oIL dropLet SIze dIStrIbutIon AS A FunctIon oF enerGY dISSIpAtIon rAte In An eXperIMentAL WAVe tAnK

The U.S. National Research Council (NRC) Committee on Understanding Oil Spill Dispersants: Efficacy and Effects (2005) identified two factors that require further investigation in chemical oil dispersant efficacy studies: 1) quantification of mixing energy at sea as energy dissipation rate and 2) dispersed particle size distribution. To fully evaluate the significance of these factors, a wave tank facility was designed and constructed to conduct controlled oil dispersion studies. A factorial experimental design was used to study the dispersant effectiveness as a function of energy dissipation rate for two oils and two dispersants under three different wave conditions, namely regular non-breaking waves, spilling breakers, and plunging breakers. The oils tested were weathered MESA and fresh ANS crude. The dispersants tested were Corexit 9500 and SPC1000 plus water for no-dispersant control. The wave tank surface energy dissipatation rates of the three waves were determined to be 0.005, 0.1, and 1 m2/s3, respectively. The dispersed oil concentrations and droplet size distribution, measured by in-situ laser diffraction, were compared to quantify the chemical dispersant effectiveness as a function of energy dissipation rate. The results indicate that high energy dissipation rate of breaking waves enhanced chemical dispersant effectiveness by significantly increasing dispersed oil concentration and reducing droplet sizes in the water column (p < 0.05). The presence of dispersants and breaking waves stimulated the oil dispersion kinetics. The findings of this research are expected to provide guidance to disperant application on oil spill responses. * Corresponding author; Ph: (902) 426-3442; fax (902) 426-6695; email: [email protected] 622 2008 InternatIonal oIl SpIll ConferenCe Scornette 2002). Velocity shear with its associated friction causes the dissipation of kinetic energy of the fluid. One may use velocity measurements in a selected water body to compute the shear and subsequently the energy dissipation rate per unit mass, ε, which varies in time and space. The effectiveness of a chemical dispersant is typically evaluated at various scales ranging from the smallest (10 cm, typical of the swirling flask test in the laboratory) to the largest (10’s to 100’s meters, typical of field scale open water dispersion tests). In terms of product selection for spill response operations, standard laboratory assays for the evaluation of oil dispersant effectiveness such as the swirling flask test have limitations due to insufficient mixing energy or limitation to account for the transport and interaction between oil and dispersant in the water column (Sorial et al. 2004a; Sorial et al. 2004b; Venosa et al. 2002). But testing on the sea is expensive and not always reproducible due to uncontrollable environmental variables. To bridge the knowledge gap, a wave tank facility was constructed for evaluating chemical dispersant effectiveness at various quantitatively defined wave conditions. The current hypothesis is that the energy dissipation rate per unit mass (ε) plays a major role in the effectiveness of a dispersant. The hydrodynamic tests (Boufadel et al. 2008; Venosa et al. 2005; Wickley-Olsen et al. 2007) have demonstrated that the nonbreaking waves and breaking waves that were generated in our test tank facility had similar energy dissipation rates as for natural waters (Delvigne and Sweeney 1988). Conservation of ε between the wave tank and actual field conditions provides support for the use of our test system to evaluate the operational effectiveness of chemical oil dispersants. The primary objective of the project is to develop protocols that can be used for the evaluation of chemical dispersant effectiveness in the wave tank facility. This paper documents the findings of experimental studies on chemical dispersant effectiveness, with emphasis on dispersed oil droplet size distribution as a function of energy dissipation rate under non-breaking and breaking-wave conditions. Details of the wave generation and characterization of the wave tank are reported elsewhere (Boufadel et al. 2008). The detailed oil distribution chemistry analysis and ultraviolet fluorometry analysis is reported in Venosa et al. (2008) and Kepkay et al. (2008). MAterIALS And MethodS Wave tank facilities Figure 1 presents the schematic of the wave tank located at the Bedford Institute of Oceanography (Dartmouth, NS). The tank is 32 m by 0.6 m and 2 m high. The average water depth during the experiments was 1.50 m. Different waves are generated by a computer-controlled flap-type wave maker situated at one end of the tank linked to an adjustable cam that controls its stroke length to alter wave height. The wave frequency (and thus wavelength) is controlled by the rotation speed of the cam that is driven by an electric motor. The computer-controlled