Article pubs.acs.org/est
Nanomaterial Transformation and Association with Fresh and Freeze-Dried Wastewater Activated Sludge: Implications for Testing Protocol and Environmental Fate Mehlika A. Kiser,*,† David A. Ladner,‡ Kiril D. Hristovski,§ and Paul K. Westerhoff† †
Civil, Environmental, and Sustainable Engineering, Arizona State University, Tempe, Arizona 85287-5306, United States Environmental Engineering and Earth Sciences, Clemson University, Anderson, South Carolina 29625-6510, United States § College of Technology and Innovation, Arizona State University−Polytechnic Campus, Mesa, Arizona 85212, United States ‡
S Supporting Information *
ABSTRACT: Engineered nanomaterials (ENMs) are an emerging class of contaminants entering wastewater treatment plants (WWTPs), and standardized testing protocols are needed by industry and regulators to assess the potential removal of ENMs during wastewater treatment. A United States Environmental Protection Agency (USEPA) standard method (OPPTS 835.1110) for estimating soluble pollutant removal during wastewater treatment using freeze-dried, heat-treated (FDH) activated sludge (AS) has been recently proposed for predicting ENM fate in WWTPs. This study is the first to evaluate the use of FDH AS in batch experiments for quantifying ENM removal from wastewater. While soluble pollutants sorbed equally to fresh and FDH AS, fullerene, silver, gold, and polystyrene nanoparticles’ removals with FDH AS were approximately 60−100% less than their removals with fresh AS. Unlike fresh AS, FDH AS had a high concentration of proteins and other soluble organics in the liquid phase, an indication of bacterial membrane disintegration due to freeze-drying and heat exposure. This cellular matter stabilized ENMs such that they were poorly removed by FDH AS. Therefore, FDH AS is not a suitable sorbent for estimating nanoparticle removal in WWTPs, whereas fresh AS has been shown to reasonably predict full-scale performance for titanium removal. This study indicates that natural or engineered processes (e.g., anaerobic digestion, biosolids decomposition in soils) that result in cellular degradation and matrices rich in surfactant-like materials (natural organic matter, proteins, phospholipids, etc.) may transform nanoparticle surfaces and significantly alter their fate in the environment.
■
INTRODUCTION Conventional wastewater treatment plants (WWTPs) and membrane bioreactor treatment plants use activated sludge to remove nutrients, metals, organic chemicals, pathogens, and suspended particles from wastewater. Activated sludge is produced as a mixed population of microorganisms, including various strains of bacteria and their extracellular products, rotifers, protozoa, and fungi, converts organic matter into cellular material.1 If a pollutant readily adsorbs to or otherwise associates with activated sludge, it will be removed from the plant during primary or secondary solids separation. Approximately one-half of biosolids generated by WWTPs in the U.S. are applied to agricultural land as soil conditioner, and the remaining fraction is either landfilled or incinerated.2 If a pollutant has a low affinity for activated sludge and is not degraded or volatilized during treatment, it will remain mostly in the aqueous phase and be released with treated effluent into surface water. Quantifying a pollutant’s affinity for activated sludge is an essential step toward predicting its fate in the environment and assessing exposure risks. One of the most fundamental methods for quantifying pollutant distribution between the solid and liquid phases of activated sludge is the batch sorption isotherm experiment. The © 2012 American Chemical Society
United States Environmental Protection Agency (USEPA) provided a standard method for testing soluble pollutant sorption to activated sludge with the publication of the OPPTS 835.1110 Activated Sludge Sorption Isotherm test guideline.3 The experimental method outlined in this guideline has become standard industry and research practice for predicting chemical removal from wastewater during biological treatment in WWTPs.2,4−8 OPPTS 835.1110 calls for the use of freezedried and heat-treated (FDH) activated sludge as sorbent. Unlike fresh activated sludge, which must be collected, processed, and used daily, FDH activated sludge can be stored for several months, allowing for a convenient and uniform supply of sorbent for batch experiments. As the production of consumer products containing engineered nanomaterials (ENMs) grows, nanomaterials are entering WWTPs in increasing amounts and have been detected in WWTP biosolids and effluent.9−11 Studies have shown that ENMs Special Issue: Transformations of Nanoparticles in the Environment Received: January 27, 2012 Accepted: February 9, 2012 Published: February 9, 2012 7046
