figure 2. This will be discussed later when analyzing the theoretical calculations. Figure 3a illustrates a cyclic voltammogram of the Au(111) surface obtained after annealing the Au specimen and measured

Figure 3. (a) Cyclic voltammogram of evaporated gold after heattreatment in aqueous 0.1 M H2SO4, dE/dt = 50 mV/s. (b) STM image of the clean Au(111) surface; it 800 pA, Ebias 200 mV.

in a O2-free aqueous acid solution. The i versus E profile is typical for an Au(111) single crystal electrode with atomically smooth (111) surfaces. Figure 3b shows a typical STM image with the typical herringbone morphology of Au(111) flat terraces. Figure 4a and d illustrates the voltamperommetric response for the reductive desorption process of SAMs of 4ATP and MDPP from the Au(111). The stripping of the SAMs from the Au surface was conducted in 0.1 M NaOH at 0.05 V/ s, respectively. The single desorption wave observed for 4-ATP at ca. −0.8 V suggests that the SAMs for this particular molecule are rather homogeneous. In contrast, the linear voltammogram of the reductive desorption of MDPP (Figure 4d) shows a rather broad and negatively shifted wave, that seems to be the result of the combination of three peaks, indicating the desorption of three forms of MDPP adsorbed on Au. The molecular structure of MDPP suggests that significant lateral chain−chain interactions (π-stacking) could be taking place once the SAMs is formed, which ultimately strongly influences the stability and the desorption process of the 15332

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

Figure 4. From top to bottom: Voltamperommetric result for the electrodesorption of Au(111)/SAMs in 0.1 M NaOH at 0.05 V/s; STM height image of the Au(111)/SAMs surfaces, it 200 pA, Ebias 1.5 V; and cross-section analysis of Au(111)/SAMs surfaces, of 4-ATF SAMs (a−c), and MDPP SAMs (d−f).

results for the Au surface coated directly with FePc is interesting: despite the fact that the bare Au(111) has catalytic activity for the reduction of O2, after adding FePc to its surface, an enhancement of the currents is observed. FePc probably covers more Au active sites as compared to the new Fe active sites existing in the FePc molecules, but this is surpassed by the superior catalytic activity of FePc. If one assumes that the FePc adsorbed flat on the Au(111) surface, forming a monolayer, each FePc molecule covers between 18 and 24 Au active sites.59 This estimation is based on STM high-resolution images obtained in high vacuum on monolayers of FePc on Au(111) that show that FePc can adopt two possible flat configurations covering between 18 and 24 Au atoms.59 The FePc monolayer leaves at least 15% of the Au surfaces uncoated, so it can be argued that when the Au(111) surface is modified with FePc some Au sites might still be active for O2 reduction. On the other hand, when 20 sites are shadowed by a FePc molecule, O2 has no access to these sites, but it now has access to a new active site, the Fe center present in the phthalocyanine located at the outermost site of the SAM. So roughly speaking, the gold surface loses 85% of its active sites by the presence of FePc, but the net result is a much higher activity. Fe sites are then very effective for promoting the O2 reduction despite being diluted by a ratio of at least ca. 20:1 as compared to the number of Au

response of the bare Au(111). ORR on gold is known to proceed only via 2-electrons to give peroxide,60,61 so the voltammetric waves correspond to a total transfer of 2electrons. After modification of the Au(111) with a SAMs of MDPP, the ORR currents are severely suppressed, indicating that most of the surface of Au(111) is coated with these SAMs, probably leaving very few open spots or pinholes for the direct reduction of O2 on the Au sites (see purple line). When the Au(111) is modified with a SAM of 4-ATP, the O2 reduction currents are suppressed even to much lower values as compared to those obtained when MDPP (see green line), showing that in this case the SAM layers are probably more compact, leaving practically no free Au sites when O2 reduction can take place. These results are in agreement with the data shown in Figure 4, which shows that SAMs containing 4-ATP molecules are denser than those formed by MDPP. The results obtained with the two different SAMs strongly indicate that they cover the Au surface to such an extent so to block the active sites for O2 reduction. Hence, as expected, the SAMs of both thiols acts as inhibitors of the reduction of molecular oxygen, and the adsorbed SAMs (which have no phthalocyanine molecules linked to them) show no catalytic activity for the reaction. The blue line shows the response of the Au(111) surface when it is modified by direct adsorption of FePc on its surface. The 15333

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

Au/FePc proceeds via 4 electrons, the current peak should be almost double as compared to that obtained with bare Au(111). However, with the Au/FePc electrode, there is only an increment of ca. 60%, so probably the ORR proceeds via two parallel mechanisms: via 2-e to give peroxide and via 4-e to give water, so the peak currents on the Au/Fe will involve a total transfer of an average of 3.3 electrons. This result suggests that a parallel process involving 2-electrons probably does take place on the uncovered Au sites. Similar results are obtained when FePc is anchored to the Au surface via 4-ATP (Au/4-ATP/ FePc), because the intensities of the currents are similar (slightly lower) to those obtained on Au/Fe, but they are still higher than those corresponding to bare Au(111). Further, a shift of almost 0.1 V is observed to more favorable potentials (less negative), so the presence of the anchoring thiol decreases the overpotential of the ORR by near 0.1 V. This effect is surprising because the presence of a spacer should decrease the overall ET rate, despite the fact that the aminothiophenol spacer has a π-conjugated system. As shown by several authors,64 ET transfer rates via conjugated systems decrease as the length of the chain increases, and this is more pronounced when nonconjugated aliphatic spacers are used.64 A more dramatic effect is observed when a bulkier spacer (Au/ MDPP/FePc) is used to anchor the FePc to the Au(111) surface. A shift of the ORR currents of more than 0.1 V to more positive potentials is observed as compared to Au/4-ATP/FePc and of more than 0.2 V as compared to bare Au. This shift in the reduction waves when using a spacer for FePc can be interpreted as a net catalytic effect for ORR induced by the axial ligand. The peak current of ORR with the FePc/MDPP/Au is also the largest of all of the systems examined and is 2.1 times larger than that obtained with bare Au, so this is a strong indication that for this system the ORR process occurs almost entirely via 4 electrons to give water. The uncovered Au sites probably do not contribute much to the measured currents at those potentials. Rotating ring-disk measurements could be very useful to corroborate this even further to quantify the amount of peroxide formed. However, the SAMs configurations might not be stable upon rotating a Au electrode. Work is in progress to check this point. It can be concluded from the data shown in Figure 6 that the presence of FePc in any configuration increases the amount of electrons per oxygen molecule transferred, and particularly for Au/MDPP/FePc, the process seems to proceed almost entirely via 4-electrons. Furthermore, the SAMs functionalized with MDPP and FePc present the highest activity for O2 reduction as compared to all of the other configurations, both in terms of number of electrons per O2 molecule transferred and by the decrease in the overpotential of the reaction by more than 0.2 V as compared to bare Au. However, when a spacer is used, the rate of ET should not increase, as the longer is the spacer, the lower is the ET rate (assuming the rate-determining step is the transfer of one electron to the dioxygen molecule, as described in the reaction mechanism later). Nevertheless, this effect seems to be overcome by another effect. One can argue that the increase in catalytic activity observed when SAMs are used can be attributed to a larger delocalized π-system that probably decreases the reorganizational energy of the whole system, decreasing the activation energy. Moreover, axial ligation to the Fe center by the spacer via a nitrogen atom could also favor the interaction of O2 with the Fe center. As discussed by Kasai et al.18 for the ligation of O2 to Fe in heme, axial ligation can affect the orientation of the O2 molecule on the Fe center. O2

