Impacts of Surface Functionalization on the Electrocatalytic Activity of Noble Metals and Nanoparticles

Low temperature fuel cells are a promising clean energy source, which can directly convert the chemical energy stored in hydrogen and small organic molecules (e.g., methanol, ethanol, and formic acid) into electrical energy with high efficiency, and thus find extensive applications in, for instance, automobiles and portable electronic devices. However, the commercialization of fuel cells is seriously impeded by the poor performance of the electrocatalysts for oxygen reduction reactions (ORR), and oxidation of small organic molecules.1, 2 The sluggish kinetics of these reactions, even catalyzed by high loading of Ptbased catalysts, results in significant losses in thermal efficiency. Therefore, great efforts have been dedicated to improving the electrocatalytic performance of noble metals in the past decades.

Generally, there are several possible interactions between reactants (and spectator molecules/ions) and modifying molecules, as depicted in Figure 1: (1) Steric blocking effect. The modifying molecules occupy part of the metal surface, which can greatly hinder the adsorption of other molecules with a relatively large size that need multiple surface sites. (2) Electrostatic effects or dipole-dipole interactions. For example, the modification of anions (cations) may enhance the co-adsorption of cations (anions). (3) Electronic structure effects. The interactions between modifying molecules and metal atoms may result in adsorbateinduced distortions of the surface lattice and partial electron transfer, which may alter the adsorption energy of a second adsorbate. In this chapter, we highlight some recent key progress in surface functionalization on massive Pt single crystal planes with calixarene and CNions, and their effects on the electrocatalytic activity and selectivity in ORR, hydrogen oxidation reactions (HOR), and methanol oxidation. We then extend the discussion to the synthesis of Pt and Pd nanoparticles capped with aryl and phosphine derivatives, and discuss the impacts of the molecule-molecule and molecule-nanoparticle interactions on the catalytic activities.

Surface modification of massive single crystal planes
Metal single crystal planes have well-defined surface atomic arrangements. Fundamentally, the use of single crystal planes as model electrocatalysts is anticipated to lead to a better understanding of the reaction mechanisms, surface processes, etc. 24,25

Modification of Pt single crystal planes with calix[4]arene
The ideal electrocatalysts for fuel cells should have both good activity and stability. It has been observed that during shutdown and startup that are inevitable for, for instance, automobile applications, the performance of cathodic Pt/C catalysts degrades quickly. 26 Under these conditions, the undesired ORR will occur on the anode side that is originally designed only for hydrogen oxidation. This reverse polarity phenomenon will push the cathode potential up to 1.5 V in some regions, and result in quick corrosion of Pt www.intechopen.com nanoparticles and carbon support. Clearly, it is very important to design selective anode catalysts that can efficiently suppress ORR while fully preserving the ability for HOR.
The Markovic group shows that chemically modified platinum with a self assembled monolayer of calix [4]arene molecules may meet this challenging requirement. 27 The structure of a calix [4]arene thiol derivative that they used is illustrated in Figure 2a. The wide rim containing sulfur atoms can firmly adsorb onto the Pt(111) terrace, leaving some edge or step sites (i.e., natural defects) and narrow terraces between the calix [4] [4]arene molecules totally suppress the oxygen adsorption/desorption (+0.5 ~ +0.9 V), and most of hydrogen adsorption/desorption (+0.05 ~ +0.4 V), as their coverage increases from 0.84 to 0.98 (from curve A to D). This suggests that the ions/water from the supporting electrolyte are unable to penetrate the narrow rim of the calix [4]arene molecules, and the only sites available for the adsorption of electrolyte components are the relatively small number of unmodified Pt step-edges and/or the small ensembles of terrace sites between the anchoring groups.
The calix [4]arene-modified Pt(111) exhibits significantly different selectivity for ORR and HOR, as shown in panels (c) and (d). That is, ORR is greatly suppressed as the coverage of calix [4]arene increases, while HOR is essentially the same. It is accepted that the critical number of bare Pt atoms required for the adsorption of O 2 is higher than that required for the adsorption of H 2 molecules and a subsequent HOR. As for the ORR, the turnover frequency (TOF) of free Pt sites that are not occupied by calix [4]arene changes little (decreasing from 9 to 6 at +0.80 V), while the TOF for HOR increases sharply from 8 to 388 at +0.1 V as the coverage of calix [4]arene increases. Clearly, the decrease of ORR activity can be attributed to blocking effect, i.e., the calix [4]arene molecules block the adsorption of O 2 and its dissociation. However, the reason for enhanced HOR on the remaining Pt sites is not clear yet, and may be related to electronic effects, and/or mass-transfer limit for naked surfaces. The observed exceptional selectivity is not unique to the Pt(111)-calix system and has also been observed on Pt(100) and polycrystalline Pt electrodes, as well as Pt nanoparticles. 27,28 The successful preparation of oxygen-tolerant Pt catalysts for HOR indicates that surface functionalization of Pt surfaces is a promising approach to tuning the selectivity of electrocatalysts, which has great applications such as methanol-tolerant ORR catalysts.