dx.doi.org/10.1021/es300339x | Environ. Sci. Technol. 2012, 46, 7046−7053
Environmental Science & Technology
Article
(NanoComposix), carboxylate-, and sulfate-functionalized yellowgreen fluorescent microspheres (Car-PS and Sulf-PS, respectively) (FluoSpheres, Invitrogen; Eugene, OR), and carboxylate-functionalized silver (Car-Ag) (Vive Nano; Toronto, ON, Canada) were purchased. Dr. Mark Wiesner of Duke University’s Center for Environmental Implications of Nanotechnology (CEINT) provided three different nanosilver suspensions: polyvinylpyrrolidone-coated (PVP-Ag) (NanoAmor, Los Alamos, NM), citrate-coated (Cit-Ag) (prepared at Duke University), and gum-arabic-coated (GA-Ag) (prepared at Duke University). Characterization data of nanoparticles used in this study are shown in Supporting Information (SI) Table SI.1. Activated Sludge Collection and Preparation. Return activated sludge was collected from a municipal conventional activated sludge wastewater treatment plant in central Arizona that treats 90% domestic and 10% industrial wastewater. The plant’s aeration basins have solid retention times of 15 days and remove nitrogen through nitrification and denitrification. The collected sludge was kept at 4 °C during transport and storage in the laboratory. Within 48 h of collection, the sludge was prepared for experimentation. To prepare FDH activated sludge, the procedure detailed in OPPTS 835.1110 was followed. Briefly, activated sludge was rinsed three times with ultrapure water. Rinsing was accomplished by centrifuging the activated sludge at RCF = 2000g for 5 min (IEC Multi, Thermo IEC; Waltham, MA) and then decanting the supernatant. Rinsed sludge was freeze-dried following manufacturer instructions (FreeZone 6 Liter, Labconco; Kansas City, MO), passed through a No. 30 (600 μm aperture) sieve (SoilTest, Inc.; Evanston, IL), and finally heat dried at 104 °C for 3 h, 12 h, 24 h, 3 d, 7 d, or 14 d. The resulting powder-like FDH activated sludge was cooled to room temperature in a desiccator. Freezedried activated sludge was also prepared without the final heatdrying step (FD activated sludge). The day before an experiment, activated sludge suspensions were prepared by mixing desiccated activated sludge powder in buffered (1 mM NaHCO3) ultrapure water and storing the suspension overnight (4 °C) to rehydrate the activated sludge. Rinsed FDH activated sludge was prepared using a portion of FDH activated sludge suspension that had been stored overnight. This portion of FDH activated sludge was rinsed nine times with buffered water following the same rinsing protocol as described above. Fresh activated sludge suspensions were prepared by rinsing activated sludge three times with buffered water as described above and then resuspending the rinsed sludge in buffered water. Fresh activated sludge was stored at 4 °C for a maximum of 24 h before being used in experiments. Batch Experiments. To compare the removal of our sorbates in fresh and FDH activated sludge, batch experiments were conducted. Glass vials containing fresh or FDH activated sludge suspension and buffer solution (1 mM NaHCO3) were spiked with a chemical solution or a nanoparticle suspension. Sample concentrations of sorbents and sorbates are listed in SI Table SI.2. Concentrations of sorbent (800 mg/L TSS) and sorbates (∼30 μg/L to 6 mg/L for soluble compounds; ∼0.1 to 3 mg/L for nanoparticles) were chosen such that differences could be distinguished in the removal of the various sorbate types in fresh and FDH activated sludge. Samples were agitated for 3 h on a platform shaker (C1, New Brunswick Scientific; Edison, NJ) and then stood upright for 2 h to simulate mixing in aeration basins and sedimentation in secondary clarifiers, respectively.22,23 After sedimentation, supernatant (the top liquid portion of the sample after sedimentation) was collected