Figure 5. (a) 100 × 100 nm2 height STM image of a surface of Au(111)/SAMs 4-ATP/FePc. (b and c) Section analysis of a section surface of Au(111)/SAMs 4-ATP/FePc surface (it 200 pA, Ebias −1.2 V).

Figure 6. Electroreduction of O2 on a clean Au(111) surface and modified with FePc, with and without 4-ATP and MDPP SAMs. Measurements conducted in O2 saturated 0.1 M NaOH, dE/dt = 0.05 V/s.

atoms covered, all this of course assuming that a monolayer does exists. It has been discussed by Scherson62 for O2 reduction on Fe porphyrins confined on graphite, from earlier data published by Shigehara and Anson,63 that very few Fe active sites are sufficient to sustain catalytic currents. It is important to point out that, as shown in many publications, FePc in alkaline media catalyzes the ORR preferentially via 4 electrons,2−4,6,7,9 so the enhancement of the catalytic currents in the presence of FePc can be attributed in part to a change in the overall mechanism of the reaction. If the reaction on the 15334

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

ces2,4,6,7,10,11,65 that seems to operate for FePc anchored to Au(111) is the following:

reduction can have two possible orientations when interacting with the Fe center: end-on (tilted) and side-on (with O2 parallel to the plane of the phthalocyanine molecule). On the one hand, the end-on orientation does not favor the cleavage of the O−O bond that is required for a 4-electrons reduction process, so this configuration should lead to a 2-electron reduction mechanism to give peroxide. On the other hand, the side-on interaction involves a transition state that has larger energy (higher activation energy), even though the side-on orientation does promote the splitting of the O−O bond involving a transition state of lower energy, which implies a lower activation energy for the process. In case that the latter configuration predominates, the overall process for O 2 reduction should proceed via the transfer of a total of 4electrons. So in our case, the axial ligation of the bulkier spacer seems to favor the side-on interaction, facilitating the cleavage of the O−O bond and the overall ORR via 4-electrons and probably lowering the activation energy as pointed out by Kasai et al.18 for heme. It is also possible that O−O binds to a Fe center and to a nitrogen atom in the phthalocyanine, as suggested by Sidik and Anderson.40 This interaction can prevent the peroxide formed from desorbing from the Fe site after the transfer of two electrons. Finally, Tafel slopes obtained from slow scan polarization curves give linear correlations of slope −0.040 V/decade (see Figure 7) for O2 reduction on FePc, for all different

[Fe(III)PcOH]ads + e− ⇆ [Fe(II)Pc]ads + OH−

fast

(4)

[Fe(II)Pc]ads + O2 ⇆ [Fe III‐‐O2−]ads

fast

(5)

[Fe III‐‐O2−]ads + e− → intermediates

rds

(6)

However, according to the calculated energy diagram of Figure 8, the formation of the adduct in step 5 will slow the reaction instead of catalyzing it because the energy barrier is larger as compared to that where the initial reactants proceed directly to the formation of the adduct, concerted with the transfer of oneelectron. So, the following mechanism seems to be more plausible and still agrees with the kinetic data: [Fe(II)Pc]ads + O2 + e− → [Fe III‐‐O2−]ads

rds

(7)

Under this new scheme, the rate-determining step 7 proceeds immediately after step 4, so again, the fast electron-transfer step is followed by a slow step. In this case, O2 adsorption on the Fe sites takes place in a concerted fashion with an electron-transfer process. If step 7 is the rate controlling, with a symmetrical energy barrier, the Tafel slope should be 0.039 V. Many authors have reported Tafel slopes in the range 0.030−0.040 V for O2 reduction in alkaline media catalyzed by Fe phthalocyanines or similar Fe macrocyclic complexes using graphite or carbon substrates.2,4,6,7,10,11,67 Figure 7 shows Tafel lines for FePc adsorbed directly on Au(111) and via molecular spacers, and in all cases the slopes are close to 0.030 V, so essentially the kinetics parameters as far as Tafel slopes are concerned do not seem to depend on the way of surface confinement of the FePc catalyst on the electrode. They give practically the same values no matter whether the substrate is graphite, carbon, gold, or gold coated with SAMs. Essentially one can state that FePc behaves as if it were in the homogeneous phase, and the electrode material behaves only as a sink or source of electrons. The enhancement in the catalytic activity when FePc is further away from the electrode surface could be possibly attributed also to a three-dimensional configuration of the catalytic centers that is usually observed for redox catalysis as discussed by Savéant.58a,66 In redox catalysis, the catalyst is present in the homogeneous phase, and the electrode surface only acts as a source or sink of electrons to regenerate the active form of the catalyst upon reduction. In the present case, this could correspond to step 4 in the scheme described above. In fact, as discussed by Savéant,58a “electrochemists are often trying to design porous electrode-so porous that they are sometimes called volumic electrodes.” However, redox catalysis is essentially an outer-sphere process where the reacting molecule and the active site only collide,66 without the formation of an adduct as in step 5 or 7. It is unlikely that O2 simply collides with the Fe centers forming superoxide O2−. The outer-sphere reduction of O2 to give O2− is only observed in very strong alkaline solutions.10 Also, an outer-sphere process is unlikely if a 4-electron reduction mechanism is observed, because O2 needs to bind to an active center, most probably by a side interaction in the present case, to promote the splitting of the O−O bond. So finally, the source of the enhancement of the catalytic activity of FePc when using conjugated bridging molecules to Au(111) is not clear at the moment but is probably related to an effect of the axial ligand that binds the FePc to the gold surface. If the axial ligand withdraws electron density form the