Modification of Pt(111) with cyanide
Cyanide (CN -) anions can adsorb strongly on metal surfaces. On the Pt(111) surface, they form a well defined (23  23)-R30 structure (inset to Figure 3a, denoted as Pt(111)-CN ad ). 29,30 The structure consists of hexagonally packed arrays, each containing six CN groups adsorbed on top of a hexagon of Pt atoms surrounding a free Pt atom. The cyanide coverage is 0.5. The cyanide adlayer on Pt(111) is remarkably stable, and can survive even after repetitive potential cycling between +0.06 and +1.10 V (vs RHE). 31 The CNgroups act as a third body, blocking Pt atoms that they occupy, but leaving other Pt atoms free onto which H, OH, or CO can still adsorb. The negative charge of the CN ad dipole can also interact with other ions or polar molecules, so that the Pt(111)-CN ad electrode has been considered as a chemically modified electrode.
Cuesta used the Pt(111)-CN ad as a model catalyst to probe the reaction pathways of methanol electrooxidation. 32 As shown by the dashed curve in Figure 3a, there is no hysteresis of the oxidation current of methanol (+0.83 V) between the positive and the negative going sweeps on this surface. In addition, the hydrogen adsorption region (+0.05 ~ +0.5 V) remains unaffected in comparison with that recorded in a blank solution (solid curve). Electrochemical in situ FTIR studies confirm no CO formation in methanol oxidation on the Pt(111)-CN ad electrode ( Figure 3b). In addition to the product of CO 2 (~ 2340 cm -1 ) www.intechopen.com and the original CN adlayer (~2100 cm -1 ), no adsorbed CO signals can be observed in the region of 1800 to 2100 cm -1 . There are four possible free Pt sites on the Pt(111)-CN ad : (i) a single Pt atom; (ii) two adjacent Pt atoms; (iii) three Pt atoms arranged linearly; (iv) three Pt atoms arranged forming a chevron with a 120° angle. No adsorbed CO formation on these reactive sites suggests that the dehydrogenation reaction of methanol to CO needs at least three contiguous Pt atoms, likely as triangle assembly (Figure 3c). In another study, the Markovic group uses the Pt(111)-CN ad as a model catalyst for ORR. 33 In acidic solutions, some anions such as SO 4 2-and PO 4 3-can adsorb strongly onto Pt surfaces (i.e., specific adsorption), and block surface sites for the adsorption of O 2 . As a result, the ORR activity of Pt catalysts is much lower in H 2 SO 4 (or H 3 PO 4 ) than in HClO 4 . 34 In the latter case, the adsorption of ClO 4 -is negligible. It has been found that if the Pt(111) electrode is modified with cyanide, the adsorption of SO 4 2-and PO 4 3-is greatly suppressed ( Figure 4 a and b, Region II), and the butterfly peaks as well as the specific adsorption of SO 4 2-almost disappear. This can be simply understood in terms of through-space electrostatic repulsive interactions between the electronegative CN ad adlayer and the negative SO 4 2- In a 0.1 M KOH solution, the behaviors change greatly. In this solution, the Pt(111)-CN ad electrode shows a 50-fold decrease in ORR activity as compared to the naked electrode. The interaction between the CN ad adlayer and the hydrated K + cations may lead to the formation of large K + (H 2 O) x -CN ad clusters, which are 'quasi-specifically adsorbed' on Pt, hence acting as large site blockers for the adsorption of reactants such as O 2 and H 2 O (see Figure 5c). 35 These studies of the interactions between the CN ad adlayer and H 2 O, oxygen species, and spectator anions shed light on the role of covalent and non-covalent interactions in controlling the ensemble of Pt active sites required for high turn-over rates of ORR and methanol oxidation, and may be helpful in the quest of identifying new directions in designing electrochemical interfaces that can bind selectively reactive and spectator molecules.

