Strongly Enhanced Interaction between Evaporated Pt Nanoparticles and Functionalized Multiwalled Carbon Nanotubes via Plasma Surface Modifications: Effects of Physical and Chemical Defects De-Quan Yang and Edward Sacher* Regroupement Que ´ be ´ cois de Mate ´ riaux de Pointe, De ´ partement de Ge ´ nie Physique, E Ä cole Polytechnique, C.P. 6079, Succursale Centre-Ville, Montre ´ al, Que ´ bec H3C 3A7, Canada ReceiVed: August 14, 2007; In Final Form: NoVember 27, 2007 Oxygen and argon plasmas were used to modify multiwalled carbon nanotubes (CNTs) to improve their interfacial interaction with subsequently deposited Pt nanoparticles. In contradistinction to what was found in the case of highly oriented pyrolytic graphite (HOPG), X-ray photoelectron spectroscopy (XPS) confirms the introduction of chemical functionalizations by oxygen plasma treatment; however, as in the case of HOPG, argon plasma treatment produced physical defects. Transmission electron microscopy (TEM) provided visual evidence of the interaction of subsequently evaporated Pt with treated CNTs, showing it to have been enhanced by both plasma treatments. XPS and TEM analyses demonstrate that the enhancement is due to similar interactions of Pt nanoparticles with both types of treated CNTs, although not to the same extent: X-ray photoelectron spectroscopy gives no evidence of chemical bonds formed for either plasma treatment. The morphology of the Pt nanoparticles changes with the deposition rate, which may be influenced by the limited availability of the CNT surface. Introduction The surface modification and functionalization of carbon nanotubes (CNTs), both single-walled (SWCNTs) and multi- walled (MWCNTs), have been important issues in their rational and predictable manipulation. The ability to control the type of functionalization permits the direct tailoring of their physical and chemical properties for specific applications. Several approaches have been used for functionalizing CNTs: (1) wet chemical processes, including strong acid oxidation, 1,2 direct reaction with fluorine and subsequent nucleophilic substitution, 3,4 electrochemical or thermal reduction of aryl diazonium salts, 5,6 the addition of radicals, nitrenes, or carbenes, 7 supramolecular complexation with detergents, proteins, or polymers, 8-10 ozo- nation and subsequent derivatization, 11-13 ultrasonication with organic materials, 14 and electrodeposition 15 and (2) dry pro- cesses, including both nonreactive and reactive plasmas 16-19 and low-energy ion beam bombardment in a vacuum. 20 These modifications can cause CNT damage 20 that may affect their electrical and mechanical properties. 21,22 As compared to wet approaches, dry vacuum processing may be easier to control, with relatively less contamination. Noble metal nanoparticles (NPs), supported on CNTs, have attracted considerable interest over the past few years, due to potential applications in the catalyst industry 23-27 (e.g., proton exchange membrane fuel cells (PEMFC) require a stronger interaction between catalyst NPs and CNTs on which they are deposited, to increase catalyst loading and control NP size). 26-29 A similar concern exists in the nanoelectronics industry because the NP-CNT interaction is directly associated with the contact resistance of electrodes made from these materials. 30 In general, the interactions of both transition and noble metals with CNTs are very weak, 31,32 similar to their interactions with highly oriented pyrolytic graphite (HOPG). 33 As a result, the CNT surface must be functionalized to enhance the adhesion of the metal NPs. We previously studied the effect of plasmas on HOPG, 34 where we found that Ar, O 2 ,N 2 , and H 2 O plasmas broke C-C bonds, producing -C • free radical defects that, on atmospheric exposure, reacted with components of air (H 2 O and O 2 ) to produce oxidized carbon species (C-OH, CdO, and COOH); these species were able to hydrogen bond to hydroxyl groups on the small amount of oxide present on the surfaces of Pt nanoparticles, even in a vacuum. We found that CNTs give similar results but with significant differences. One such difference concerns the action of an O 2 plasma: upon atmo- spheric exposure, the amount of oxidation such a treatment introduced on the HOPG surface 34 was within experimental error of the amounts introduced on Ar and N 2 treatments, making it obvious that the three plasma treatments functioned in the same manner: to break C-C bonds, producing -C • free radicals. However, in the case of MWCNTs, the amount of oxidation on O 2 plasma treatment was found to be 2.5 times that introduced on Ar and N 2 treatments. Clearly, the O 2 plasma treatment of MWCNTs serves not only to produce -C • free radicals but to also react with some of them, giving oxidation products even before atmospheric exposure. Here, we present these results: we contrast the roles of physical surface defects (i.e., free radicals) on MWCNTs, created by Ar plasma treatment (which we compare with Ar + beam irradiation), and that of chemical defects (i.e., oxidized carbon functional groups), introduced by O 2 plasma treatment, on their interaction with evaporated Pt NPs. These physical and chemical defects act as both nucleation and binding sites for the Pt NPs in this study. We present transmission electron microscopy (TEM) evidence of nanoparticle morphologies and surface coverages and X-ray photoelectron spectroscopy (XPS) * Corresponding author. E-mail: edward.sacher@polymtl.ca; tel.: (514) 340-4711, ext. 4858; fax: (514) 340-3218. 4075 J. Phys. Chem. C 2008, 112, 4075-4082 10.1021/jp076531s CCC: $40.75 © 2008 American Chemical Society Published on Web 02/22/2008