Environ. Sci. Technol. 2004, 38, 2657-2663
Factors Affecting Linear Alkylbenzene Sulfonates Removal in Subsurface Flow Constructed Wetlands Y U M I N G H U A N G , †,§ A N A L A T O R R E , † D A M I AÅ B A R C E L O Ä , † J O A N G A R C IÄ A , ‡ PAULA AGUIRRE,‡ RAFAEL MUJERIEGO,‡ AND J O S E P M . B A Y O N A * ,† Department of Environmental Chemistry, IIQAB-CSIC, Jordi Girona 18-26, E-08034 Barcelona, Spain, and Environmental Engineering Division, Department of Hydaulics, Coastal and Environmental Engineering, School of Civil Engineering, Technical University of Catalonia, Jordi Girona 1-3, E-08034 Barcelona, Spain
The behavior of linear alkylbenzene sulfonate (LAS) and sulfophenyl carboxylate (SPC) biointermediates in a pilot subsurface flow constructed wetland (SFCW) is reported for the first time. The effects of wetland configuration and operation on their treatment efficiency were investigated. The pilot SFCW constituted by eight beds of 55 m2 with different aspect ratios (1 × 1; 1.5 × 1; 2 × 1; 2.5 × 1), two water depths (i.e., 0.47 and 0.27 cm) at 5 cm below surface and two medium sizes (i.e., D60 ) 10 mm and 3.5 mm) planted with Phragmites sp. That SFCW pilot treats urban wastewater (i.e., 200 inhabitants) and was operated at four hydraulic loading rates (HLRs) (20, 27, 36, and 45 mm d-1). Influent and effluent sampling was carried out from May 2001 to January 2002 with a weekly pattern. Main results were as follows: (i) water depth has a major influence on the performance of SFCW for the LAS removal, and HLR shows significant effect on SPC evolution; (ii) water temperature has a significant effect on the LAS evolution; (iii) biodegradation of LAS and SPC can occur under sulfate-reducing environment and mixed conditions (i.e., sulfate-reducing and denitrification), but aerobic respiration cannot be excluded; and (iv) C13 LAS homologues were generally removed in higher extent than the shorter alkyl chain counterparts. In the most appropriate conditions, LAS and SPC can be biodegraded up to 71% and 11%, respectively, in the pilot SFCW evaluated.
Introduction Linear alkylbenzene sulfonates (LAS) are the most widely used synthetic anionic surfactants. They account for 28% of the total production (ca. 3-4 million t/yr) of synthetic surfactants in Western Europe, Japan, and the United States (1, 2). Due to its high-volume use in laundry and cleaning * Corresponding author phone: +34-93-4006100; fax: +34-932045904; e-mail:
[email protected]. † IIQAB-CSIC. § Present address: Institute of Environmental Chemistry, Southwest Normal University, Chongqing. 400715, People’s Republic of China. ‡ Technical University of Catalonia. 10.1021/es034821q CCC: $27.50 Published on Web 03/20/2004
2004 American Chemical Society
products, LAS is a ubiquitous water contaminant (3-5). The U.S. Geological Survey identified LAS in all bottom sediments in the Mississippi River at concentrations ranging from 0.01 to 20 mg kg-1 and in 21% of water samples at concentrations ranging from 0.1 to 28.2 µg L-1 (3). Two recent surveys indicated that drinking waters have detectable LAS and its main degradation product, sulfophenyl carboxylates (SPC), in Europe and Brazil (4, 5). It is widely accepted that LAS can be biodegraded under aerobic conditions in many environmental compartments (6). The generally accepted LAS biodegradation pathway begins with the ω-oxidation of the terminal methyl of the alkyl chain and continues with the shorting of the alkyl chain by successive β-oxidations (two carbon atoms) generating various SPCs (7). Some authors consider the ω-oxidation to be the phase that controls the rate of the entire LAS biodegradation process, particularly in systems with low oxygen (7). SPC with longer carboxylic chains are rapidly degraded to shorter chain SPCs. Despite the use of constructed wetland technology to treat wastewater (8) and the fact that nonpoint sources of pesticides in surface waters (9) is well-documented, little information exists regarding to the feasibility of wetland-based systems to remove LAS and its SPC biointermediates from wastewaters. Inaba (10) reported a 20-kg annual mass removal of LAS in a surface flow wetland system (planted TyphaPhragmites) receiving wastewater from a residential area (ca. 100 inhabitants). The inlet