associate with solids (i.e., microorganisms) in activated sludge, and that ENM removal during wastewater treatment is controlled by the extent of ENM association with the solid phase of activated sludge.9,12−14 A standardized method of quantifying nanomaterial association with the solid phase, or conversely, removal from the liquid phase, in activated sludge does not currently exist. The USEPA recently published an interim technical guide for evaluating the environmental fate of nanomaterials using OPPTS 835.1110 with the hope that “experienced scientists will find it helpful and will contribute to the further development and validation of this approach.”15 However, the OPPTS guideline for testing soluble compounds was neither developed nor validated for use with nanomaterials. The goal of this study was to evaluate the use of FDH activated sludge, a standard sorbent in batch experiments for chemicals, for quantifying ENM removal from wastewater. Nanomaterial removal in fresh and FDH activated sludge was compared by conducting batch experiments with fresh activated sludge produced following protocol we developed in previous work14 and with FDH activated sludge prepared following the method outlined in OPPTS 835.1110. Furthermore, the effect of FDH activated sludge processing steps (freeze-drying, heat treatment) on ENM removal from wastewater was studied. The degradation of biomass in activated sludge and the consequent release of soluble biosurfactants were investigated as a mechanism for transforming ENM surface properties. This study contributes to the development of a reliable standard method for evaluating nanoparticle removal from wastewater and has important implications for the transformation and fate of nanomaterials in the environment.
■
MATERIALS AND METHODS Chemical Solutions and Nanomaterial Suspensions. Three compounds were chosen to serve as model soluble contaminants to test in batch experiments. Methylene blue (MB) (Fisher Scientific; Pittsburgh, PA) is an organic, cationic dye that readily stains bacteria because of its affinity for negatively charged cellular constituents such as acidic polysaccharides and nucleic acids.16 Used as a dye in material industries and as a biological stain, MB has also been widely employed for several decades as a model sorbate in adsorption studies.17,18 17αethinylestradiol (EE2) (Sigma-Aldrich; St. Louis, MO), a synthetic steroid estrogen and the active ingredient in contraceptive pills, is only partially removed in conventional activated sludge WWTPs and is implicated in the endocrine disruption of aquatic organisms.4,19 Ionic silver (Ag+) from silver nitrate (AgNO3) (Sigma-Aldrich; St. Louis, MO) is antimicrobial and used as a biological stain for scanning electron microscopy and protein demonstration in PAGE gels. Ag+ readily binds to thiol groups in membrane proteins and disrupts protein function.20 Silver has been quantified along the treatment train of WWTPs and found to strongly associate with solids, resulting in more than 94% removal of silver over the course of treatment.21 Nine nanomaterial suspensions were used to compare removal in fresh and FDH activated sludge and expose relationships between nanomaterial properties and removal. Nonfunctionalized fullerene (aq-nC60) suspension was prepared by magnetically stirring 99.9% C60 powder (MER Corporation; Tucson, AZ) in ultrapure water without exposure to light for several months and then filtering the golden-brown suspension through 0.7 μm glass-fiber filters (Whatman; Maidstone, UK). Suspensions of tannic-acid-capped nanogold (TA-Au) in a range of sizes (NanoComposix; San Diego, CA), PVP-coated gold (PVP-Au) 7047
dx.doi.org/10.1021/es300339x | Environ. Sci. Technol. 2012, 46, 7046−7053
Environmental Science & Technology
Article
1 mL/min with a mobile phase of 50% acetonitrile and 50% toluene. The method detection limit and percent recovery for aq-nC60 in secondary effluent using LLE with UV/vis spectroscopy has been found to be 3−4 μg/L and 93 ± 2%, respectively.25 Nanoparticle hydrodynamic diameters were measured from five one-minute runs using phase analysis light scattering (PALS) (90 Plus, Brookhaven Instruments Corp.; Holtsville, NY). Zeta potentials were determined on a dynamic light scattering (DLS) particle sizer (NICOMP 380 ZLS, Particle Sizing Systems; Santa Barbara, CA). Activated sludge stocks were characterized by measurement of total suspended solids (TSS)26 and chemical oxygen demand (COD) (Hach, Loveland, CO). Filtered (0.45-μm polysulfone) activated sludge supernatants were also characterized by measuring COD, UV−vis absorbance at 280 nm (DR 5000, Hach), surface tension (Tensiomate 21, Fisher Scientific; Waltham, MA), and protein concentration (BCA Protein Assay Kit, Thermo Scientific; Waltham, MA). Solid samples were imaged after exposure to fluorescent Sulf-PS microspheres using both bright-field and epifluorescence microscopy, as described by Kiser et al.14