Figure 7. Tafel Plot for for O2 reduction in 0.1 M NaOH on clean Au(111) surface, FePc adsorbed on Au(111), FePc-linked-SAMs of 4ATP on Au(111), and FePc-linked-SAMs of MDPP on Au(111).

configurations suggesting that the rate-determining step involves the transfer of one electron after a fast one-electron transfer step. These slopes are typical values also observed for O2 reduction on graphite electrodes modified with Fe phthalocyanines1,4,6,7,65 so the mechanism of the reaction seems to be dictated solely by the FePc and not by the form of modification of the surface, no matter if the substrate is graphite, gold, or gold coated with SAMs of different thicknesses. Finally, it is important to note that with our results we are not suggesting that all of the self-assembled monolayers are capped with FePc molecules. We have shown that uncapped SAMs (without FePc molecules) do not show catalytic activity, and they should not contribute to the catalytic currents observed. A generally accepted mechanism for O2 reduction by Fe phthalocyanines confined to graphite or carbon surfa15335

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

half of those observed for FePc (see Figure 6), indicating that 2-electrons instead of 4-electrons are involved in the O2 reduction process. The currents are for all cases above 400 μA (for bare Au and for Au/CuPc and Au/MDPP/CuPc). As it has been pointed out above, bare gold is a well-known 2electron catalysts for ORR, so by comparison it can be stated that all configurations of CuPc catalyze the reduction of O2 via 2-electrons. Bare Au (dashed line in Figure 9) shows higher activity than when modified with CuPc. Only the Au/4-ATP/ CuPc electrode shows larger currents, but the peak is narrower. The differences in reactivity of all different configurations are rather small, but, taking as a criterion the peak potential, it can be said that the activity decreases as follows: bare Au > Au/ CuPc Au/4-ATP/CuPc ≥ Au/MDPP/CuPc. Despite the small differences, it can be said that both Au/4-ATP/CuPc and Au/ MDPP/CuPc show similar reactivities but lower than that of Au/CuPc (currents shifted to more negative values by 50 mV). Thus, when CuPc is anchored to gold using SAMs, the activity decreases as compared to the case when CuPc is directly attached to Au. It is important to point out that the ORR reduction waves of Au/CuPc, Au/4-ATP/CuPc, and Au/ MDPP/CuPc are probably contaminated with currents coming from ORR occurring on naked regions of the gold surface not covered by the different molecular arrangements. In the case of FePc, this “contamination” of the currents by reduction of O2 on possible unnocupied sites of Au is less important because FePc is much more active than bare Au and the currents are observed at much lower overpotentials. So, if one assumes that the fractions of uncovered Au in Au/CuPc, Au/4-ATP/CuPc, and Au/MDPP/CuPc are similar, then the trend in reactivities observed is real. A more pessimistic view could consider that when CuPc is attached directly to gold, it leaves more open spaces than when coated with SAMs, because the thiol molecules are probably not all capped with CuPc. In this case, the higher activity of Au/CuPc could be attributed to a higher contribution of the bare gold to the overall reduction currents observed. However, despite all of these possible considerations, it can be concluded that the effect of the SAMs spacer on the catalytic activity is minimal, in contrast to the larger effects observed for FePc. It is important to emphasize that the differences in catalytic activity between FePc and CuPc have also been explained in terms of intrinsic redox activity. For example, FePc exhibits the M(III)/M(II) reversible processes but CuPc does not.2,10 Redox processes that occur on CuPc involve the ligand and not the metal center, as mentioned above. FePc possesses energy levels with large metal character between the energy of the HOMO and the energy of the LUMO of the phthalocyanine ligand, but CuPc does not have these intermediate energy levels with metal character. It is important to point out that to obtain catalytic activity, the frontier orbital of the metallophthalocyanine needs to have some d character. Indeed, the most active MPcs2,10 (having Cr, Mn, Fe, and Co as metal centers) have frontier orbitals with d character, whereas in Ni and Cu phthalocyanines, the frontier orbitals have more ligand character.58b This has been graphically illustrated with theoretical calculations58c through the comparison of the frontier orbitals of CoPc with the corresponding orbitals of CuPc where the former shows a welldefined dz2 orbital sticking out of the plane of the phthalocyanine, whereas CuPc does not and shows very low activity for O2 reduction. There is experimental evidence to support this using tunneling electron microscopy. By this technique, Hipps et al.58d have found a strong d-orbital

Fe center, it will favor the reaction. This is a known effect for phthalocyanines that have electron-withdrawing groups located on the phthalocyanine ligand.6,10 Cu phthalocyanine presents low activity for the reduction of O2.2,10 Of the series CrPc, MnPc, FePc, CoPc, NiPc, and CuPc, the latter is the least active, and this has been attributed to the fact that the frontier orbitals of this particular catalyst have very low metal character, in contrast to FePc.10 So the electronic interaction between O2 and the CuPc is expected to be very weak. Further, if CuPc molecules are anchored to gold using the molecular spacers that we have discussed for FePc, the electronic communication between the gold and CuPc should be weaker as compared to that for FePc; hence the behavior of the supported CuPc should be different as compared to the isolated CuPc. So we have examined the catalytic activity of three different configurations of CuPc, CuPc/Au, CuPc/4ATP/Au, and CuPc/MDPP/Au, as previously discussed for FePc. Figure 9 illustrates the electrochemical response of these

Figure 8. Energy profile of O2 reduction on MPc−L (M = Fe, Cu; L = 4-ATP, MDPP).

Figure 9. Electroreduction of O2 on a clean Au(111)surface and modified with CuPc, with and without SAMs of thiophenols. Measurements conducted in O2 saturated 0.1 M NaOH, dE/dt = 0.05 V/s.

three configurations for O2 reduction under the same conditions used for FePc. It can be clearly seen that two main effects are noticed: (i) the current maxima are about one15336

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

dependence on the images of metal phthalocyanines. Unlike CuPc where the metal appears as a hole, CoPc shows the highest point in the molecular image, while the benzene regions showed the same height in both systems. So, essentially real images are in agreement with those predicted by theoretical calculations.10,58c We carried out theoretical calculations for a better understanding of the observed experimental data described above. Optimized structures (Figure 2) and electronic structure properties of models of free metallophthalocyanine and linked to a gold cluster are listed in Tables 1−3. These tables

Table 3. Some Geometric Parameters of the Systems (Distances in Å and Angles in deg) and Energy Intermediate Specie HO2 (ΔE3, kcal/mol), When the Complex Is Formed and Reduced with and without the Gold Cluster: H2O− MPc−L−Au26, H2O−MPc−L, H2O−MPc−Au26, and H2O− MPc, Respectively (M = Fe, Cu; L = 4-ATP and MDPP)