Surface-functionalization of metal nanoparticles with improved electrocatalytic properties
Surface functionalization of massive single crystal electrodes offers a unique structural framework for the fundamental studies of fuel cell electrocatalysis. Yet, from the practical points of view, the preparation of surface-functionalized metal nanoparticles is necessary. In fact, most catalysts are in the form of nanosized particles dispersed on varied substrate supports. Herein, we highlight two cases, i.e., aryl-and phosphine-stabilized Pt and Pd nanoparticles, and their applications in fuel cell electrocatalysis.

Surface functionalization by aryl groups
The modification of electrode surfaces with aryl groups through electrochemical reduction of diazonium salts has been studied rather extensively. 36 A variety of electrode substrates, such as carbon, silicon, copper, iron, and gold, have been used. The covalent attachment of the aryl fragments onto the electrode surfaces is via the following mechanism 37 Ar-N≡N + + e  Ar  + N 2 (1) At a negative enough potential, diazonium ions are electrochemically reduced to aryl radicals releasing nitrogen molecules (step (1)). The radicals are very active, and they are readily bonded onto the electrode surfaces (step (2)). Theoretical calculations predict that the adsorption energies vary from 25 to 106 kcal mol -1 on different substrates. 37 The strong bond strength suggests that the interaction is chemical in nature, i.e., covalent bond.
In recent years, aryl-stabilized metal nanoparticles through metal-carbon (M-C) covalent bonds, synthesized by the co-reduction of metal precursors and diazonium salts, have received great interest. [37][38][39] The formation mechanism is similar to that in electrochemical reduction. 36,38 For instance, for Au nanoparticles, surface enhanced Raman scattering (SERS) have provided direct evidence for the formation of Au-C covalent bonds on the nanoparticle surface. 40 Mirkhalaf and co-workers are the first to report the synthesis of Au and Pt nanoparticles stabilized by 4-decylphenyl fragments through the co-reduction of the corresponding metal precursors and diazonium salt. 38 The resulting Pt nanoparticles show a smaller core + Ar   -Ar www.intechopen.com diameter (~ 3 nm) than the Au nanoparticles (8.1 nm), due to the stronger bonding of the aryl group on Pt than on Au. The 1 H NMR spectra of the Pt nanoparticles in CD 2 Cl 2 solution only show some very board peaks from the 4-decylphenyl ligands. The broadening of the NMR peaks is ascribed to the fast spin-spin relaxation of atoms close to the metal core, and confirms the direct linkage of the 4-decylphenyl fragments onto Pt surfaces.
Using a similar method, our group has synthesized Pd nanoparticles stabilized by a variety of aryl moieties, including biphenyl, ethylphenyl, butylphenyl, and decylphenyl groups. 39,41 Of these, the average core size of the Pd nanoparticles stabilized by butylphenyl group (Pd-BP) is 2.24 ± 0.35 nm (Figure 6a and b), 41 which is much smaller than that of commercial Pd black catalyst (10.4 ± 2.4 nm, Figure 6c and d). Thermogravimetric analysis (TGA) indicates that the organic component is about 19% with a main mass-loss peak at 350 C (Figure 7a). On the basis of the weight loss and the average core size of the Pd nanoparticles, the average area occupied by one butylphenyl ligand on the Pd nanoparticle surface is estimated to be about 21 Å 2 , close to the typical value (~20 Å 2 ) for long-chain alkanethiolate adsorbed on metal nanoparticles. 