concentration (totaled about 5 mg L-1) was reduced by over 90% in summer and over 40% in winter at an average HLR of approximately 50 mm d-1. Barber et al. (11) reported that LAS removal depended on the hydraulic retention time (HRT) and the biogeochemical environment in various wetland configurations. However, at longer HRT, geochemical conditions could change from aerobic to anaerobic, which in turn influenced the rate of LAS biodegradation. In fact, it was found that there was little attenuation of LAS in wetland systems with low oxygen concentrations whereas LAS removal is nearly complete in oxygenated systems (11), consistent with the aerobic biodegradation being the primary LAS removal mechanism (12). Although some operational factors such as HRT or HLR in the previously mentioned wetlands systems were tested, a detailed study on the effects of wetland configuration and operation on efficiency of LAS and its SPC biodegradation intermediates treatment has not been investigated. The goal of the present study was to investigate the behavior of LAS and its SPC biodegradation intermediates in a subsurface flow constructed wetland (SFCW) system and the effects of wetland configuration and operation on efficiency of treatment. The specific objectives were (i) to evaluate the effect of HLR, aspect ratio, water depth, and medium size on the evolution of LAS and its SPC biodegradation intermediates in constructed wetlands; (ii) to explore the effects of physicochemical characteristics [dissolved oxygen (DO), redox potential, temperature, and concentrations of electron acceptors such as sulfate and nitrate] of the water in wetlands on the fate of LAS and its SPC biodegradation intermediates; and (iii) to model the evolution of LAS in the present SFCW system.
Materials and Methods Site Study. Experiments were carried out in a pilot plant serving an urban development of 200 inhabitants (Les Franqueses del Valle`s, Barcelona, Spain). It consists of eight pilot-scale beds lined with high-density polyethylene and filled with gravel. A schematic diagram of the constructed VOL. 38, NO. 9, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2657
FIGURE 1. Schematic diagram of SFCW pilot plant at the Les Franqueses del Valle` s, Barcelona, Spain. Brief description of beds can be found in the site study of the Materials and Methods section. wetlands is shown in Figure 1. All the beds have approximately the same surface area (54-56 m2), and their aspect ratio (length to width) varies in pairs. The two beds of the shape named A have an aspect ratio of 1×1, of B are 1.5×1, of C are 2×1, and of D ared 2.5×1. Beds of type 1 contain coarse gravel (D60 ) 10 mm, Cu ) 1.6), while type 2 beds contain fine gravel (D60 ) 3.5 mm, Cu ) 1.7). Beds of the type D (the longest) were constructed shallower in order to evaluate the effect of the water depth. Water level was adjusted to 5 cm under the surface, giving an average water depth of approximately 0.5 m for beds of the type A-C and of 0.27 m for D. In all beds, three piezometers (i.e., perforated tubes φ ) 10 cm) were inserted in the middle part of the bed witdth and uniformly distributed throughout the length of the bed, which allowed us to obtain intermediate samples. The SFCWs were planted with phragmites sp (3 plants m-2) in March 2001. Details of the design are given elsewhere (13). Fullsurface coverage was attained at the end of August 2001. For this study, four different total flows were used, giving four HLRs: 20, 27, 36, and 45 mm d-1. Each wetland was automatically supplied with a selected HLR of urban wastewater previously treated in an Imhoff tank. That tank is connected with another in series with a pump from which wastewater is conveyed to a repartition chamber. The primary effluent is then split by means of a weir with eight holes and flows to each constructed wetland. Between the pump and the repartition chamber there is a valve and flowmeter. Sampling campaigns with an approximately weekly pattern were carried out from May 2001 to January 2002. Grab samples were taken using plastic containers from the primary effluent (in the distribution chamber) and the effluents of the eight beds. Two additional sampling campaigns (July and August when the HLR was 36 mm d-1) were done to obtain dissolved oxygen (DO) and redox potential (E) profiles in all the piezometers. DO was measured using an YSI 58 portable oxymeter provided with a membrane for low concentrations and E using a Crison 506 with a platinum electrode. During these two campaigns, influent and effluent samples were taken and analyzed for chemical oxygen demand (COD), ammonia, nitrite, nitrate, and sulfate using conventional methods (14) to qualitatively evaluate the relative importance of the biochemical reactions involved in the degradation of organic matter in each bed type. Methane atmospheric emission from wetland bed was concurrently collected in those of type A and D by using inverted funnels that were inserted into the gravel. Sampling was performed with gastight syringes from the headspace piercing through a septum and determined by gas chromatography-flame ionization detector or hot wire detector depending on its concentration. Chemicals and Reagents. Commercial LAS were supplied by Petroquimica Espan ˜ ola S.A. in a single standard mixture 2658
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 9, 2004
containing a relative composition of the different homologues as follows: C10, 3.9%; C11, 37.4%; C12, 35.4%; and C13, 23.1%. SPC standards were obtained from the University of Cadiz (Spain). HPLC grade water, dichloromethane, acetonitrile, and acetic acid were purchased from Merck (Darmstadt, Germany). All solvents were filtered through a 0.45-µm filter before use. Triethylamine was from Fluka (Steinheim, Switzerland). Analytical Techniques for LAS and SPC Determination. The detailed method for LAS and SPC analysis is described elsewhere (15). Before extraction, samples were filtered through a 0.45-µm membrane filter (Whatman, England) and then acidified to pH 3. Cartridges (ENV+, 200 mg, Isolute, IST, Germany) were conditioned with 6 mL of methanol and 6 mL of water (pH 3). Water samples (200 mL) were extracted using a Baker LSE apparatus (J. T. Baker, Deventer, The Netherlands) connected to a vacuum system set at -15 psi with an automated off-line solid-phase extraction manner. Then the cartridges were dried under vacuum for 20 min to avoid hydrolysis and eluted with 9 mL of a mixture of methanol-dichloromethane (9:1). These extracts were evaporated to dryness with a Reacti-Vap3 (Pierce) operating under a gentle stream of nitrogen and redissolved to 1 mL with methanol, in which 20 µL was injected to the LC-MS system for quantification. Detection was carried out using a HP 1040 M diode array UV-Vis detector coupled in series with the LC-MSD HP 1100 mass selective detector, equipped with an atmospheric pressure ionization source in electrospray interface (ESI) in the negative mode. The ions used for LAS homologue quantitation were at m/z 297, 311, 325, and 339 and a confirmation ion at m/z 183. For SPC homologues, m/z at 215, 229, 243, 257, 313, 327, and 341 were used. Recoveries ranged from 73 to 93% and from 61 to 97% for LAS and SPCs, respectively. The reproducibility (RSD) was from 3 to 9%. Data Analysis. All mean values given in tables are shown (1 SD. Statistical analyses were carried out with the SYSTAT package. In a first step, the one-way ANOVA method was used to check the influence of each considered factor (HLR; type of bed, which includes altogether aspect ratio and water depth; and medium size) on each effluent water quality parameter. After that, a three-way ANOVA was used to evaluate interactions among factors. LAS and SPC data were log-transformed before ANOVA analysis. The Bonferroni method was used to test all pairwise comparisons of marginal averages as a tool to detect the pairs of means that differed significantly. To model the evolution of LAS in the present SFCW system, two models (namely, zero-order and first-order kinetics) were evaluated. The first-order area rate constants (k1) were calculated assuming plug-flow and a zero background
TABLE 1. Concentrations of Selected Parameters in Influent and Effluents of Pilot SFCWsa for the HLRb Tested and Its Estimated Percentage of Organic Matter Removed parameter L-1
influent
DO, mg (Jul/Aug)c E,c mV SO42-, mg of S L-1 (Jul/Aug) NO3-, mg of N L-1 NH3, mg of N L-1 (mean value ( SD) sulfate reductiond (%, Jul/Aug) denitrificatione (%, Jul/Aug) CH4 production (mg m-2 d-1)
A1
A2
B1
B2
C1
C2
D1
D2