Table 1. Some Geometric Parameters of the Systems (Distances in Å and Angles in deg) and Adsorption Energy (ΔEad,kcal/mol), When the Complex Is Formed with and without the Gold Cluster: O2−MPc−L−Au26, O2−MPc−L, O2−MPc−Au26, and O2−MPc, Respectively (M = Fe, Cu; L = 4-ATP and MDPP) systems O2 O2−FePc−4ATP− Au26 O2−CuPc−4ATP− Au26 O2−FePc−MDPP− Au26 O2−CuPc−MDPP− Au26 O2−FePc−4ATP O2−CuPc−4ATP O2−FePc−MDPP O2−CuPc−MDPP O2−FePc−Au26 O2−CuPc−Au26 O2−FePc O2−CuPc

S−Au

M− Au

Nax−M

O−M

O−O

ΔEad −22.5

2.60

2.25

1.88

1.21 1.27

2.67

3.63

2.96

1.23

−8.2

2.79

1.99

1.96

1.25

−13.1

2.77

2.51

3.03

1.22

−2.8

2.19 3.32 2.00 2.49

1.94 2.81 1.95 3.04 1.74 3.07 1.92 3.07

1.27 1.24 1.26 1.22 1.26 1.22 1.26 1.21

−23.0 −4.6 −13.0 −2.5 −32.8 −3.5 −43.2 −2.9

3.96 3.84



O2−FePc−4ATP− Au26 − O2−CuPc−4ATP− Au26 − O2−FePc−MDPP− Au26 − O2−CuPc−MDPP− Au26 − O2−FePc−4ATP − O2−CuPc−4ATPa − O2−FePc−MDPP − O2−CuPc−MDPP O2−FePc−Au26 O2−CuPc−Au26 O2−FePc O2−CuPc a

M− Au

Nax−M

O−M

O−O

ΔE2

2.62

2.26

1.89

1.27

71.1

2.65

3.67

2.92

1.24

76.7

2.78

1.99

1.96

1.27

24.9

2.77

2.45

3.05

1.23

29.6

2.24

1.96

1.28

109.0

2.01 2.44

1.96 3.12 1.88 3.00 1.88 3.08

1.28 1.23 1.26 1.23 1.28 1.22

38.8 49.2 23.4 27.1 42.1 61.3

S−Au

4.03 3.96

S−Au

M− Au

Nax−M

O−M

O−O

ΔE3

HO2−FePc−4ATP− Au26 HO2−CuPc−4ATP− Au26 HO2−FePc−MDPP− Au26 HO2−CuPc−MDPP− Au26 HO2−FePc−4ATP HO2−CuPc−4ATP HO2−FePc−MDPP HO2−CuPc−MDPP HO2−FePc−Au26 HO2−CuPc−Au26 HO2−FePc HO2−CuPc

2.64

2.64

2.17

1.81

1.44

−70.3

2.63

2.63

3.86

2.35

1.37

−57.5

2.77

2.77

2.03

1.83

1.43

−27.8

2.77

2.77

2.66

2.77

1.33

−7.2

2.15 3.26 2.03 2.65

1.85 2.20 1.82 2.72 1.76 2.28 1.76 2.32

1.45 1.40 1.43 1.34 1.46 1.35 1.46 1.34

-101.8 −80.6 −41.3 −26.7 −23.3 −8.6 −40.9 −41.1

4.01 3.95

summarize the steps of the mechanism proposed in the experimental work. Figure 8 shows a schematic energy profile that describes the proposed mechanism, including the slow step of the overall process. The structural results are consistent with the data reported in the literature.18,40−42 In general, our optimized structures agree very well with experimental observations and other theoretical publications.44−46 There are two main possible coordination modes for the O2 molecule toward the metal center in the phthalocyanine: end-on or sideon configuration that depends on the nature of metal center and the structure of the macrocyclic ligand. In this work, we have obtained for all modes an end-on binding in the first step (Figure 2) as in the following two reaction steps. Selected geometric parameters are summarized in Table 1 for the first step, which corresponds to the adsorption of O2 on the central metal atoms of the MPc molecules. The obtained S−Au distances are close to the typical bond length, depicting a stabilization of the S−Au interaction of ∼40 kcal/mol as it was reported in a previous study.39 The distance between the axial nitrogen and the metal center in the phthalocyanine (Nax−M distance) denotes a typical coordination bond length for Fe, but for Cu is dramatically longer and weaker, suggesting that no formal bond is involved. When we use 4-ATP as the anchor ligand, the results showed that O2 adsorption energy on FePc/ 4-ATP is about 23.0 and 22.5 kcal/mol when the complex is free and when it is attached to the gold surface, respectively. On the other hand, when the ligand is MDPP, O2 adsorption energy on FePc is about 13.0 and 13.1 kcal/mol when the complex is free and when it is attached to the gold surface, respectively. Such energy magnitudes are all weak. This is evident from the O−O bond distance, which is still short at this stage and is associated with a double bond. The O−M distance corresponds to a bond length for complex formation as previously described in literature.43−46 Moreover, the O2 adsorption energy on CuPc is even weaker, with values between 2.5 and 8 kcal/mol, showing a longer O−M bond distance for the 4-ATP and MDPP ligands. It is interesting to observe the effect of interaction of O2 without ligand spacer and

Table 2. Some Geometric Parameters of the Systems (Distances in Å and Angles in deg) and Reduction Energy (ΔE2, kcal/mol), When the Complex Is Formed and Reduced with and without the Gold Cluster: −O2−MPc−L− Au26, −O2−MPc−L, −O2−MPc−Au26, and −O2−MPc, Respectively (M = Fe, Cu; L = 4-ATP and MDPP) systems

systems

No minimum was obtained. 15337

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

gold cluster in the systems Au26−MPc−O2 and MPc−O2. For FePc, a higher interaction energy is obtained as compared to the same system with ligand spacers. Meanwhile, the energy values in systems with Cu showed a lower magnitude, similar to the original systems. The results of the study of the second step, which involves the reduction of the system, are described in Table 2. It is possible to appreciate that most of the geometric parameters do not change, with the exception of the O−O and O−M distances. The energy of the reduction reaction, which we have called ΔE2, shows that it is an endergonic process. The complexes without the gold surface have a higher barrier. For complexes with MDPP, the energetic barriers are about onehalf of that obtained with 4-ATP. For the reduced models MPc−Au26 and O2−MPc, there is a noticeable decrease of the energy barrier. The third step describes the formation of the HO 2 intermediate. This species is formed from the protonation reaction of the previous step involving the adduct. The geometric parameters and reaction energies are shown in Table 3, where it is possible to notice that the O−O bond distance is associated with a single bond. All of the reactions were exergonic. As indicated above, the three steps are described in Figure 8. The energy barrier is located between the first and second steps, and the barrier heights are summarized in Table 4, from which it can be seen that the