42  www.intechopen.com is estimated to be 122 m 2 g -1 for Pd-BP nanoparticles, which is about 3.6 times higher than that of commercial Pd black (33.6 m 2 g -1 ). The theoretical ECSA of Pd-BP nanoparticles is calculated to be 223 m 2 g -1 , on the basis of the average core size (2.24 nm, Figure 6). This indicates that over 50% of the particle surface was accessible under electrochemical conditions. Due to the remarkably high ECSA and surface functionalization, the Pd-BP nanoparticles exhibit a high mass activity in the electrocatalytic oxidation of formic acid. Figure 8c depicts the voltammograms of Pd-BP and Pd black catalysts recorded in a 0.1 M HCOOH + 0.1 M HClO 4 solution at a potential scan rate of 100 mV s -1 . The peak current density, normalized to the Pd mass, is as high as 3.39 A mg -1 Pd on the Pd-BP nanoparticles, which is about 4.5 times higher than that of the Pd black (0.75 A mg -1 ). In addition, the peak potential for HCOOH oxidation is also negatively shifted by 80 mV. The Pd-BP nanoparticles also possess higher stability of the voltammetric current than Pd black for formic acid oxidation. Figure 8d shows the current-time curves recorded at 0.0 V for 600 s. The Pd-BP catalysts maintain a mass current density that is 2.7 to 4.8 times higher than that of Pd black. The high electrocatalytic activity of Pd-BP nanoparticles may be correlated with butylphenyl www.intechopen.com functionalization of the nanoparticle surface. As indicated by cyclic voltammograms where the currents are normalized to ECSA (Figure 8b), the adsorption of (bi)sulphate ions at around 0 V diminishes substantially on the Pd-BP electrode, 44 and the initial oxygen adsorption at around +0.50 V in the positive-going scan is also blocked, in comparison to Pd black. Such a discrepancy of surface accessibility and reactivity indicates that butylphenyl group may provide steric hindrance for (bi)sulphate adsorption, and also change the electronic structure of Pd.
Butylphenyl modified Pt (Pt-BP) nanoparticles have also been prepared by the co-reduction of diazonium salt and H 2 PtCl 4 with NaBH 4 as the reducing agent. 45 TEM images (Figure 9) show that the as-prepared Pt-BP nanoparticles are well-dispersed and uniform, and well clear (111) lattice fringes can be observed. The average core size of Pt-BP nanoparticles is determined to be 2.93  0.49 nm, which is comparable to the commercial Pt/C (3.1 nm) catalyst that is used in the controll experiment. The IR transmission spectrum of Pt-BP nanoparticles also exhibits the characteristic peaks of the butylphenyl group, e.g., stretching vibration of the C-H moiety on the phenyl ring (3029 cm -1 ), saturated C-H stretches in the butyl group (2955, 2929, 2858 cm -1 ), aromatic ring skeleton vibration (1577 cm -1 ), and out-of-plane C-H deformation vibration for parasubstituted aromatic rings (827 cm -1 ). The amount of ligands on the Pt-BP nanoparticles is quantitatively determined by TGA measurements to be about 27%. Therefore, the average area occupied by one butylphenyl ligand can be estimated to be about 5.8 Å 2 , considerably lower than that observed with the aforementioned Pd nanoparticles. 