Journal of Hazardous Materials 290 (2015) 60–68

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Influence of surface heterogeneities on reversibility of fullerene (nC60 ) nanoparticle attachment in saturated porous media Chongyang Shen a , Mengjia Zhang a , Shuzhen Zhang a , Zhan Wang a , Hongyan Zhang b , Baoguo Li a , Yuanfang Huang a,∗ a b

Department of Soil and Water Sciences, China Agricultural University, Beijing 100193, China Department of Applied Chemistry, China Agricultural University, Beijing 100193, China

h i g h l i g h t s • • • • •

We examine reversibility of fullerene nC60 particle attachment on collector surfaces. Attachments were more in sand than those in glass beads. Significant detachments were observed in sand upon reduction of ionic strength. The detachment was due to removal of surface chemical heterogeneities. Surface curvature plays a critical role in both attachment and detachment.

a r t i c l e

i n f o

Article history: Received 19 December 2014 Received in revised form 23 February 2015 Accepted 25 February 2015 Available online 26 February 2015 Keywords: Fullerene C60 Nanoparticle Transport Porous media Surface heterogeneity

a b s t r a c t This study systematically investigated influence of surface roughness and surface chemical heterogeneity on attachment and detachment of nC60 nanoparticles in saturated porous media by conducting laboratory column experiments. Sand and glass beads were employed as a model collectors to represent a different surface roughness. The two collectors were treated by washing with only deionized water or by using acids to extensively remove chemical heterogeneities. Results show that both attachment and detachment were more in the acid-treated sand than those in the acid-treated glass beads. The greater attachment and detachment were attributed to the reason that sand surfaces have much more nanoscale asperities, which facilitates particle attachment atop of them at primary minima and subsequent detachment upon reduction of ionic strength. No detachment was observed if the water-washed collectors were employed, demonstrating that the couple of chemical heterogeneity with nanoscale roughness causes irreversible attachment in primary minima. Whereas existing studies frequently represented surface rough asperities as regular geometries (e.g., hemisphere, cone, pillar) for estimating influence of surface roughness on Derjaguin–Landau–Verwey–Overbeek (DLVO) interaction energies, our theoretical calculations indicate that the assumptions could underestimate both attachment and detachment because these geometries cannot account for surface curvature effects. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The fullerene C60 has been widely used in various fields (e.g., biotechnology, electronics, optics, and cosmetics) due to its unique physical and chemical properties [1–3]. The wide use of fullerene C60 inevitably causes its release into the environment during their production, application, and disposal [4–7]. Despite very low solubility, the fullerene C60 can form stable nanoscale aggregates (nC60 ) because the fullerene C60 surface can acquire negative charge in

∗ Corresponding author. Tel.: +86 1062732963; fax: +86 1062733596. E-mail address: [email protected] (Y. Huang). http://dx.doi.org/10.1016/j.jhazmat.2015.02.067 0304-3894/© 2015 Elsevier B.V. All rights reserved.

aqueous system [8,9]. The fullerene nC60 nanoparticles, however, were reported to be toxic to a variety of living organisms [10–15]. In addition, the nC60 nanoparticles can absorb other contaminants and enhance their transport in porous media [16–18]. Therefore, understanding of the transport of nC60 nanoparticles in porous media is of importance for preventing groundwater from being contaminated by the nC60 nanoparticles and their associated contaminants. Attachment is one of the primary factors controlling particle transport in porous media [19]. Considerable studies [4,5,9,20–27] have been conducted to investigate the mechanisms controlling attachment of nC60 nanoparticles on collector surfaces for the past decade. The attachment of nC60 nanoparticles was

C. Shen et al. / Journal of Hazardous Materials 290 (2015) 60–68

found be qualitatively in agreement with the prediction by Derjaguin–Landau–Verwey–Overbeek (DLVO) theory [28]. Specifically, increasing ionic strength increases attachment of nC60 nanoparticles because DLVO energy barrier is lower at higher ionic strength, thus, the nC60 nanoparticles are easier to overcome the energy barrier and attach in primary minima. In addition to ionic strength, the attachment of nC60 nanoparticles is influenced by factors such as flow velocity [9,20], sunlight [24], collector surface coating [22,23,29], the presence of macromolecules in the pore solutions [5,21,22,24,28], and the preparation methods of nC60 nanoparticles [21]. The detachment of nC60 nanoparticles from collector surfaces, however, has received very limited attention. Particularly, Chen and Elimelech [28] conducted quartz crystal microbalance experiments to examine the detachment of nC60 nanoparticles during transients in solution chemistry. The nC60 nanoparticles were first attached onto silica-coated quartz surfaces at 30 mM NaCl or 0.6 mM CaCl2 . The system was then flushed with 1 mM NaCl or deionized (DI) water. The detachment of nC60 nanoparticles was found to be minor. Therefore, the attachment of nC60 nanoparticles was regarded to be chemically irreversible in the literature. Chen and Elimelech [28], however, only examined detachment of nC60 nanoparticles that were attached under favorable conditions (i.e., in the absence of DLVO repulsive energies). The detachment of nC60 nanoparticles that were initially attached under unfavorable conditions has not been investigated to date. In addition, Chen and Elimelech [28] adopted a parallel flow chamber system and the quartz substrates used were relatively smooth. The influence of surface roughness on both attachment and detachment of fullerene nC60 nanoparticles in porous media is not very clear. This study conducted laboratory column experiments to systematically investigate the influence of surface roughness on attachment and subsequent detachment of nC60 nanoparticles in saturated porous media. Sand and glass beads were used as model collectors to represent different surface roughness. We showed that attachments of nC60 nanoparticles were more in sand than in glass beads. Considerable detachments of nC60 nanoparticles attached in NaCl at high ionic strengths (e.g., ≥0.01 M) in sand columns were observed upon reduction of ionic strength. The results are in contrast to previous observations [28,30] and challenge the common belief that attachment of nC60 nanoparticles is chemically irreversible. The discrepancy was attributed to the reason that the collectors used in this study were extensively treated to remove the chemical heterogeneities. Findings from this study have implication to design of surfaces for reversible attachment of nC60 nanoparticles in engineered systems.