formed. The Pc ring increases its withdrawing capability when the complexes interact with Au26 via the spacer, and this agrees with previous work6,10,67 that has shown that electronwithdrawing groups like Cl located on the ligand increase the catalytic activity of Fe phthalocyanine, as compared to unsubstituted FePc. So, this effect of the spacer could explain the higher catalytic activity of FePc when linked to gold via an axial ligand. These quantities change according to the type of bridging ligand attached to both the MPc molecule and the Au surface. The nature of the O2−MPc (M = Fe, Cu) has been rationalized by using the Morokuma−Ziegler energy decomposition (MZ-EDA) scheme.55 Previously, we have shown that the inclusion of the gold surface with the MPcL moiety (L = 4aminothiophenol and 4-mercaptopyridine) decreases the MPc−L interaction, leading to a slightly ionic interaction, in addition to its role as electron source.39 In this work, we focus into the O2 interaction and its variation through the three catalytic steps considered in eqs 4−6. The 4-ATP/FePc−O2 interaction with a calculated strength of −23.0 kcal/mol exhibits a covalent character (70.2%68) as a consequence of the backbonding toward the partially filled 2π* molecular orbital of O2 (Figure 10). In contrast, the O2− FePc(MDPP) interaction is only −13.0 kcal/mol, mainly due to a decrease in the backbonding, thus decreasing the orbital stabilizing term to 60.3% of the overall stabilizing terms (i.e., electrostatic and covalent interaction). For the 4-ATP systems, the inclusion of the gold surface slightly decreases the strength of the O2 adsorption, whereas for the MDPP counterpart the catalytic site remains almost unchanged. This can be attributed to the length of the MDPP ligand, which decreases the influence from the gold surface at this stage (eq 1). The copperbased system exhibits a weaker O2 adsorption due to almost negligible backbonding (Table 5), which accounts for its low capability to promote the catalytic reduction of O2. We also focused our analysis on the results provided by the 4-ATP/FePc and its interaction between the adsorbed O2 and the catalyst, because MDPP and its gold surface models exhibit a similar behavior along the steps of the catalysis. The reduction of the system (eq 2) leads to an increase in the population of the 2π* molecular orbital of O2 (initially of 2 e) from about 2.49 e for 4-ATP/FePc-O2 to 2.71 e, for 4-ATP/FePc-O2 destabilizing the system (Figure 10), and thus promoting the formation of 4-ATP/FePc-HO2. In 4-ATP/FePc-O2, the orbital stabilizing contribution to the stabilizing energy increases from 70.2% to 75.8%, for the 4-ATP/FePc−O2 interaction. However, as a consequence of the destabilization of the system (Figure 8), the 4-ATP/FePc−O2 interaction strength decreases to −16.3 kcal/mol mainly due to an increase in the destabilizing Pauli repulsion term,69 within the MZ-EDA scheme. Next, the formation of 4-ATP/FePc−HO2 stabilizes the system leading to an adsorbate−catalyst interaction of about 20.3 kcal/mol, with character similar to that of the initial 4-ATP/FePc−O2 adduct (71.3% orbitalary character). Such a system can undergo similar steps to promote the O2 conversion to H2O2 or 2H2O. For the MDPP case, the studied reaction steps are quite similar, showing a minor energy of the rate-determining step (slow step) because of a lesser destabilization of the MDPP/ FePc−−O2 adduct, due to that ligands stabilize the extra charge. The inclusion of the gold surface greatly decreases the calculated energies for the slow step (Table 4), due to its influence over the interaction of the adsorbate−catalyst due to a modulation of the catalyst center, decreasing the destabiliza-

Table 4. Energy of the Slow Step (kcal/mol)

a

systems

ΔE

O2−FePc−4ATP−Au26 O2−CuPc−4ATP−Au26 O2−FePc−MDPP−Au26 O2−CuPc−MDPP−Au26 O2−FePc−4ATP O2−CuPc−4ATPa O2−FePc−MDPP O2−CuPc−MDPP O2−FePc−Au26 O2−CuPc−Au26 O2−FePc O2−CuPc

48.6 68.5 11.8 26.8 86.0 25.8 46.7 −9.4 23.6 −1.1 58.4

No minimum was obtained.

FePc complexes show lower barriers than the CuPc complexes. Moreover, when complexes are linked to gold via the MDPP axial ligand, the barriers are lower, thus enhancing their catalytic activity. For the reduced models MPc−Au26 and O2−MPc, Fe systems show a trend almost without an energy barrier. On the other hand, systems with CuPc produce higher barriers. These latest results demonstrate the importance of the ligand spacer and the cluster of gold. To obtain a deeper insight on the electronic charge rearrangement caused by the interaction of MPc moiety with the gold substrate, the Natural Population Analysis (NPA) analysis based on the PBE density was performed for all of the series here studied (Table 5). It is possible to observe a charge transfer from the negatively charged anchor ligands (MDPP and 4-ATP) toward the MPc and the gold cluster (Au26) when the Au26−L−MPc assembly is formed. The charge in the gold cluster is mainly due to a charge transfer from the sulfur and bridging spacers. The metal centers (Fe and Cu) show a small change in their charges when the bridging complexes are 15338

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

Table 5. Natural Population Atomic (NPA) Analysis of the Systems

a

systems

Au

M

Pc

L

S

Nax

O

O2−FePc−4ATP−Au26 − O2−FePc−4ATP−Au26 HO2−FePc−4ATP−Au26 O2−CuPc−4ATP−Au26 − O2−CuPc−4ATP−Au26 HO2−CuPc−4ATP−Au26 O2−FePc−MDDP−Au26 − O2−FePc−MDDP−Au26 HO2−FePc−MDDP−Au26 O2−CuPc−MDDP−Au26 − O2−CuPc−MDDP−Au26 HO2−CuPc−MDDP−Au26 O2−FePc−4ATP − O2−FePc−4ATP HO2−FePc−4ATP O2−CuPc−4ATP − O2−CuPc−4ATPa HO2−CuPc−4ATP O2−FePc−MDDP − O2−FePc−MDDP HO2−FePc−MDDP O2−CuPc−MDDP − O2−CuPc−MDDP HO2−CuPc−MDDP O2−FePc−Au26 − O2−FePc−Au26 HO2−FePc−Au26 O2−CuPc−Au26 − O2−CuPc−Au26 HO2−CuPc−Au26 O2−FePc − O2−FePc HO2−FePc O2−CuPc − O2−CuPc HO2−CuPc