41 Such a low occupied area of aryl group on Pt surfaces suggests there may exist some multilayer polyaryl structure. 46 It is well established that Pt catalysts for formic acid electrooxidation are easily poisoned by adsobed CO (CO ad ) intermediates. 25 Interestingly, the functionalization of Pt nanoparticles by the butylphenyl groups greatly suppresses the CO poisoning. Figure 10 compares the cyclic voltammograms of formic acid oxidation on the Pt-BP (red curve) and commercial Pt/C catalysts (black curve). On the Pt/C, along with increasing electrode potential, a weak and broad oxidation peak appears at around +0.30 V, followed by a sharp peak at +0.60 V in the forward scan. The former is attributed to the direct oxidation of formic acid to CO 2 , while the latter is assigned to the oxidation of the CO ad that is generated from the spontaneous dissociative adsorption of formic acid on Pt surfaces. 48 After the stripping of CO ad , formic acid can then be readily oxidized on the clean Pt surface, as indicated by a very intensive current peak in the reverse scan. In contrast, on the Pt-BP nanoparticles, the hysteresis of the oxidation current of formic acid between forward and reverse scans is much less significant, and only a hump can be observed in the potential region around +0.60 V for CO ad oxidation. This fact indicates that the CO poisoning pathway has been greatly suppressed on the Pt-BP nanoparticles. As a result, the peak current density of formic acid oxidation on the Pt-BP (1.03 A mg -1 Pt ) is about 4 times larger than that on the Pt/C (0.25 A mg -1 Pt ) in the forward scan at round +0.30 V. The suppression of CO poisoning for formic acid on the Pt-BP nanoparticles is further confirmed by electrochemical in situ FTIR spectroscopic measurements ( Figure 11). With Pt/C (panel b), the spectra show a downward band at 2343 cm -1 that is assigned to CO 2 , the oxidation product of formic acid, and another at around 2050 cm -1 to linearly bonded CO www.intechopen.com (CO L ), the poisoning intermediates. Clearly, unlike naked Pt, there is negligible adsorbed CO on the Pt-BP electrode (panel a), confirming that the functionalization of butylphenyl group have effectively blocked the poisoning pathway of formic acid oxidation. As a result, Pt-BP can yield more CO 2 than the Pt/C at a same potential. It has been known that there are at least two effects responsible for the improved electrocatalysts for formic acid oxidation: third-body effects and electronic structure effects. 49,50 It is generally accepted that the dehydration of formic acid to form CO ad is a site www.intechopen.com demanding reaction, and several neighboring Pt sites are required. 50 Thus, this pathway can be blocked if Pt surfaces are covered by foreign atoms (third body). Clearly, the high catalytic activity of the Pt-BP nanoparticles may be, at least in part, reasonably attributed to third-body effects, i.e., the butylphenyl group provides steric hindrance for formic acid to form CO ad .
The two examples listed above demonstrate that surface functionalization of noble metal nanoparticles with aryl groups is a promising route towards the preparation of electrocatalysts with improved performance for fuel cell electrocatalysis.

