Environ. Sci. Technol. 2010, 44, 6396–6402
Impact of Plastics on Fate and Transport of Organic Contaminants in Landfills J O V I T A M . S A Q U I N G , * ,† CARL D. SAQUING,‡ DETLEF R. U. KNAPPE,† AND MORTON A. BARLAZ† Department of Civil, Construction, and Environmental Engineering, Box 7908, North Carolina State University, Raleigh, North Carolina 27695-7908 and Department of Chemical and Biomolecular Engineering, Box 7905, North Carolina State University, Raleigh, North Carolina 27695-7905
Received April 19, 2010. Revised manuscript received June 26, 2010. Accepted June 29, 2010.
Factors controlling organic contaminant sorption to common plastics in municipal solid waste were identified. Consumer plastics [drinking water container, prescription drug bottle, soda bottle, disposable cold cup, computer casing, furniture foam, carpet, vinyl flooring, formica sheet] and model polymers [highdensity polyethylene (HDPE), medium-density polyethylene, lowdensity polyethylene, poly(vinyl chloride) (PVC)] were characterized by X-ray diffractometry, differential scanning calorimetry, and elemental analysis. The material characterization was used to interpret batch isotherm and kinetic data. Kp values describing toluene sorption to rubbery or “soft” polymers could be normalized by the amorphous polymer fraction (famorphous) but not by the organic carbon fraction (foc). Diffusion coefficients (D) describing the uptake rate of toluene by rubbery plastics (HDPE, drinking water container, prescription drug bottle) were similar (D ≈ 10-10 cm2/s), indicating that pure HDPE can be used as a model for rubbery plastics. Toluene diffusivity was similar among glassy or “hard” plastics (PVC, soda bottle, computer casing, disposable cold cup; D ≈ 10-12 cm2/ s) but lower than for rubbery plastics. Plastics in landfills are potential sinks of hydrophobic organic contaminants (HOCs) because of their higher affinity for HOCs compared to lignocellulosic materials and the slow desorption of HOCs from glassy plastics.
Introduction Plastic products and packaging materials are ubiquitous in our everyday lives, and the plastic content of discarded municipal solid waste (MSW) has increased from ∼0.5% to ∼17% between 1960 and 2007 (1). Much of this plastic ends up in landfills as it is estimated that approximately 54% of municipal solid waste (MSW) is disposed of in landfills (1). Due to differences in classification schemes, it is difficult to compare the plastic content of MSW generated in various countries, but available information suggests that MSW generated in Japan, Korea, Canada, and Mexico consists of * Corresponding author phone: (919) 760-0994; fax: (919) 5157908; e-mail:
[email protected]. † Department of Civil, Construction, and Environmental Engineering. ‡ Department of Chemical and Biomolecular Engineering. 6396
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
3-13% plastics (2). Furthermore, the increasing proportion of plastics in MSW has been associated with economic progress in developing Asian regions (3). Along with biopolymer composites [e.g., office paper (OP), newsprint (NP), food and yard wastes], synthetic polymers or plastics are an increasingly important component of the sorbent organic matter (OM) in landfills. The association of hydrophobic organic contaminants (HOCs) with sorbent OM retards contaminant transport and reduces HOC availability for chemical and biological degradation (4-6). Sorption affinity for alkylbenzenes was observed to be at least an order of magnitude higher in model rubbery [high-density polyethylene (HDPE)] and glassy [poly(vinyl chloride) (PVC)] plastics relative to MSW biopolymers (7). Likewise, sorption uptake of phenanthrene was at least a factor of 10 higher in model plastics [HDPE, polypropylene (PP), PVC] than in marine sediments (8). Contaminant partitioning to plastics is also important in the marine environment because plastic debris can serve as a transport medium for HOCs (e.g., polychlorinated biphenyls (PCBs), PAHs, and pesticides), some of which are ingested by marine life (9). The major types of plastics discarded in landfills by resin are HDPE (18%), low-density polyethylene (LDPE, 21%), PP (16%), PVC (6%), polystyrene (PS, 9%), and polyethylene terephthalate (PET, 11%) [Figure S1 of the Supporting Information (1)]. In terms of product-based categories, the distribution of discarded plastics is 35% durable goods, 23% nondurables (e.g., plates, trash bags), and 42% containers and packaging materials (1). The distribution of these product-based categories in each plastic resin is important as it provides insights on the relative thickness