−0.52 −1.07 −0.58 −0.63 −1.23 −0.54 0.01 −0.48 −0.07 −0.08 −0.64 −0.08

0.86 0.79 0.84 1.21 1.20 1.23 0.78 0.76 0.82 1.23 1.24 1.24 0.74 0.76 0.79 1.19

−0.07 −0.43 −0.02 −0.02 −0.25 −0.05 −0.08 −0.26 0.03 −0.03 −0.06 −0.06 −0.56 −0.98 −0.40 −0.54

−0.16 −0.19 −0.17 −0.25 −0.33 −0.18 0.25 0.01 0.23 0.11 −0.21 0.16 −0.15 −0.61 −0.29 −0.27

−0.25 −0.26 −0.28 −0.23 −0.27 −0.20 −0.32 −0.35 −0.31 −0.31 −0.32 −0.31 −0.18 −0.44 −0.25 −0.21

−0.79 −0.79 −0.78 −0.78 −0.80 −0.77 −0.41 −0.42 −0.42 −0.51 −0.54 −0.48 −0.76 −0.78 −0.77 −0.77

−0.24 −0.31 −0.68 −0.11 −0.18 −0.66 −0.18 −0.26 −0.65 0.00 −0.09 −0.47 −0.29 −0.40 −0.75 −0.19

1.24 0.78 0.74 0.82 1.24 1.24 1.24 0.89 0.97 0.96 1.21 1.22 1.27 1.05 0.82 0.98 1.21 1.19 1.27

−0.35 −0.10 −0.43 0.01 −0.03 −0.34 −0.06 0.21 0.16 0.31 0.11 0.06 0.18 0.26 −0.66 0.22 0.00 −0.99 0.08

−0.21 0.29 −0.24 0.19 0.05 −0.55 0.11

−0.19 −0.11 −0.26 −0.13 −0.14 −0.26 −0.13

−0.76 −0.41 −0.43 −0.43 −0.52 −0.55 −0.49

−0.84 −0.19 −0.33 −0.66 −0.01 −0.11 −0.50 −0.17 −0.23 −0.67 0.00 −0.08 −0.58 −0.26 −0.34 −0.68 0.00 −0.01 −0.53

−0.04 −0.93 −0.09 −0.11 −0.98 −0.05

No minimum was obtained.

of the processes because of its influence over the M−O2 interaction.

4. CONCLUSIONS The preparation of self-assembled monolayers (SAMs) of 4aminothiophenol (4-ATP) and 1-(4-mercaptophenyl)-2,6diphenyl-4-(4-pyridyl)pyridinium tetrafluoroborate (MDPP) functionalized with iron phthalocyanine (FePc) and copper phthalocyanine (CuPc) adsorbed on gold (111) electrodes has been achieved. The catalytic activity of these SAMs/MPc was examined for the reduction of O2 in 0.1 M NaOH aqueous solution and compared to that of bare gold and of gold coated directly with MPc molecules. The electrocatalytic studies carried out with FePc/4-ATP/Au and FePc/MDPP/Au electrodes show that the O2 reduction takes place preferentially by the transfer of 4-electron to give water in contrast to a 2electron transfer process observed on the bare gold. The modified electrode obtained by simple adsorption of FePc directly to the Au(111) surface still promotes the 4-electron reduction process but shows less activity than the electrodes involving SAMs and FePc molecules positioned in the outmost

Figure 10. Close-up of the back-donation toward the 2π* molecular orbital of O2, denoting the participation from the anchoring ligand.

tion of the reduced system, due to an electron-withdrawing effect over the FePc moiety favoring such interaction (Table 5). As can be seen, in the overall catalysis steps, the gold surface, in addition to its electron-source role, decreases the energy barrier 15339

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

357−372. (g) Wood, T. E.; Tan, Z.; Schmoeckel, A. K.; O’Neill, D.; Atanasoski, R. J. Power Sources 2008, 178, 510−516. (h) Jaouen, F.; Herranz, J.; Lefrevre, J. M.; Dodelet, J.-P.; Kramm, U. I.; Herrmann, I.; Bogdanoff, P.; Maruyama, J.; Nagaoka, T.; Garsuch, A.; Dahn, J. R.; Olson, T.; Pylypenko, S.; Atanassov, P.; Ustinov, E. A. ACS Appl. Mater. Interfaces 2009, 1, 1623−1639. (i) Zagal, J. H.; Ponce, I.; Baez, D.; Venegas, R.; Pavez, J.; Paez, M.; Gulppi, M. Electrochem. Solid-State Lett. 2012, 15, B90−B92. (9) Baker, R.; Wilkinson, D. P.; Zhang, J. Electrochim. Acta 2009, 54, 3098−3102. (10) Zagal, J. H.; Griveau, S.; Silva, J. F.; Nyokong, T.; Bedioui, F. Coord. Chem. Rev. 2010, 254, 2755−2791. (11) van den Brink, F.; Barendrecht, E.; Visscher, W. Recl. Trav. Chim. Pays-Bas 1980, 99, 253−262. (12) Tarasevich, K. A. R. a. M. R. Elektrokhimiya 1986, 22, 1155− 1170. (13) Yeager, E. Electrochim. Acta 1984, 29, 1527−1537. (14) Cook, A. H. J. Chem. Soc. (Resumed) 1938, 1761−1768. (15) Murtomäki, L.; Barker, M. H.; Manzanares, J. A.; Kontturi, K. J. Electroanal. Chem. 2003, 560, 95−103. (16) Melnyk, A. C.; Kildahl, N. K.; Rendina, A. R.; Busch, D. H. J. Am. Chem. Soc. 1979, 101, 3232−3240. (17) George, P. Biochem. J. 1948, 43, 287−295. (18) Tsuda, M.; Diño, W. A.; Nakanishi, H.; Kasai, H. Chem. Phys. Lett. 2005, 402, 71−74. (19) Araki, K.; Wagner, M. J.; Wrighton, M. S. Langmuir 1996, 12, 5393−5398. (20) Suto, K.; Yoshimoto, S.; Itaya, K. Langmuir 2006, 22, 10766− 10776. (21) Roberts, G. G. Langmuir-Blodgett Films; Plenum Press: New York, 1990, (22) Ulman, A. An Introduction to Ultrathin Organic Films: From Langmuir-Blodgett to Self-Assembly; Academic Press: New York, 1991. (23) Kampas, F. J.; Yamashita, K.; Fajer, J. Nature 1980, 284, 40−42. (24) Curran, D.; Grimshaw, J.; Perera, S. D. Chem. Soc. Rev. 1991, 20, 391−404. (25) Yu-Hong Tse, P. J.; Lam, H.; Zhang, J.; Pietro, W. J.; Lever, A. B. P. J. Porphyrins Phthalocyanines 1997, 1, 3−16. (26) Griveau, S.; Pavez, J.; Zagal, J. H.; Bedioui, F. J. Electroanal. Chem. 2001, 497, 75−83. (27) Pavez, J.; Páez, M.; Ringuedé, A.; Bedioui, F.; Zagal, J. H. J. Solid State Electrochem. 2005, 9, 21−29. (28) Okunola, A. O.; Nagaiah, T. C.; Chen, X.; Eckhard, K.; Schuhmann, W.; Bron, M. Electrochim. Acta 2009, 54, 4971−4978. (29) Manivannan, A.; Nagahara, L. A.; Hashimoto, K.; Fujishima, A.; Yanagi, H.; Kouzeki, T.; Ashida, M. Langmuir 1993, 9, 771−775. (30) Zhang, Z.; Hou, S.; Zhu, Z.; Liu, Z. Langmuir 1999, 16, 537− 540. (31) Somashekarappa, M. P.; Keshavayya, J.; Sampath, S. Pure Appl. Chem. 2002, 74, 1609−1620. (32) Eckermann, A. L.; Feld, D. J.; Shaw, J. A.; Meade, T. J. Coord. Chem. Rev. 2010, 254, 1769−1802. (33) Kalyuzhny, G.; Vaskevich, A.; Ashkenasy, G.; Shanzer, A.; Rubinstein, I. J. Phys. Chem. B 2000, 104, 8238−8244. (34) Mezour, M. A.; Cornut, R.; Hussien, E. M.; Morin, M.; Mauzeroll, J. Langmuir 2010, 26, 13000−13006. (35) Ozoemena, K. I.; Nyokong, T. Talanta 2005, 67, 162−168. (36) Ozoemena, K. I.; Nyokong, T. J. Electroanal. Chem. 2005, 579, 283−289. (37) Ozoemena, K. I.; Nyokong, T. Electrochim. Acta 2006, 51, 2669−2677. (38) (a) Nyokong, T.; Bedioui, F. J. Porphyrins Phthalocyanines 2006, 10, 1101−1115. (b) Ponce, I.; Silva, J. F.; Oñate, R.; Caroli Rezende, M.; Páez, Ma. A.; Pavez, J.; Zagal, J. H. Electrochem. Commun. 2011, 13, 1182−1186. (c) Ponce, I.; Silva, J. F.; Oñate, R.; Caroli Rezende, M.; Páez, Ma. A.; Pavez, J.; Zagal, J. H. Electrochem. Commun. 2012, 17, 100.

