Environ. Sci. Technol. 2009, 43, 1354–1359
Influence of Ionic Strength, pH, and Cation Valence on Aggregation Kinetics of Titanium Dioxide Nanoparticles REBECCA A. FRENCH,† ASTRID R. JACOBSON,‡ BOJEONG KIM,† SARA L. ISLEY,§ R. LEE PENN,§ AND P H I L I P P E C . B A V E Y E * ,| Center for NanoBioEarth, Department of Geosciences, Virginia Tech, 4044 Derring Hall, Blacksburg, Virginia 24061, Plants, Soils and Climate Department, Utah State University, 4820 Old Main Hill, Logan, Utah 84322, Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, and SIMBIOS Centre, Abertay University, Bell Street, Dundee DD1 1HG, U.K.
Received September 16, 2008. Revised manuscript received December 9, 2008. Accepted December 17, 2008.
The extensive use of titanium dioxide nanoparticles (nano-TiO2) in many consumer products has raised concerns about possible risks to the environment. The magnitude of the threat may depend on whether nano-TiO2 remains dispersed in the environment, or forms much larger-sized aggregates or clusters. Currently, limited information is available on the issue. In this context, the purpose of the present article is to report initial measurements of the morphology and rate of formation of nanoTiO2 aggregates in aqueous suspensions as a function of ionic strength and of the nature of the electrolyte in a moderately acid to circumneutral pH range typical of soil and surface water conditions. Dynamic light scattering results show that 4-5 nm titanium dioxide particles readily form stable aggregates with an average diameter of 50-60 nm at pH ∼4.5 in a NaCl suspension adjusted to an ionic strength of 0.0045 M. Holding the pH constant, but increasing the ionic strength to 0.0165 M, leads to the formation of micron-sized aggregates within 15 min. At all other pH values tested (5.8-8.2), micron-sized aggregates form in less than 5 min (minimum detection time), even at low ionic strength (0.0084-0.0099 M with NaCl). In contrast, micronsized aggregates form within 5 min in an aqueous suspension of CaCl2 at an ionic strength of 0.0128 M and pH of 4.8, which is significantly faster than observed for NaCl suspensions with similar ionic strength and pH. This result indicates that divalent cations may enhance aggregation of nano-TiO2 in soils and surface waters. Optical micrographs show branching aggregates of sizes ranging from the 1 µm optical limit of the microscope to tens of micrometers in diameter.
nano-TiO2 is used as antimicrobial, antibiotic, and antifungal agents, as ultraviolet (UV) blockers, antiscratch additives, and catalysts. It is routinely found in sunscreens, coatings, plastics, soaps, nanofibers and nanowires, bandages, alloys, and textiles. The U.S. Food and Drug Administration approved its use as a color additive in food, drugs, cosmetics, and contact lenses. Because of its high specific surface area and sorption capacity for ionic and nonionic species (3-5), the use of nano-TiO2 has also been promoted as a scavenger of inorganic and organic contaminants in water treatment plants and in the remediation of polluted subsurface environments (6, 7). This booming demand for nano-TiO2 has spurred significant public alarm about its possible impact on the environment as a result of accidental spills during manufacturing and transport, or its presence in waste, sewage, and runoff. Several panels, environmental organizations, and various researchers (1, 2) are concerned that nanomaterials transported in the environment may contaminate surface and groundwater resources and pose toxicological risks to microorganisms, animals, and humans. In vitro studies have shown that nano-TiO2 produces reactive oxygen species that can induce oxidative damage in bacteria (8), crustaceans (9), and diverse mammal cell types (10, 11). The form of the nanoparticles to which cells are exposed influences their toxicity. For example, aggregate size is a determining factor in the uptake and response of a cell to nano-TiO2 (11) and in the bioavailability of nanoparticles to plant roots, algae, and fungi (12). Unfortunately, the potentially important effects of aggregation have often not been considered in experiments despite mounting evidence that aggregation is the rule rather than the exception. Recent articles stress that even protracted sonication of nano-TiO2 in suspensions (10, 13), or special precautions to ensure full dispersion, cannot prevent the majority of suspended nanoparticles from forming relatively large clusters or aggregates. This clustering, already documented for nano-TiO2 and other metal oxide nanoparticles (14-16) is consistent with the principles of colloidal chemistry and is expected to be strongly influenced by the ionic strength and pH of the aqueous environments in which nano-TiO2 is placed (15). One would expect this clustering to have important repercussions for practical uses of nano-TiO2 and for its health and environmental impact. At this point, however, very little information is available on the influence of ionic strength and pH on the extent and particularly on the kinetics of the aggregation of nano-TiO2. In this context, we present the first experimental attempt to both characterize the morphology and quantify the rate of formation of nano-TiO2 aggregates as a function of the ionic strength and pH of suspensions and of the nature of the electrolyte present (NaCl versus CaCl2). All the parameters considered fall within ranges likely to be encountered in nature, specifically in situations where nano-TiO2 enters into contact with surface soils.