or diffusional length scale, which controls sorption and desorption kinetics (Figure S1, Supporting Information). For example, the thickness of HDPE in a film and a milk container is 0.02 and 1.85 mm, respectively. The physical and chemical sorbent properties of a plastic material are dependent on the resin type. The addition of plasticizers, additives, and fillers in plastic products, however, can modify the sorbent properties of a material (10). In previous (de)sorption studies of HOCs, pure HDPE was used as a model for rubbery or “soft” plastics while pure PVC was used as a model for glassy or “hard” plastics (7, 11, 12). However, the suitability of HDPE and PVC to represent rubbery and glassy consumer plastics, respectively, requires validation. Another limitation of previous studies is that HOC (de)sorption in model plastics was measured with neat compounds or at very high aqueous concentrations (13-16). However, the aqueous concentrations of HOCs detected in the environment (e.g., landfills, marine environments) are generally at trace levels (17) and are not likely to swell plastics or plasticize glassy materials (18). Because the plastic content and type of plastics can influence HOC release and longterm persistence in landfills, it is important to have a better understanding of factors controlling sorption to common MSW plastics. The objectives of this research were to (1) determine the rubbery and glassy behavior of representative consumer plastics and (2) determine and compare sorption equilibrium and kinetic parameters of representative consumer plastics and model polymers at dilute HOC aqueous concentrations.
Materials and Methods Plastic Materials. A drinking water container (HDPE), prescription drug bottle (PP), soda bottle (PET), disposable cold cup (PS), and PVC pipe were chosen to represent the major plastic resins present in MSW (1). In addition, plastic 10.1021/es101251p
2010 American Chemical Society
Published on Web 07/22/2010
computer casing [acrylonitrile butadiene styrene (ABS)], furniture foam [polyurethane (PU)], nylon carpet [polyamide 6 (PA6)], vinyl flooring, and a formica sheet [melamine formaldehyde (MF)] were selected to represent other plastic components, including those important in building and demolition debris. Computer casings were obtained from the North Carolina State University surplus warehouse. Building debris components were obtained from a hardware store. Pure forms of HDPE, medium-density polyethylene (MDPE), LDPE, and PVC (Catalog Nos. 42,799-3; 33,211-9; 428043, 18,958-8, Sigma-Aldrich) were tested for their ability to serve as model polymers representing plastic materials. Except as noted, all plastic materials were ground separately in a Wiley mill to pass a 0.25 mm screen. For the drinking water container and prescription drug bottle, a 0.841 mm screen was used to prevent clogging of smaller screens. Container caps and labels were removed prior to grinding. The vinyl flooring and formica sheet were dried at 100-105 °C after grinding and stored in vacuum desiccators until use. The other ground materials were dried at room temperature and stored in vacuum desiccators to maintain prior thermal history (19, 20). HDPE, MDPE, LDPE, and PVC polymers were procured in pellet or powder form and used without further processing. Material Characterization. The drinking water container, prescription drug bottle, soda bottle, and disposable cold cup were characterized by X-ray diffractometry (XRD) using a Philips X’Pert PRO MRD HR diffractometer over an angular range of 5-85° (2θ) in intervals of 0.01° (2θ) and with a step duration of 1 s. Ground plastics were dispersed and attached to the sample holder with double-sided tape. For some pure polymers (PP, PET), published XRD data were used (21). C, H, and N analyses were performed with a CHN analyzer (Perkin-Elmer PE 2400 CHN Elemental Analyzer, PerkinElmer Corp.), and O and Cl contents were calculated based on the chemical composition of the monomers. As carpet backing and vinyl flooring commonly contain calcium carbonate filler, these materials were acid washed to eliminate inorganic carbon prior to analysis (22). Ash content was measured by combustion at 550 °C for 2 h. The glass transition (Tg) and melting (Tm) temperatures and % crystallinity (Xc) were determined by differential scanning calorimetry (DSC) on a TA Instruments DSC Q2000 equipped with thermal analysis data acquisition software. Details on Tg, Tm, and Xc determination are provided in the Supporting Information. The morphology and particle size of the plastic materials were observed by scanning electron microscopy (SEM) (Hitachi