2. Materials and methods 2.1. Fullerene particle suspensions and porous media Fullerene (C60 , 99.9%, purified by sublimation) was purchased from Sigma–Aldrich (St. Louis, MO). The stable aqueous suspensions of nC60 were prepared using the method in previous studies

61

[31,32]. This method is based on transfer of fullerenes from a nonpolar solvent (toluene) to a polar one (water) under ultrasonic conditions. Briefly, a 5 mL mixture of toluene and nC60 at a concentration of 1 mg/L was mixed with 20 mL doubly DI water using ultrasonic treatment for about 3 h. The toluene was allowed to evaporate off after mixing. The suspension was then filtered through a 0.45 ␮m microfilter. Using a TOC-meter (TOC-VCPN , Shimadzu, Japan), the concentrations of the resulting nC60 stock suspensions were determined to be about 9 mg/L. The nC60 stock suspensions were diluted in degassed NaCl solutions to yield nC60 suspensions at different ionic strengths. The concentration of nC60 in aqueous suspensions was determined by UV–vis spectrophotometry (DU Series 800, Beckman Instruments, Inc., Fullerton, California) at a wavelength of 344 nm. Using dynamic light scattering (Zetasizer Nano ZS, Malvern Instruments Ltd., Southborough, Massachusetts), the mean diameters of the nC60 particles were determined to be 242 nm, 253 nm, and 281 nm at 0.001 M, 0.01 M, and 0.1 M NaCl, respectively. Quartz sand and glass beads with sizes ranging from 300 to 355 ␮m were used as model collector grains. The sand was sieved from Accusand 40/60 (Unimin Corporation, Le Sueur, Minnesota) with a stainless steel mesh. The glass beads were obtained from Mo-Sci Corporation (Rolla, Missouri). The procedure from Zhuang et al. [33] was used to extensively remove metal oxides and other impurities from the sand. Surface roughness of the cleaned sand and glass beads was measured in our previous study [34] using a bioscope atomic force microscopy (Veeco Instruments Inc., Plainview, New York) mounted on an Axiovert 200 inverted fluorescent microscope (Carl Zeiss Inc.). The electrophoretic mobilities of the nC60 nanoparticles and collectors (i.e., sand and glass beads) were measured at different ionic strengths at pH 7 using a ZetaPlus analyzer (Brookhaven Intruments Corp., Holtesville, New York) and were converted to zeta potentials using the Henry equation. The obtained zeta potentials of nC60 nanoparticles and collectors were presented in Table 1. 2.2. Column attachment and detachment experiments Colloid transport experiments were conducted in acrylic columns packed with the sand or glass beads. The column was 3 cm in diameter and 10 cm long with a top and a bottom plate. The collectors were wet-packed in deionized (DI) water with vibration to minimize any layering and air entrapment in the column. The porosities of the packed beds were determined to be 0.33 and 0.38 for the packed sand and glass beads, respectively (based on a density of 2.65 g/cm3 for sand and 2.5 g/cm3 for glass beads). Analytical reagent-grade NaCl (Sigma–Aldrich, Le Sueur, Missouri) and DI water were used to prepare electrolyte solutions at desired ionic strengths. All the electrolyte solutions were buffered to pH 7 by addition of NaHCO3 and Na2 CO3 . Column experiments were performed over a range of ionic strengths (0.001, 0.01, and 0.1 M) at an approach velocity of 1.2 × 10−5 m/s. For each experiment, background electrolyte solution was first delivered to the column upward for at least 20 PVs (Pore volumes) to standardize the ionic strength and pH of the system. Then a three-step

Table 1 Zeta potentials of sand, glass beads, and nC60 nanoparticles, and calculated maximum energy barriers (Umax ), secondary-minimum depths (Usec ) and distance for the particles at different solution ionic strengths. Usec Zeta potential (mV)

Umax (kT)

Depth (kT)

Distance (nm)

I (M)

Sand

Glass beads

nC60

Sand

Glass beads

Sand

Glass beads

Sand

Glass beads

0.001 0.01 0.1

−49.3 −41.5 −32.7

−43.8 −38.4 −30.6

−23.9 −16.6 −8.6

113.5 49.3 –

109.0 47.2 –

0.0006 0.54 –

0.0006 0.56 –

94.2 21.9 –

94.0 21.5 –

62

C. Shen et al. / Journal of Hazardous Materials 290 (2015) 60–68

(DL) interaction, and the Born (BR) repulsion between element dS and dA, respectively. Details about use of the SEI technique to calculate DLVO interaction energies can be found in Supplementary material. 3. Results 3.1. Attachment of nC60 nanoparticles Fig. 2 presents the representative breakthrough curves for the nC60 nanoparticles in sand and glass bead packed columns. The normalized effluent particle concentration C/C0 was plotted as a function of pore volume. Each breakthrough curve includes three segments corresponding to the three experimental phases. In phase 1, the nC60 nanoparticles were retained in porous media at different ionic strengths. Because the ratios of particle diameter to collector diameter (