Materials and Methods Introduction In the past decade, the range of applications of nano-TiO2 has expanded at an extremely fast pace (1, 2). Currently, * Corresponding author e-mail:
[email protected]. † Virginia Tech. ‡ Utah State University. § University of Minnesota. | Abertay University. 1354
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 5, 2009
Preparation and Characterization of Nano-TiO2. Nano-TiO2 was synthesized using a sol-gel method (17). The particles were diluted to obtain stock suspensions of known concentration (see Supporting Information (SI) for details). The phase composition and average size of the nano-TiO2 were characterized using X-ray diffraction (XRD) with Rietveld refinements (17). The point-of-zero-charge (PZC) was determined using electrophoretic mobility measurements. The nano-TiO2 particles were imaged using cryo-TEM on a Tecnai 10.1021/es802628n CCC: $40.75
2009 American Chemical Society
Published on Web 01/21/2009
12 Biotwin transmission electron microscope and a Gatan model 791 multiscan camera. The XRD, electrophoretic mobility, and cryo-TEM methods are explained in more detail in the SI Materials and Methods section and Figures S1 and S2. Suspensions with Different Ionic Strength and Cation Valence. By addition of ACS grade reagents and DI water, the nano-TiO2 stock suspension (80-83 mg L-1 TiO2 for NaCl experiments and 80 mg L-1 TiO2 for CaCl2 experiment) was adjusted to pH ∼4.5 with 0.005 M HCl and to an ionic strength of 0.001 M with NaCl and CaCl2. Aliquots (∼100 mL) of the resulting suspensions were sonicated for 1 h in an ultrasonic bath and allowed to cool to room temperature for at least 3 h prior to experiments. The suspensions were then filtered through a 0.45 µm filter to remove large aggregates. For the duration of the experiments, these stock suspensions retained a stable average aggregate size of 50-60 nm (SI Figures S3 and S7). Aliquots containing 25 mL of these suspensions were mixed with 25 mL of respectively 0.008 M, 0.016 M, 0.024 M, and 0.032 M NaCl solutions or a 0.008 M CaCl2 solution in plastic vials containing a Teflon-coated stir bar to reach ionic strengths of 0.0045 M, 0.0085 M, 0.0125 M, and 0.0165 M for the NaCl experiments and 0.0128 M for the CaCl2 experiment, respectively. The final nano-TiO2 concentration was approximately 40-42 mg L-1 for the NaCl experiment and 42 mg L-1 for the CaCl2 experiments. Because the pH of the suspensions was not buffered, there was some drift in the pH values in both the NaCl and CaCl2 samples. However, the pH values of all the samples and replicate suspensions (SI Figures S4, S5, and S8) of the two salts fell between 4.5 and 4.8 indicating that pH was not the dominating factor controlling the aggregation rate. In all cases, including 2 min of constant stirring, exactly 5 min elapsed between the moment the salt solutions were added to the stock nano-TiO2 suspension and the start of the DLS measurements. Suspensions with Different pH. The 83 mg L-1 nanoTiO2 suspensions were prepared in buffer solutions adjusted to pH 6.8, 7.5, and 8.2 by addition of 0.010 M HEPES (N2-hydroxyethylpiperazine-N-2-ethanesulphonic acid) (Sigma Chemical Co.), titrated with freshly prepared NaOH. The resulting ionic strength and average aggregate size of these suspensions depended on the amount of NaOH added to the buffer (SI Figure S11). The pH 5.8 suspension was prepared in DI water and adjusted to 0.001 M with NaCl. The suspensions were sonicated for 1 h in an ultrasonic bath and cooled to room temperature for at least 3 h prior to DLS measurements. A 25 mL aliquot of these suspensions was mixed with 25 mL of buffered NaCl solution to reach a final ionic strength of 0.0084-0.0099 M. The final nano-TiO2 concentration was 42 mg L-1. The DLS measurement procedure was the same as that for the ionic strength experiments. Dynamic Light Scattering (DLS) Measurements. The size of nano-TiO2 aggregates in various suspensions was measured using DLS on a Zetasizer Nano-ZS (Malvern Instruments). The CONTIN algorithm was used to convert intensity autocorrelation functions to intensity-weighted hydrodynamic aggregate diameter distributions on the basis of the Stokes-Einstein relationship for spherical particles. All DLS analyses were carried out at 25 °C. Preliminary tests showed that the final nanoparticle concentration of the suspensions (40-42 mg L-1) was adequate to avoid multiple scattering. Data points were collected approximately every 2 minsthe exact time was recorded by the instrumentsfor the first 35 min of DLS analysis. Time zero was defined as the time of the first DLS measurement. Subsequent measurements were taken at 2 h, 4 h, and 8 h for the first day and then every 24 h for the next four days or until aggregation had reached the micron range limit of DLS. To ensure a representative sample
FIGURE 1. Cryo-TEM image of 4.4-5.5 nm nano-TiO2. Scale bar represents 100 nm. for the aliquot subjected to DLS analysis, suspensions were agitated vigorously using a vortex mixer, for ∼10 s, 2 min prior to removing an aliquot for DLS analysis. Suspensions were stored at room temperature between DLS measurements. Suspension pH was measured at the end of the DLS measurements to avoid possible aggregation due to interaction with the pH electrode. The contour plots used to depict the DLS data (Figures 2-5 and SI Figures S6, S8, and S10) were generated using SigmaPlot 10.0 (Systat Software, Inc.), which uses simple interpolation between data points. The data were not smoothed prior to plotting.