portion of the self-assembled monolayers. The activity of the electrodes increases as follow: Au < Au/FePc < Au/4-ATP/ FePc < Au/MDPP/FePc with the highest activity when FePc molecules are more separated from the Au surface. In contrast, the less active CuPc shows almost the same activity in all three configurations. The enhancement in the electrocatalytic activity in the presence of axial ligands that are also acting as spacers can be attributed to an electron-withdrawing effect of the axial ligand on the Fe center. Electron-withdrawing groups located on the phthalocyanine are known to increase the catalytic activity for ORR. Theoretical calculations indicate the importance of the backbonding into the formation of the adduct, showing the catalytic role of the supporting gold surface mediated by anchoring ligands. It also shows that the formation of an adduct between FePc and O2 before the electron transfer process will slow the reaction, so it is likely that adduct formation occurs in a concerted fashion at the rate-determining step and involving the transfer of one-electron. This agrees with the kinetic parameters (Tafel slopes close to 0.040 V/decade) obtained from the experiments.



ASSOCIATED CONTENT

* Supporting Information S

Additional experimental details. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (J.P.); [email protected] (F.M.); [email protected] (R.A.-P.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work has been funded and prompted by the Millenium Nucleus P07-006-F. We thank Fondecyt Projects 1100773, 11100450, 1090627, 1110758, 11100027, and 1100162, UNAB-DI-28-12/R, and UNAB-DI-17-11/R. I.P. and R.O. acknowledge Conicyt doctoral fellowships. J.F.S. is grateful for Bicentenario grant PDA-23.



REFERENCES

(1) Vasudevan, P.; Santosh; Mann, N.; Tyagi, S. Transition Met. Chem. 1990, 15, 81−90. (2) Zagal, J. H. Coord. Chem. Rev. 1992, 119, 89−136. (3) Lipkowski, J.; Ross, P. N. Electrocatalysis; Wiley-VCH: New York, 1998, (4) Zagal, J. H. Macrocycles. Handbook of Fuel Cells; John Wiley & Sons, Ltd.: New York, 2010. (5) Li, Z.; Liu, B. J. Appl. Electrochem. 2010, 40, 475−483. (6) Zagal, J. H.; Silva, J. F.; Páez, M. In N4-Macrocyclic Metal Complexes; Zagal, J. H., Bedioui, F., Dodelet, J. P., Eds.; Springer: New York, 2006; pp 41−82. (7) Alkire, R. C.; Kolb, D. M.; Lipkowski, J.; Ross, P. Electrochemical Surface Modification: Thin Films, Functionalization and Characterization; Wiley-VCH: New York, 2008. (8) (a) Dodelet, J. P. In N4-Macrocyclic Metal Complexes; Zagal, J. H., Bedioui, F., Dodelet, J. P., Eds.; Springer: New York, 2006; pp 83− 147. (b) Bezerra, C. W. B.; Zhang, L.; Lee, K.; Liu, H.; Marques, A. L. B.; Marques, E. P.; Wang, H.; Zhang, J. Electrochim. Acta 2008, 53, 4937−4951. (c) Bashuam, R.; Zelenay, P. Nature 2006, 443, 63−66. (d) Wu, G.; More, K. L.; Johnston, C. M.; Zelenay, P. Science 2011, 22, 443−447. (e) Chlistunoff, J. J. Phys. Chem. C 2011, 115, 6496−6507. (f) Othman, R.; Dicks, A. L.; Zhu, Z. Int. J. Hydrogen Energy 2012, 37, 15340

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341

The Journal of Physical Chemistry C

Article

Organometalics 1986, 5, 1428−1434. (c) Lexa, D.; Savéant, J. M.; Su, K.-B.; Wang, D. L. J. Am. Chem. Soc. 1987, 109, 6464−6470. (67) (a) Zagal, J. H.; Cárdenas-Jirón, G. I. J. Electroanal. Chem. 2000, 489, 96−100. (b) Zagal, J. H.; Gulppi, M.; Isaacs, M.; Cárdenas, G.; Aguirre, M. J. Electrochim. Acta 1998, 44, 1349−1357. (68) Within the Morokuma−Zigler scheme, the ΔEorb represent 70.2% of the total stabilizing energy (calculated as ΔEorb/(ΔVelstat + ΔEorb)%). (69) For a recent example of the role of the Pauli repulsion term into the bonding, see: Jakobsche, C. E.; Choudhary, A.; Miller, S. J.; Raines, R. T. J. Am. Chem. Soc. 2010, 132, 6651−6653.