from each sample and analyzed. To quantify the effect of the liquid phase of FDH activated sludge on nanoparticle removal, a batch experiment following the same procedure as described above was conducted using fresh, FDH, and rinsed FDH activated sludge. To quantify reversibility of nanoparticle association with the solid phase in activated sludge, the volume of liquid remaining over the settled activated sludge was discarded and replaced with the same volume of fresh buffer solution. These samples were agitated for 3 h and settled for 2 h, and then supernatants were collected and analyzed. For all experiments, controls (without NPs; without activated sludge) were made and analyzed alongside samples. At least 15% of samples were conducted in replicate. Assuming that volatilization and biodegradation were negligible, the percent removals of the ENMs were determined from the difference between the measured ENM concentrations in the supernatants of controls without activated sludge and the ENM concentrations in the supernatants of samples with both activated sludge and ENMs. Analytical Methods. Quantification of EE2 was accomplished using HPLC coupled with fluorescence detection at excitation and emission wavelengths of 280 and 310 nm, respectively, which we described elsewhere.24 The method detection limit of EE2 using HPLC with fluorescence detection was found to be 283 ng/L.24 Methylene blue was analyzed by UVvis absorbance at 664 nm (DR 5000, Hach; Loveland, CO). The detection limit of the UV-vis spectrophotometer is 0.01 units of absorbance. Ionic silver was measured using inductively coupled plasma optical emission spectroscopy (ICP-OES) (iCAP 6000 Series, Thermo Scientific; Cambridge, UK). Samples with nanosilver and nanogold were digested with nitric acid and aqua regia, respectively, in order to transform nanoparticles into ions. Silver and gold concentrations in the digested samples (Car-, Cit-, GA-, and PVP-Ag; TA- and PVPAu) were measured using ICP-OES, which had a detection limit for gold and silver of at least 1 μg/L. All of the silver- or gold-containing samples that were measured, including both no-activated-sludge controls and supernatants after exposure to activated sludge, had metal concentrations of at least 100 μg/L, which is approximately 100 times greater than the detection limit. Concentrations of yellow-green fluorescent Car- and SulfPS NPs were indirectly determined by measuring sample fluorescence (excitation/emission maxima = 505/515 nm) (LS 50B Luminescence Spectrometer, PerkinElmer; Waltham, MA). Concentrations at least 100 times lower than our initial sample concentrations could be accurately determined on the luminescence spectrometer with a variability of less than 5% in the measurement. For aq-nC60, samples were prepared for analysis using liquid−liquid extraction (LLE). The optimal LLE condition was selected as follows: 10 mL sample, 10 mL toluene, and 25 mL glacial acetic acid (GAA). After 2 h of agitation, vials were stood upright for 60 min, during which time toluene separated from the rest of the mixture and formed a layer on top. 0.5 mL of toluene of each LLE sample was collected in an HPLC vial and then evaporated under a nitrogen stream. After evaporation to dryness, the sample was reconstituted with 0.5 mL of toluene and then sonicated in an ultrasonication bath (100W) for 5 min. The vial was filled with 0.5 mL of acetonitrile for HPLC analysis. HPLC analysis using a wavelength of 336 nm was performed on a Water Alliance Separate Module and a UV-vis detector (Waters 2475, 2695, and 2996; Milford, MA). The analytical column was a Discovery C18, 150 mm × 4.6 mm, packed with 5 μm particles (Supelco, U.S.). The chromatographic separation was performed at a constant flow rate of
■
RESULTS AND DISCUSSION Comparison of Nanomaterial Removal with Fresh and FDH Activated Sludge. SI Figure SI.1 shows removal percentages of MB, EE2, and Ag+ with 800 mg/L total suspended solids (TSS) of fresh and FDH activated sludge. Assuming biodegradation and volatilization are negligible, the percent of sorbate removed from the liquid phase is the same as the percent of sorbate associated with the solid phase of activated sludge. EE2 removal with the two activated sludge types were similar: 48 ± 4% of EE2 was removed with fresh activated sludge, and 43 ± 14% was removed with FDH activated sludge. Ag+ removal with FDH activated sludge also reasonably represented the extent of sorption to fresh activated sludge. 