Article pubs.acs.org/est
Bioturbation Delays Attenuation of DDT by Clean Sediment Cap but Promotes Sequestration by Thin-Layered Activated Carbon Diana Lin,† Yeo-Myoung Cho,† David Werner,‡ and Richard G. Luthy*,† †
Department of Civil and Environmental Engineering, Stanford University, Stanford, California 94305, United States School of Civil Engineering and Geoscience, Newcastle University, Newcastle upon Tyne NE1 7RU, England, U.K.
‡
S Supporting Information *
ABSTRACT: The effects of bioturbation on the performance of attenuation by sediment deposition and activated carbon to reduce risks from DDT-contaminated sediment were assessed for DDT sediment-water flux, biouptake, and passive sampler (PE) uptake in microcosm experiments with a freshwater worm, Lumbriculus variegatus. A thin-layer of clean sediment (0.5 cm) did not reduce the DDT flux when bioturbation was present, while a thin (0.3 cm) AC cap was still capable of reducing the DDT flux by 94%. Bioturbation promoted AC sequestration by reducing the 28-day DDT biouptake (66%) and DDT uptake into PE (>99%) compared to controls. Bioturbation further promoted AC-sediment contact by mixing AC particles into underlying sediment layers, reducing PE uptake (55%) in sediment compared to the AC cap without bioturbation. To account for the observed effects from bioturbation, a mass transfer model together with a biodynamic model were developed to simulate DDT flux and biouptake, respectively, and models confirmed experimental results. Both experimental measurements and modeling predictions imply that thin-layer activated carbon placement on sediment is effective in reducing the risks from contaminated sediments in the presence of bioturbation, while natural attenuation process by clean sediment deposition may be delayed by bioturbation.
■
INTRODUCTION Sediment-bound hydrophobic organic contaminants (HOCs) from legacy industrial chemicals, such as dichloro-diphenyltrichloroethane (DDT), polychlorinated biphenyls, and polycyclic aromatic hydrocarbons, continue to pose ecological and human health risks to a number of water bodies. When the sources of contamination are eliminated and the site is netdepositional, then natural depositional processes may be an effective means of burying the older contaminated sediment, gradually reducing the risks from contamination toward overlying water and benthic biota. Natural attenuation by sediment deposition will be affected by deposited sediment physical and chemical properties, erosion and resuspension, and bioturbation from benthic organisms. Bioturbation may disrupt this process of natural capping and tends to increase sedimentto-water flux1,2 of buried contaminants through the enhanced movement of sediment and porewater. Areas being managed for legacy contamination often need assessment of whether or not natural recovery is occurring adequately and timely enough to meet water quality standards and protect human health. If natural attenuation is determined to be insufficient, appropriate remedial actions need to be evaluated, and in situ remediation using activated carbon (AC) is considered a promising option.3 The technology is based on changing the sediment geochemistry to enhance HOCs sorption by means of AC amendment.4−12 HOCs weakly sorbed to sediment are © 2013 American Chemical Society
transferred to the strong sorbent, AC, with the HOC becoming less bioavailable. Various laboratory and field studies validate the benefits of in situ AC amendment, showing reduced porewater concentrations and uptakes by benthic organisms.3 While studies have shown13,14 that active mixing of AC with contaminated sediment helps to realize the full remediation potential of AC remediation in the short term, actively mixing the AC into the sediment is often not feasible for deep water sediments or ecologically sensitive areas. In this case the delivery of AC as a thin layer or in a pelletized form on top of sediment10,15 may be more efficient deployment options. In this AC application scenario, benthic activity is harnessed to mix sorbents down into the biologically active layer and facilitate AC-sediment contact.15 The aim of this paper is to obtain a mechanistic understanding of the effects of bioturbation on the mass transfer of HOCs at the sediment-water and sedimentporewater interface for a simulated natural attenuation process by clean sediment deposition and remedial action by a thin layer AC cap. Received: Revised: Accepted: Published: 1175
September 14, 2013 December 10, 2013 December 20, 2013 December 20, 2013 dx.doi.org/10.1021/es404108h | Environ. Sci. Technol. 2014, 48, 1175−1183
Environmental Science & Technology
■
Article