Surface functionalization by phosphine
Organophosphine ligands can strongly bind onto metal surfaces, especially for Pt group metals. 51,52 Pietron and co-workers carry out electrochemical studies to examine the impacts of triphenylphosphine triphosphonate (TPPTP, Figure 12a) on the electrocatalytic activity of Pt nanoparticles for ORR. 53 The TPPTP-capped Pt nanoparticles are prepared through ligand exchange reactions between glycol-stabilized Pt nanoparticles and TPPTP for 2 to 4

TPPT P
www.intechopen.com days. 54 The core size of the Pt nanoparticles is about 2.1 nm. Studies based on X-ray absorption spectroscopy including XANES and EXAFS indicate that the coverage of TPPTP is about 0.3 ML, and the binding to Pt surfaces is via the P atom (Pt-P-P<tp) at +0.54 V (RHE), and approximately one-half of this converts to a Pt-O-P<tp linkage at +1.0 V. 55 Electrochemical cyclic voltammograms (Figure 12b) show the features of weakly adsorbed hydrogen on TPPTP-caped Pt nanoparticles (dotted curve). They attribute this observation to the surface ligands that change the electronic energy of the Pt frontier orbitals and hence alter the hydrogen adsorption electrochemistry. More importantly, the TPPTP ligands weaken the oxygen adsorption on Pt, as evidenced by the positive shift of 40 mV for both onset potential and reduced desorption peak of oxygen species, as compared to that on a "bare" Pt/VC catalyst with a similar core size of 2.4 nm (solid curve). The weakening of the Pt-O bonds may improve the ORR electrocatalytic activity, since the reduced desorption of oxygen species is the rate-determined step for ORR on Pt. 7 Figure 12c and d show the polarization curves and Tafel plots of area-specific ORR activities for these Pt catalysts (along with a somewhat larger "bare" Pt/VC sample of about 4.4 nm in diameter) in O 2saturated 0.1 M HClO 4 with a rotating disc electrode. The area-specific activity of TPPTPcaped Pt is enhanced by 22% in comparison with naked Pt; however, the mass-specific activity is inferior due to lower specific electrochemical area, as shown in Figure 12c. Two factors may contribute to the enhancement in area-specific activity. One is the weakening of Pt-O bonds, as mentioned above, and the other is hydrophobic effect of the aryl groups in TPPTP, which may repel water and inhibit Pt-oxide formation.
Thiol molecules have also be used to functionalize Pt nanoparticles, [56][57][58] but no considerable improvement of the ORR activity has been achieved.

Conclusion
In this chapter, we highlight some new progress in the study of the impacts of deliberate chemical functionalization of massive single crystal planes and noble metal nanoparticles with specific molecules, ions and molecular fragments on their catalytic performance in fuel cell electrochemistry. Through the interactions between the modifying groups and reactants (or spectator molecules/ions), and ligand-mediated electronic effects, the electrocatalytic activity and selectivity may be greatly tuned. This approach may have extensive applications in analytical, synthetic and materials chemistry as well as in chemical energy conversion and storage, and selective fuel production.
Further studies in this field may include the following aspects: (1) selection and optimization of functional molecules that can greatly change surface electronic structure of the metal substrates, but not occupy too many metal surface sites; (2) improvement of the stability of the modifying molecules, especially under the electrochemical conditions for ORR; (3) exploration of functional molecules that have synergistic effects for target reactions; and (4) identification of modifying molecules that possess special physicochemical properties, such as high solubility for O2, great hydrophilic or hydrophobic characteristics, and an environment similar to enzyme, that are optimal for fuel cell electrocatalysis.

Acknowledgments
The authors are grateful to the National Science Foundation and the ACS-Petroleum Research Funds for financial supports of the research activities. In a classical approach materials science is mainly dealing with interatomic interactions within molecules, without paying much interest on weak intermolecular interactions. However, the variety of structures actually is the result of weak ordering because of noncovalent interactions. Indeed, for self-assembly to be possible in soft materials, it is evident that forces between molecules must be much weaker than covalent bonds between the atoms of a molecule. The weak intermolecular interactions responsible for molecular ordering in soft materials include hydrogen bonds, coordination bonds in ligands and complexes, ionic and dipolar interactions, van der Waals forces, and hydrophobic interactions. Recent evolutions in nanosciences and nanotechnologies provide strong arguments to support the opportunity and importance of the topics approached in this book, the fundamental and applicative aspects related to molecular interactions being of large interest in both research and innovative environments. We expect this book to have a strong impact at various education and research training levels, for young and experienced researchers from both academia and industry.