95 ± 1% and 91 ± 1% of Ag+ was removed with fresh and FDH activated sludge, respectively. Levels of EE2 and Ag+ sorption to activated sludge are similar to values reported in other studies.4,21 MB sorbed about 15% less to FDH activated sludge than to fresh activated sludge. Overall, sorption percentages of the model soluble contaminants were comparable between fresh and FDH activated sludge. Figure 1 shows removal results for nine different types of ENMs by fresh and FDH activated sludge. For fresh activated sludge, two groups of nanomaterials are evident. Nanomaterial removals in Group I (CIT-Ag, PVP-Ag, PVP-Au, Car-Ag, and GA-Ag) ranged from 39 to 62%. Nanomaterials in Group II (TA-Au, Car-PS, Sulf-PS, and aq-nC60) had higher removals (92 to 94%) by fresh activated sludge. All of the nanomaterial types in our study had lower association with solids in FDH activated sludge than in fresh activated sludge. No removal of Cit-Ag, PVP-Ag, PVP-Au, or TA-Au was detected with FDH activated sludge. For those nanomaterials that did associate with solids in FDH activated sludge, removals ranged from 7 to 24%, which was less than their removal by fresh activated sludge. The range of removals by fresh activated sludge is attributed to differences in ENM materials and surface properties. While a trend was not discernible between percent removal and the zeta potentials or sizes of the different types of nanomaterials used in this study, a batch experiment using only one type of nanoparticle (TA-Au) ranging from 5 to 100 nm in diameter showed that size, density, and number concentrations are factors in 7048
dx.doi.org/10.1021/es300339x | Environ. Sci. Technol. 2012, 46, 7046−7053
Environmental Science & Technology
Article
Figure 1. Percent removal of nanoparticles after exposure to 800 mg/L TSS fresh or FDH-14d activated sludge; error bars represent ± standard deviation of duplicate samples.
removal (SI Figure SI.2). Thus, removal appears to be a function of nanoparticle material, surface functionalization, size, and zeta potential. Looking at only one of these factors, such as zeta potential or size, may not be sufficient for making a priori estimates of the removal of ENMs in fresh activated sludge. Group I includes surface functionalizations (SI Figure SI.3) and materials resulting in lower activated sludge affinities. Carboxylates are low molecular weight polar molecules often used as surface-functionalization agents to make particles more stable in water. PVP is a hydrophilic synthetic polymer that has been used as a coating to make membrane filters hydrophilic and as a carrier matrix to improve dissolution of hydrophobic drugs.27 Gum arabic, a natural polysaccharide with hydroxyl and carboxyl functional groups, is used to stabilize nanoparticles in aqueous solutions; the hydrophilicity of polysaccharides has been shown to increase nanoparticle residence time in blood and inhibit particle coating by plasma components.28 ENMs in Group II have properties that favor association with fresh activated sludge. Tannic acid (TA) (SI Figure SI.3) is a hydrolyzable tannin. Tannins, high-molecular-weight polyphenols produced by plants, strongly bind to proteins and other macromolecules to form insoluble complexes.29 Numerous studies have found that tannins bind to bacterial cell membranes and extracellular structures such as fimbriae.29,30 Though stabilized in water through long-term stirring, aq-nC 60 aggregates are relatively hydrophobic31 and have been shown to amass around bacterial cells.14,32 After simulating the translocation of C60 across a lipid bilayer, Qiao et al.33 suggest that C60 sorption into the lipid bilayer is driven by hydrophobic interactions between C60 and the lipid tails of the bilayer. The sulfate-functionalized polystyrene nanoparticles used in this study are hydrophobic. According to the manufacturer’s description, Sulf-PS nanoparticles will passively and nearly irreversibly adsorb almost any type of protein. Interestingly, carboxylate-modified polystyrene nanoparticles were removed to the same extent as Sulf-PS nanoparticles and to a greater degree than carboxylate-modified silver nanoparticles in fresh activated sludge. Car-PS and Sulf-PS nanoparticles have the same particle material (hydrophobic polystyrene), while Car-PS and Car-Ag nanoparticles have the same zeta potential and surface groups (carboxylate), though the degree of surface functionalization of these two nanoparticle types may be different. Without further study, conclusions cannot be made about which of these factors controlled removal. Fresh and FDH activated sludge were qualitatively compared using bright-field and epifluorescence microscopy (Figure 2).
