Influence of the initial state of carbon nanotubes on their colloidal stability under natural conditions Irène Schwyzer a, b , Ralf Kaegi b , Laura Sigg b , Arnaud Magrez c , Bernd Nowack a, * a Empa e Swiss Federal Laboratories for Materials Science and Technology, CH-9014 St. Gallen, Switzerland b Eawag e Swiss Federal Institute of Aquatic Science and Technology, CH-8600 Dübendorf, Switzerland c EPFL e Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland The colloidal stability of CNTs varies a lot depending on the initial state of the CNTs (dry vs. pre-dispersed), the applied dispersant for pre-suspension, and the composition of the medium. article info Article history: Received 4 January 2011 Received in revised form 11 February 2011 Accepted 26 February 2011 Keywords: Carbon nanotubes Surfactants (anionic, non-ionic, cationic) Natural organic matter (humic/fulvic acid) Long-term colloidal stability abstract The colloidal stability of dry and suspended carbon nanotubes (CNTs) in the presence of amphiphilic compounds (i.e. natural organic matter or surfactants) at environmentally realistic concentrations was investigated over several days. The suspensions were analyzed for CNT concentration (UVevis spectros- copy), particle size (nanoparticle tracking analysis), and CNT length and dispersion quality (TEM). When added in dry form, around 1% of the added CNTs remained suspended. Pre-dispersion in organic solvent or anionic detergent stabilized up to 65% of the added CNTs after 20 days of mild shaking and 5 days of settling. The initial state of the CNTs (dry vs. suspended) and the medium composition hence are critical determinants for the partitioning of CNTs between sediment and the water column. TEM analysis revealed that single suspended CNTs were present in all suspensions and that shaking and settling resulted in a fractionation of the CNTs with shorter CNTs remaining predominantly in suspension. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Along with the forecasted increase in manufacturing and use of carbon nanotubes (CNTs), unintended and accidental releases of CNTs to the environment are becoming likely (Nowack and Bucheli, 2007). Exposure might occur during the production, use and disposal of CNTs and CNT-containing products and is largely determined by the matrices in which these materials are found (Köhler et al., 2008; Wiesner et al., 2009). CNTs may be released into the aquatic environment from point sources, such as factories, landfills and wastewater effluents or from nonpoint sources, such as storm-water runoff, attrition from composites and from wet depo- sition from the atmosphere (Wiesner et al., 2006). According to first modeling approaches the concentrations of CNTs in the aquatic environment are expected to be in the ng/L range (Mueller and Nowack, 2008) and sediments are predicted to be the major envi- ronmental sink (Gottschalk et al., 2009; Koelmans et al., 2009). From an ecological point of view, the partitioning between sediment and the water column is a critical determinant of suspendability with regard to potential impacts of CNTs on the ecosystem. The formation of agglomerates and their subsequent settling initiates partitioning. It is assumed that the better de-bundled and individually suspended the CNTs are, the longer they remain in the water column. Upon release of CNTs into the aquatic environment, water composition and properties (e.g. ionic strength, pH) affect the behavior of CNTs. Pristine CNTs are extremely hydrophobic and cannot easily be dispersed in distilled water (Girifalco et al., 2000). However, it has been shown that natural organic mater (NOM), such as humic and fulvic acid, present in natural systems (Hyung et al., 2007; Kennedy et al., 2009) enhances the solubility of CNTs at environmentally relevant concentrations. The dissolved organic carbon (DOC) concentrations in natural waters range from 0.5 to 30 mg/L (Thurman, 1985). On average, humic acid and fulvic acid accounts for 10% and 40% of the DOC, respectively (Thurman, 1985). It has been proposed that the hydrophobic moieties of NOM adsorb to the hydrophobic surfaces of the CNTs whilst hydrophilic moieties stretch towards the solution and thereby enhance the dispersability of CNTs (Hyung et al., 2007; Kennedy et al., 2009). Another possible interaction is hydrogen bonding between oxygen-containing functionalities on the NOM and polar functionalities (e.g. eCOOH, eOH) on the graphene surface of CNTs (Wang et al., 2009). Even pristine CNTs often contain oxygen-containing functional groups, formed during purification or by incidental exposure to oxidizing agents after arrival in the environment (Cho et al., 2008). NOM * Corresponding author. E-mail address: nowack@empa.ch (B. Nowack). Contents lists available at ScienceDirect Environmental Pollution journal homepage: www.elsevier.com/locate/envpol 0269-7491/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2011.02.044 Environmental Pollution 159 (2011) 1641e1648