MATERIAL AND METHODS Experimental Setup. Microcosm experiments were conducted in Pyrex glass columns (6.0 cm inner diameter, 13.3 cm tall) with a Teflon base and conducted at room temperature (21 °C). Each microcosm contained underlying contaminated sediment (collected from a freshwater lake in northern Italy in May 2012). The sediment had 1.8 ± 0.3 ppm ΣDDT (i.e., 4,4′-DDT, 2,4′-DDT, 4,4′-DDE, 2,4′-DDE, 4,4′DDD, 2,4′-DDD, and DDMU) and total organic carbon of 1.9%. Approximately 200 g of wet sediment was placed in each glass column. Microcosms were set up in three modes: 1) control microcosm, comprising contaminated sediment only, 2) a simulated natural deposition microcosm, comprising a contaminated sediment sublayer, with a thin layer (approximately 0.5 cm) of background sediment (collected from the same lake in Italy in 2010, containing 0.03 ± 0.005 ppm ΣDDT; TOC 1.5%), and 3) a simulated AC in situ treatment, comprising a contaminated sediment sublayer, with a thin layer (approximately 0.3 cm) of virgin activated carbon (TOG AC, Calgon Carbon, Pittsburgh, ground and sieved to 75−150 μm particle diameter). For each sediment treatment method, we conducted a set of microcosm experiments with L. variegatus (n = 6) and without worms (n = 3). Porewater and flux measurement devices were placed in all microcosms (TOC art). Each sampling measurement is discussed in detail below. All columns were gently aerated with a glass pipet throughout the experiment. The aerator was positioned above the flux sampler to avoid additional resuspension. Each microcosm was filled with control sediment to a height of 5.5 cm and then carefully filled with reconstituted freshwater.16 The microcosms were gently aerated at the water surface, and the suspended sediment was allowed to settle. One week after control sediment placement and overlying water aeration, forty worms (approximately 0.5 g wet weight) were added to each microcosm using a plastic pipet. This resulted in a worm density of 104 worms m−2, which is comparable to previous bioturbation and bioaccumulation experiments.10,17,18 Benthic community field sampling at the field site showed worm densities between 2 to 6 × 103 worms m−2. Sediment turnover rates in the field may be different from rates measured in the laboratory due to a variety of conditions including organism species, dissolved oxygen concentration, available carbon, and temperature.19 The capping material was soaked in water overnight to obtain a sediment or AC slurry, and the slurry was carefully applied on top of the control sediment with a disposable pipet to obtain a visually even layer covering the entire control sediment the day after worms were deployed. This method of sediment placement and settling was chosen to avoid altering sediment properties and represent settling in field conditions. The clean sediment cap, i.e., the simulated natural deposition layer thickness (0.5 cm), was chosen to be representative of measured sedimentation rates in the field, which were between 0.2−0.8 cm y−1. Sediment profile images taken at the field site showed bioturbation depths of 6−8 cm; we therefore chose a cap thickness that would ensure penetration of the cap in the laboratory microcosm where the bioturbation layer was 2 cm. The AC addition (2.5 g) corresponded to a 2.5% dose by dry sediment weight of AC amendment to the sediment in the columns. We started with an unmixed system and relied on bioturbation to homogenize AC with contaminated sediment.
The following day, the porewater sampler device was placed into the sediment, making sure that the top of the porewater device was at the sediment surface, and the flux measurement device was also placed in each microcosm. Microcosms were lighlty capped to minimize evaporation. Experimental Organisms. L. variegatus was chosen for these microcosm studies because previous tests with L. variegatus worms have assessed contaminant biouptake and biouptake kinetics from sediments,10,16,20−22 the performance of AC amendment,20 and bioturbation.18,23 Worms were obtained from a local aquarium store (Seascapes, Mountain View, CA) and maintained in a glass aquarium with aerated, clean, synthetic freshwater at 15 °C prior to use. Worms (length 2−4 cm) to be used in the experiment were separated and allowed to acclimate to room temperature for 3 h before deployment. Storing the worms at lower temperatures before the start of the experiment may have repressed feeding activities.19 The length of the worms was chosen to ensure that the worms would penetrate the treatment cap in order to observe how bioturbation affects the effectiveness of the cap. Worms were analyzed for background DDT content (0.15 ± 0.1 ppm dw of ΣDDT). DDT Bioaccumulation. The ratio of organic carbon content in sediment to dry weight of worms was more than 50:1.16 Forty mL of water was gently exchanged from each microcosm every other day, and water quality (temperature, DO, ammonia, and pH) was checked. Differences in measured parameters were not correlated with different test conditions; no worm deaths or signficant changes in activity were observed. After 28 d of exposure in the microcosms, worms were retreived from the sediment, allowed to depurate in clean, aerated, synthetic freshwater for 6 h, and then placed in glass vials and frozen at −15 °C. Samples were lyophilized and then extracted (30 mg) in a 1:1 mixture of hexane-acetone following U.S. EPA Ultrasonic Extraction method, cleaned, and analyzed for ΣDDT following analytical procedures described previously.9 Two worms from each sample were retained for lipid analysis.24,25 Water-Sediment DDT Flux. Flux measurement devices were created using low-density polyethylene (PE) sheets of 51 μm thickness (Brentwood Plastics, St. Louis, MO) and precleaned.26 The flux measuring device consisted of ten layers of PE sheets (1.3 g total, each circular sheet was 6 cm in diameter) held in place using a stainless steel frame positioned four cm above the sediment surface. The flux devices were placed one day after cap placement and were retrieved 28 d after exposure. The flux (F, μg m−2 day−1) of ΣDDT was calculated from the flux devices as F=
M A·t
(1)
where M is the ΣDDT measured in the flux measurement devices (μg of ΣDDT), A is the area of the sediment (0.0028 m2), and t is the exposure time (28 d). This calculated flux represents the time-average flux for the 28 d exposure period. Preliminary experiments had been conducted to validate flux measurements using the PE sampler as an infinite sink (SI), and similar devices using passive samplers have been validated to measure flux.27,28 After removal, the PE sheets were individually rinsed with deionized water, wiped clean with Kimwipe, and extracted in 40 mL of hexane for 24 h, cleaned, concentrated, and analyzed for ΣDDT.9 1176
dx.doi.org/10.1021/es404108h | Environ. Sci. Technol. 2014, 48, 1175−1183
Environmental Science & Technology
Article
Table 1. Mass Transfer Model Input Parametersa parameter time frame modeled subvolume (cube) dimensions sediment cap thickness in sediment cap scenario AC cap thickness in AC cap scenario time interval between cube exchanges in control and sediment cap scenario time interval between cube exchanges in AC cap scenario water-phase diffusion coefficient for 4,4′-DDD water-side diffusion boundary layer thickness for 4,4′-DDD Sediment properties sediment interparticle porosity sediment intraparticle porosity sediment particle density sediment grain size radius sediment porewater tortuosity bulk sediment-water partitioning coefficient for control sediment bulk sediment-water partitioning coefficient for clean cap sediment fast release rate from sediment slow release rate from sediment mass fraction of 4,4′-DDD initially associated with ratesc sediment concentration of 4,4′-DDD (control sediment) sediment concentration of 4,4′-DDD (sediment cap) Activated carbon properties AC particle radius AC solid-phase density AC porosity AC-water partition coefficient
parameter annotation
value
source
days (d) dz (cm) thicknessd (cm) thicknessAC (cm) dtbio (s)
28 0.02 0.5 ± 0.05 0.3 ± 0.03 24,000 ± 2,400
estimated estimated estimated
dtbio (s) Daq (cm2 s−1) DBL (cm)
48,000 ± 4,800 4.5 × 10−6 0.01 ± 0.0008
Vfsw (-) ps (-) ds (g cm−3) grainsize (cm) Tortsw Kd (cm3 g−1)
0.5 0.1 2 0.035 0.5 8.9 × 103 ± 2 × 102
Kd (cm3 g−1) ratess (s−1) ratesc (s−1) fslow (-) Cs,control (g cm−3) Cs,cap (g cm−3)
7.3 × 103 ± 5 × 102 1.9 × 10−6 ± 3.8 × 10−7 1.5 × 10−8 ± 3 × 10−9 0.63 9.38 × 10−7 5.00 × 10−9
rac (cm) dac (g cm−3) pac (-) KAC (cm3 g−1)
0.0053 1.96 0.55 2.95 × 108 ± 2.95 × 107
estimated 38
measured by alabaster technique27,33 measured 30,39
measured measured assumed same as Vfsw measured from archived sample measured measured from archived measured from archived measured from archived measured measured
sediment
sediment sediment sediment
measured 5,30 5,30 38
a
Uncertainty analysis conducted assuming normal distribution of uncertain model input parameters around specified mean and standard deviation values.
Porewater DDT Concentration Profile. The in situ porewater concentration profile measuring device comprised four PE strips made from the same material employed in the flux measurement devices. PE sheets were cut to 0.5 cm width and wrapped around clips used as a frame (Acco Regal Clips). Clips were placed in the sediment one day after cap placement and removed after 28 d of exposure in the sediment. After removal, the PE sheets were individually rinsed with deionized water and wiped clean with Kimwipe, and each sample was extracted in 40 mL of hexane for 24 h, cleaned, concentrated, and analyzed for ΣDDT.9 Inserting the flux devices disturbed a small area of the sediment (