Figure 2. Bright-field and epifluorescence images of 800 mg/L TSS fresh (left-hand column) and FDH-14d (right-hand column) activated sludge. Epifluorescence images in the bottom row correspond to the bright-field images in the preceding row. Activated sludge was exposed to 5 mg/L of 20 nm sulfate-functionalized polystyrene nanoparticles.
Well-defined components of activated sludge are visible in fresh activated sludge samples, such as spheres and filaments of different sizes and densities. In contrast, FDH activated sludge lacks distinct features and instead appears as thick, amorphous clumps of matter. The epifluorescence images show the fresh activated sludge sample as noticeably more fluorescent than the FDH sample, indicating greater association of fluorescent SulfPS nanoparticles with fresh activated sludge. Although fluorescence can be seen in the FDH sample, this is due to activated sludge autofluorescence from light exposure; an activatedsludge-only control (without exposure to nanoparticles) showed similar background color intensity as the FDH sample. The results of our batch sorption experiments and imaging indicate 7049
dx.doi.org/10.1021/es300339x | Environ. Sci. Technol. 2012, 46, 7046−7053
Environmental Science & Technology
Article
Figure 3. Effect of activated sludge processing (freeze-drying and heat drying) on nanoparticle removal. Notations within the FDH columns indicate duration of heat drying. Columns for each NP type are ordered from least to greatest degree of processing. Error bars represent ± standard deviation of duplicate samples.
FDH activated sludge contained 59 ± 0.7 μg protein/mL and 185 mg COD/L, whereas supernatant of the same concentration of fresh activated sludge only had 1.6 ± 0.4 μg protein/mL and 22 mg/L COD, evidence that freeze-drying and heat treatment result in the release of proteins into the liquid phase of FDH activated sludge suspensions. Nanomaterial Transformation by Biosurfactants. To test the effect of proteins and other cellular components in solution on nanomaterial removal, a batch experiment was conducted using Car-Ag nanoparticles and fresh, FDH, and rinsed FDH activated sludge suspensions. We verified that the rinsed FDH activated sludge had much less cellular material in solution than FDH activated sludge by measuring the absorbance and COD of filtered (0.45 μm) supernatant of each type of sorbent (800 mg/L TSS of activated sludge). Proteins in solution absorb ultraviolet light with absorbance maxima at 280 and 200 nm, and the relationship of absorbance to protein concentration is linear.37 The absorbances at 280 nm of the filtered supernatants of fresh, rinsed FDH, and FDH activated sludge were 0.017, 0.028, and 0.235. Furthermore, both fresh and rinsed FDH activated sludge had COD values under the instrument measuring range, whereas FDH activated sludge had 76 ± 14 mg/L COD. Car-Ag nanoparticle removal in FDH activated sludge was 36.5 ± 0.9% less than Car-Ag removal in fresh activated sludge. However, Car-Ag removal in rinsed FDH activated sludge was only 7.5 ± 0.7% less than Car-Ag removal in fresh activated sludge. The transformative and stabilizing effect of the cellular material in solution is demonstrated by the facts that (1) Car-Ag removal was similar (