wave-generator can produce progressive waves with desired wave parameters. The wave energy was removed at the opposite end from the wave-maker by the wave absorbers to minimize reflection and interference of incoming progressive waves from the past waves. The recurrent breaking waves can be generated at the same location by superimposing a wave of one frequency on another wave of a different frequency, causing the wave to increase in height and break. The energy dissipation rate per unit mass (ε) was used to characterize the intensity of the breaker. It was evaluated by the correlation function method (Kaku et al. 2006) using a time series of velocity measurements with Acoustic Doppler Velocimeter (SonTec/YSI, Inc. San Diego, CA) at different locations in the tank. Hydrodynamic characterization for the wave tank was reported elsewhere (Boufadel et al. 2008; Venosa et al. 2005; Wickley-Olsen et al. 2007). Experimental procedures The chemical dispersant effectiveness for two crude oils [Medium South American (MESA) and Alaska North Slope (ANS)] and two dispersants (Corexit 9500, and SPC1000) plus water as nodispersant control was investigated under three wave conditions (regular non-breaking wave, spilling breaking wave, and plunging breaking wave). A three-factor mixed-level factorial experiment was designed to include 18 treatments with triplicate runs for each treatment, resulting in 54 runs for the entire design (Table 1). Treatments were applied in random order to minimize the impacts of confounding factors such as temperature, salinity, and wind on the dispersant effectiveness of crude oil. For each experiment, seawater was pumped directly from the Bedford Basin through a double layer sock-filter (Atlantic Purification Ltd, Dartmouth, NS, Canada) with a pore size of 25 and 5 μm for the coarse and fine filters, respectively. The background temperature, salinity, and fluorescence intensity and particle size distribution were recorded before the experiment started. For each experiment, 300 ml of crude oil was gently poured onto the surface of the water within a 40cm (inner diameter) tubular ring (constructed of NSF-51 reinforced clear PVC) 10 m from the wave generation paddle. Immediately after oil addition, 12 ml of dispersant (water for the control) was sprayed conically on top of the oil through a pressurized nozzle (60 psi; 0.635mm i.d.) at a dispersantto-oil ratio (DOR) of 1:25. The ring was then lifted immediately prior to contact with the first wave. The desired wave conditions were operated continually during the 2h experiment. Samples were collected by using a set of 100ml syringes connected with stainless steel manifold from four horizontal locations (8, 12, 16, and 20 m downstream from the wave maker), three depths (5, 75, and 140 cm from the average water surface), and four time points (5, 30, 60, 120 min). These samples were subdivided for oil chemistry analysis using dichloromethane (DCM) extraction followed by ultraviolet spectrophotometer (UVS) analysis (Venosa et al. 2002) (Venosa et al., 2008) and ultraviolet fluorescence analysis (Kepkay et al. 2002) (Kepkay et al. 2008) of dispersed oil concentration. Dispersed oil droplet size distribution was measured using a laser in-situ scattering and transmissiometer (LISST-100X, Type C, Sequoia Scientific, Seattle, WA). For this instrument, there were 32 particle size intervals logarithmically placed from 2.5 – 500 μm in diameter, with the upper size in each bin 1.18 times the lower. Particle size distribution was expressed as the average volumetric concentration of oil droplets within each interval of the size range. The data acquisition was conducted at real time operation mode throughout each experiment, with an average of 10 measurements for each sample being measured every 3 seconds. The in-situ dispersed oil droplet size distribution was measured at three different depths (45, 80, and 125 cm from the average water surface) on one horizontal location (16 m downstream from the wave-maker) over four time periods (0-30, 30-60, 60-90, 90-120 min). Therefore for each measurement, there were four10-min continuous recordings with 30-min interval throughout each experiment. Results and Discussion Dispersant effectiveness as a function of energy dissipation rate The average energy dissipation rate (ε) at the surface mixing zone was determined to be approximately 0.005, 0.1, and 1 m2/ s3 through hydrodynamic experiments conducted separately in the wave tank (Boufadel et al. 2008). Throughout the oil dispersion experiments, the effects of chemical dispersants and breaking waves dictated the oil distribution at the surface and in the water column. Without dispersant, oil slicks were dispersed poorly at low ε under regular non-breaking waves; and most of the oil was transported to the end of the wave tank (at the surface in front of the wave absorbers) within 5 min and persisted at this place afterwards. At higher energy dissipation rates, the oil was dispersed at