Solubilization and Purification of Single-Wall Carbon Nanotubes in Water by in Situ Radical Polymerization of Sodium 4-Styrenesulfonate Shuhui Qin, † Dongqi Qin, † Warren T. Ford,* ,† Jose E. Herrera, ‡ Daniel E. Resasco, ‡ Sergei M. Bachilo, § and R. Bruce Weisman § Department of Chemistry, Oklahoma State University, Stillwater, Oklahoma 74078; School of Chemical Engineering and Materials Science, University of Oklahoma, Norman, Oklahoma, 73019; and Department of Chemistry, Rice University, Houston, Texas 77005 Received February 16, 2004 Revised Manuscript Received April 8, 2004 Because of outstanding electrical conductivity, ther- mal conductivity, and mechanical strength, single-wall carbon nanotubes (SWNT) have enormous potential in field emission displays, supercapacitors, molecular com- puters, and ultrahigh-strength materials. 1,2 For optimal performance in most applications, the SWNT should be separated into individual tubes or bundles of only a few tubes. However, the as-prepared SWNT contain impuri- ties of metal catalyst particles and amorphous carbon, and because of strong van der Waals attraction, the SWNT pack into bundles that aggregate into tangled networks. Dissolution of SWNT in water, which is important because of potential biomedical applications and biophysical processing schemes, has been facilitated by surfactants and polymers and by chemical modifica- tion. 3-11 Here we report a method by which pristine SWNT are solubilized, separated from catalyst particles, and separated from excess dispersant to produce SWNT with grafted poly(sodium 4-styrenesulfonate) (PSS) as an aqueous solution that is stable indefinitely. The method is illustrated in Scheme 1. As in some other functionalizations of SWNT, 12-14 the process requires no pretreatment. Debundling and functionalization of SWNT are achieved in one step with no high shear mixing or heavy sonication, which break down SWNT to shorter lengths. 5,15 A mixture of 40 mg of pristine HiPco SWNT, 4.0 g of sodium 4-styrene- sulfonate (NaSS), and 40 mg of potassium persulfate as a free radical initiator was stirred at 65 °C for 48 h. Catalyst residues and amorphous carbon were removed by gentle centrifugation, and excess unbound PSS was removed by ultrafiltration and ultracentrifugation. The final solution contained 68 mg of SWNT-PSS in 100 mL of water. A detailed procedure is in the Supporting Information. Elemental analysis (CHS) corresponded to 45 wt % of PSS in the SWNT/PSS composite. The 1 H NMR spectra of the PSS in the ultrafiltrate and the SWNT-PSS were the same, which suggests that the molecular weight of the attached PSS is high. In control experiments, stirring pristine SWNT with preprepared PSS or with potassium persulfate and sodium p-tolu- enesulfonate but no monomer by the method used during the polymerization or sonicating in a cleaning bath failed to disperse the SWNT. SWNT also can be dispersed into water by surfactants and by high shear mixing or sonication with PSS, but large excesses of the surfactants or PSS are required. 5,15 We attribute the stability of the SWNT with such a small amount of PSS to covalent bonding of the polymer to the SWNT. One attached polymer coil protects a large area of the SWNT surface from van der Waals attraction to other SWNT. By analogy to the addition of polystyryl radicals at diffusion-controlled rates to aggregates of [60]fullerene in solution, 16 bundles of SWNT should also react with polymer radicals, although at lesser rate constants because of the lesser strain of the sidewalls of SWNT than of [60]fullerene. Additions of nonpoly- meric radicals to SWNT are well-known. 17 The distributions of diameters and lengths of the functionalized SWNT were analyzed by tapping mode atomic force microscopy (AFM). Figure 1 shows contour lengths from several hundred nanometers to several micrometers and a representative diameter of 1.2 nm. The range of diameters of pristine HiPco SWNT is 0.6- 1.3 nm. 18 Larger area AFM images and transmission electron microscopy (TEM) images at resolution too low to detect individual tubes show bundles of the SWNT- PSS that are much smaller than the bundles of the pristine SWNT (Figures S1 and S2). TEM shows none of the catalyst particles that were abundant in the pristine SWNT (Figures S2a and 2b). The Raman spectrum of the functionalized SWNT in Figure 2a shows a disorder (D) band at 1315 cm -1 in addition to the radial breathing band (180-260 cm -1 ) and tangential (G) band at 1590 cm -1 . The intensity of the D band is indicative of the degree of covalent functionalization of the nanotube framework. The radial breathing bands are shifted an average of 5 cm -1 to higher frequency by functionalization with the PSS, which indicates debundling during the polymerization. † Oklahoma State University. ‡ University of Oklahoma. § Rice University. * Corresponding author: e-mail wtford@okstate.edu. Figure 1. AFM height image of SWNT-PSS (3 μm × 3 μm). The arrows point to a 1.2 nm height difference. Scheme 1 3965 Macromolecules 2004, 37, 3965-3967 10.1021/ma049681z CCC: $27.50 © 2004 American Chemical Society Published on Web 04/28/2004