S-3200 and an FEI XL30 SEM, both operating at an accelerating voltage of 5 kV). The plastic materials were coated by a K-550X sputter coater with Au/Pd ≈ 100 Å thick to reduce electron charging. The particle size of plastic samples was obtained by direct measurement of the particle diameter or length from several SEM images. The average particle size was determined from the minimum diameter or length of 30-80 particles. Model HOCs. Toluene was selected as the model HOC because it is frequently detected in landfill leachates, sorbs to refuse, and biodegrades in decomposing refuse (23). Moreover, its relatively small molecular size (5.6 Å, ChemSketch) makes it possible to achieve sorption equilibrium at reasonable times (e.g., model LDPE ≈ model MDPE > drinking water container (HDPE) > model HDPE (p ) 0.05). With the exception of the soda bottle, glassy single-resin materials exhibited higher Kp values than the rubbery plastics (Table 3). The crystallinity of the soda bottle is higher than that of other glassy polymers and likely contributed to its relatively low Kp. Solubility parameters for the different plastic materials (Table 2) did not differ greatly from that of toluene (δHa ) 18.2, δd ) 18.0, δp ) 1.4, δh ) 2.0 MPa0.5) (39). The rubbery plastics except for the furniture foam (PU) are nonpolar (Table 2), suggesting that nonspecific dispersion interactions were the dominant sorbate/sorbent interaction. Although the molecular interactions between the glassy plastics and toluene possibly included some dipole and H-bonding interactions (compare polar and H-bonding terms of Hansen solubility parameter in Table 2 with δp ) 1.4 and δh ) 2.0 MPa0.5 for toluene), the sorbate/sorbent interactions were dominated by dispersion (compare dispersion term of Hansen solubility parameter in Table 2 with δd ) 18.0 MPa0.5 for toluene) (39). Thus, the large difference in affinity for toluene between glassy and rubbery plastics studied here cannot be attributed to sorbent/sorbate compatibility or specific sorbent/sorbate interactions alone. Other HOCs such as bisphenol A (δHa ) 24.4, δd ) 19.2, δp ) 5.9, δh ) 13.8 MPa0.5) and PCE (δHa ) 19.2, δd ) 18.3, δp ) 5.7, δh ) 0.0 MPa0.5) that have the potential for dipole and H-bonding interactions will behave differently from toluene (39). For example, sorption uptake tests with bisphenol A as the sorbate showed considerable sorption affinity for furniture foam (PU) and the disposable cold cup (PS) after a week of contact time, while the model HDPE did not measurably sorb bisphenol A even after 53 days of contact (Table S2, Supporting Information). The Hansen solubility parameter for bisphenol A (δHa ) 24.4 MPa0.5) is closer to that of PU (δHa ) 21.2-24.9 MPa0.5) and PS (δHa ) 19.3-23.4 MPa0.5) than that of HDPE (δHa ) 18.1 MPa0.5) (39). Given that bisphenol A is polarizable and capable of H-bonding interactions, it was more compatible with PU and PS than with nonpolar HDPE. Thus, both sorbate and sorbent properties are important in understanding sorption affinity of plastics for HOCs. As shown in Table 3, normalization of Kp by the amorphous fraction reduced the variation among the semicrystalline plastics (PE, PP, PET). This result indicates that Xc is an important property for the sorption uptake of HOCs by plastics. The Kp/famorphous values for toluene are statistically similar for the drinking water container (HDPE), prescription drug bottle (PP), soda bottle (PET), model MDPE, and model LDPE (p ) 0.05). The difference in Kp/famorphous values for pure HDPE and the drinking water container was statistically significant (p ) 0.05) but small. Normalization of Kp by foc was not meaningful (Table 3), suggesting that unlike in soils and sediments, foc is not a controlling factor for HOC sorption to plastics. (De)Sorption Kinetics of HOCs in Plastics. Polymer diffusion coefficients (D) were used to compare HOC (de)sorption rates among different plastics. As shown in Table 4, the diffusivity of toluene is rapid in rubbery plastics (10-10 cm2/s) but slow in glassy plastics (10-12 cm2/s), a result that is consistent with the characteristics of rubbery and glassy polymers. Because of their rigid structure, glassy polymers have a much reduced relaxation speed relative to rubbery polymers, resulting in slow diffusion of the penetrating HOC. In this study, the Drubbery/Dglassy ratio was 102, which is in agreement with literature reports that range from 102 to 108 for solute molecules in the size range of toluene (5-7 Å diameter) (18, 43). The diffusivities of toluene in different plastic materials were measured at dilute aqueous concentrations (i.e.,