(39) Ponce, I.; Silva, J. F.; Oñate, R.; Miranda-Rojas, S.; MuñozCastro, A.; Arratia-Pérez, R.; Mendizabal, F.; Zagal, J. H. J. Phys. Chem. C 2011, 115, 23512−23518. (40) Anderson, A. B.; Sidik, R. A. J. Phys. Chem. B 2004, 108, 5031− 5035. (41) Jensen, K. P.; Roos, B. O.; Ryde, U. J. Inorg. Biochem. 2005, 99, 45−54. (42) Silaghi-Dumitrescu, R.; Silaghi-Dumitrescu, I. J. Inorg. Biochem. 2006, 100, 161−166. (43) Shi, Z.; Zhang, J. J. Phys. Chem. C 2007, 111, 7084−7090. (44) Wang, G.; Ramesh, N.; Hsu, A.; Chu, D.; Chen, R. Mol. Simul. 2008, 34, 1051−1056. (45) Chen, R.; Li, H.; Chu, D.; Wang, G. J. Phys. Chem. C 2009, 113, 20689−20697. (46) Sun, S.; Jiang, N.; Xia, D. J. Phys. Chem. C 2011, 115, 9511− 9517. (47) Reichardt, C.; Che, D.; Heckenkemper, G.; Schäfer, G. Eur. J. Org. Chem. 2001, 2001, 2343−2361. (48) Ricca, A.; Bauschlicher, C. W., Jr. Chem. Phys. Lett. 2003, 372, 873−877. (49) Bauschlicher, C. W., Jr.; Ricca, A. Chem. Phys. Lett. 2003, 367, 90−94. (50) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865−3868. (51) Ahlrichs, R.; Bär, M.; Häser, M.; Horn, H.; Kölmel, C. Chem. Phys. Lett. 1989, 162, 165−169. (52) Andrae, D.; Häußermann, U.; Dolg, M.; Stoll, H.; Preuß, H. Theor. Chem. Acc. 1990, 77, 123−141. (53) Bergner, A.; Dolg, M.; Küchle, W.; Stoll, H.; Preuss, H. Mol. Phys. 1993, 80, 1431. (54) Huzinaga, S. J. Phys. Chem. 1965, 42, 1293−1302. (55) (a) Morokuma, K. J. Phys. Chem. 1971, 55, 1236−1244. (b) Ziegler, T.; Rauk, A. Theor. Chim. Acta 1977, 46, 1−10. (56) Amsterdam Density Functional (ADF) Code; Vrije Universiteit: Amsterdam, The Netherlands, 2010. (57) Cao, R., Jr.; Díaz-García, A. M.; Cao, R. Coord. Chem. Rev. 2009, 253, 1262−1275. (58) (a) Savéant, J. M. Elements of Molecular and Biomolecular Electrochemistry: An Electrochemical Approach to Electron Transfer Chemistry; Wiley-Interscience: New York, 2006. (b) Rosa, A.; Baerends, E. J. Inorg. Chem. 1994, 33, 584−595. (c) Cardenas-Jiron, G. I.; Gulppi, M. A.; Caro, C. A.; Del Rio, R.; Paez, M.; Zagal, J. H. Electrochim. Acta 2001, 46, 3227. (d) Hipps, K. W.; Lung, X.; Wang, X. D.; Manzur, U. J. Phys. Chem. B 1996, 100, 11207−11210. (59) (a) Kim, Y.-T.; McCarley, R. L.; Bard, A. J. J. Phys. Chem. 1992, 96, 7416−7421. (b) Tao, Y.-T.; Wu, Ch.-Ch.; Eu, J.-Y.; Lin, W.-L. Langmuir 1997, 13, 4018−4023. (c) Cheng, Z. H.; Gao, L.; Deng, Z. T.; Jiang, N.; Liu, Q.; Shi, D. X.; Du, S. X.; Guo, H. M.; Gao, H. J. J. Phys. Chem. C 2007, 111, 9240−9244. (d) Cheng, Z. H.; Gao, L.; Deng, Z. T.; Liu, Q.; Jiang, N.; Shi, X.; He, X. B.; Du, S. X.; Gao, H.-J. J. Phys. Chem. C 2007, 111, 2656−2660. (60) Zurilla, R. W.; Sen, R. K.; Yeager, E. J. Electrochem. Soc. 1978, 125, 1103−1109. (61) Damjanovic, A.; Genshaw, M. A.; Bockris, J. O. M. J. Electroanal. Chem. Interfacial Electrochem. 1967, 15, 173−180. (62) Zhu, H.; Tolmachev, Y. V.; Scherson, D. A. J. Phys. Chem. C 2010, 114, 13650−13656. (63) Shigehara, K.; Anson, F. C. J. Phys. Chem. 1982, 86, 2776−2783. (64) (a) Weber, K.; Hockett, L.; Creager, S. J. Phys. Chem. B 1997, 101, 8286−8291. (b) Creager, S.; Yu, C. J.; Bamdad, C.; O’Connor, S.; MacLean, T.; Lam, E.; Chong, Y.; Olsen, G. T.; Luo, J.; Gozin, M. J. Am. Chem. Soc. 1999, 121, 1059−1064. (c) Sachs, S. B.; Dudek, S. P.; Hsung, R. P.; Sita, L. R.; Smalley, J. F.; Newton, M. D.; Feldberg, S. W.; Chidsey, C. E. D. J. Am. Chem. Soc. 1997, 119, 10563−10564. (65) (a) Zagal, J.; Bindra, P.; Yeager, E. J. Electrochem. Soc. 1980, 127, 1506−1517. (b) Zagal, J. H.; Páez, M.; Tanaka, A. A.; dos Santos, J. R.; Linkous, C. J. Electroanal. Chem. 1992, 339, 13−30. (66) (a) Lexa, D.; Mispelter, J.; Saveant, J. M. J. Am. Chem. Soc. 1981, 103, 6806−6812. (b) Lexa, D.; Savéant, J. M.; Wang, D. L. 15341

dx.doi.org/10.1021/jp301093q | J. Phys. Chem. C 2012, 116, 15329−15341