Results and Discussion Particle Characterization. The quantitative phase composition of the nano-TiO2 was determined using Rietveld refinements and the results were characteristic of well-crystallized, nanosized anatase/brookite with no apparent amorphous material (SI Figure S1). X-ray diffraction results show that the sample is composed of anatase and brookite with no apparent amorphous material, and the quantitative phase composition, determined using the Rietveld method, is 63.3 ( 0.5% anatase and 36.7 ( 0.5% brookite. The average particle sizes, determined using Scherrer line broadening analysis, are 5.5 ( 0.2 nm for anatase and 4.4 ( 0.9 nm for brookite. The standard deviations of these sizes (0.2 nm for anatase and 0.9 nm for brookite) represent the measurement error associated with three replicate measurements on the same sample and do not reflect an underlying particle size distribution. Cryo-TEM images of the particles (Figure 1) corroborate the XRD size results to some extent, but the particle size distribution is broad. Electrophoretic mobility measurements (SI Figure S2) indicate that the nanoparticles have a point-of-zero-charge (PZC) in the vicinity of pH 6.8, in agreement with previous reports of the pHpzc for anatase nano-TiO2 of 6.3-6.5 (4), 6.85 (16), and 6.7 (18). Initial Aggregation of Suspensions. Even though only 5 min elapsed between the mixing of the salt solutions with the stable 50-60 nm nano-TiO2 aggregate suspensions and the first DLS observations, experimental results suggest that substantial subsequent aggregation occurs in the interim at all pHs and ionic strengths employed except for the lowest ionic strength of 0.0045 M NaCl at pH 4.5. Its starting size distribution is narrow, centered around 50 to 60 nm (Figure 2a). In suspensions at higher ionic strengths the mean VOL. 43, NO. 5, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1355
FIGURE 2. Aggregation of nano-TiO2 at pH 4.5 and different ionic strengths over a period of 5 days (120 h). The ionic strengths, adjusted with NaCl, are respectively (a) 0.0045 M and (b) 0.0085 M. Colors represent percent intensity of light scattered by aggregates.
FIGURE 3. Aggregation of nano-TiO2 at pH 4.5 and different ionic strengths over a period of 35 min. The ionic strengths, adjusted with NaCl, are respectively (a) 0.0125 M and (b) 0.0165 M. aggregate size is substantially larger and the size distribution is broader (Figures 2b and 3). At an ionic strength of 0.0165 M aggregate sizes extend from 20 to 1000 nm with a mean around 150 nm (Figure 3b). At higher pH, above the level for the PZC as well as in the suspensions with a CaCl2 electrolyte background, aggregates in suspension at the onset of the DLS measurements are much bigger with an average size in the micron range (Figures 4 and 5 and SI Figures S8 and S10). 1356
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 5, 2009
Aggregation Kinetics at pH below Point-of-Zero-Charge (pHpzc). The first series of experiments involved suspensions initially targeted to have a pH of 4.5. In some cases, after drifting, the final pH ends up in the vicinity of 4.8. Still, this is roughly two pH units below the pHpzc of the nano-TiO2 and the particle surfaces are therefore bearing a positive charge, compensated by the Cl- ions in solution. According to existing theories of colloidal stabilitysin this case, given the low surface charge density of nano-TiO2, the old DLVO theory, and the more accurate Sogami-Ise theory (19-21) lead to similar predictionssany increase in the ionic strength of the electrolyte solution in which the particles are suspended should compress the diffuse layer (DL) associated with the particles, decrease their zeta potential, diminish interparticle repulsion, and therefore promote aggregation. Observations of such an effect of ionic strength on aggregation at pH 4.5 has also been made for magnetite nanoparticles (22). At the lowest ionic strength (0.0045 M) evaluated in our experiments, DLS measurements (Figure 2a) do not show significant aggregation beyond that which occurred prior to DLS monitoring. The aggregate size distribution remains centered around 50-60 nm over the next few days. This is the same particle size range determined for the initial nanoTiO2 suspensions with an ionic strength of 0.001 M. At an ionic strength of 0.0085 M (Figure 2b) the average aggregate size increases throughout the first few hours of DLS monitoring to eventually reach a stable plateau with a mean aggregate size around 500 nm. This same leveling-off pattern is manifested at higher ionic strengths to such an extent that only DLS measurements in a much shorter time span, e.g., the first 35 min, can be reported due to TiO2 precipitation and DLS particle size limitations (Figure 3). At the highest ionic strength (0.0165 M) studied, aggregates reach the micron scale barely 15 min after the beginning of DLS observation. Thereafter, the aggregate size distribution in suspension becomes multimodal. A portion of the aggregates have sizes around 2-3 µm, which is close to the limit of detection for DLS. Some of the aggregates have a size around 0.5-1 µm and a few keep their initial size with a mean around 60 nm. This latter group seems to become negligible (
