INTRODUCTION

Floating oil slicks entrain seawater and bubbles from the water column and can potentially form an emulsion. The emulsified oil (i.e., the emulsion) has a significantly increased viscosity (Boufadel et al. 2019; Hedgpeth et al. 2021; Prince and Lessard 2004), which poses extra challenges to the response techniques.

It is well established that resins and asphaltenes enhance oil emulsion formation and stability (Fingas and Fieldhouse 2006). Asphaltenes form a network structure within a thin oil film that separates approaching water bubbles (Czarnecki et al. 2012). Meanwhile, the surfactant-like polar resins tend to maintain small water droplets (1 – 20 μm) in oil (Fingas 1995; Fingas and Fieldhouse 2012). Other surfactant materials such as organometallics, and nitrogen, sulfur and oxygen (NSO) compounds are also important in preventing water-water coalescence within the emulsion (Payne 1982).

Many abiotic processes play important roles in the emulsification processes, including solar photo-oxidation, mixing energy, sheening, etc. Photo-oxidation happens normally after an oil spill, and fundamentally changes the properties of the oil residue (Aeppli 2022). Hydrocarbons are oxidized to a variety of oxygen-containing products (Overton et al. 1980) which likely have surfactant properties that accelerate the emulsification process and stabilize the formed emulsions (Brandvik and Daling 1991). While the photo-oxidized NSO compounds would have enhanced water solubilities and might be expected to be removed from the thick oil, increased diffusion coefficients as a result of elevated viscosities and specific gravities may prevent these materials from being lost (Payne 1982). Nonetheless, the photo-oxidation happens more extensively in thinner oil slicks (Katz et al. 2022) as sunlight penetrates the oil slick (Ward et al. 2018) and oxygen diffuses (Lichtenthaler et al. 1989), while many lab simulations failed to accomplish this condition. Freeman and Ward (2022) set up 200 μm oil thickness to study the photo-dissolution apparent quantum yields as a function of wavelengths and photons, which in their case simulated the photooxidation corresponding to 2 – 3 weeks light exposure in Gulf of Mexico in summertime. Sheen leads to the removal of many of the lower molecular weight materials and enhances dissolution and evaporation (Payne 1982). Thicker masses in oil slicks exhibit higher wind resistance and thus higher frictional resistance at the oil water interface compared to the sheen (Zodiatis et al. 2017), which leads to the trail and furthermore, disappearance of sheen (Grose et al. 1979; Ranieri et al. 2013; White et al. 2019). Nonetheless, the spilled oil thickness is normally found around 100 μm or less (Carolis et al. 2012; Leifer et al. 2012; Svejkovsky et al. 2016), though heavily emulsified oil could be more than 200 μm (Daling et al. 2003; French-McCay et al. 2021; OC et al. 2012).

Moreover, challenges remain in terms of the actual behavior of emulsion at sea. For example, the surfactants (i.e., the polar resins and some photo-oxidized products) that are inherently present in a spilled oil remain in the flask in most lab-scale investigations, which would impact the subsequent behavior of the emulsion. While at sea, the surfactants are likely to leave the oil, which impacts the emulsification of and/or the stability of emulsion. In an inter-lab comparison on emulsification processes in different vessels (French-McCay et al. 2022), the viscosity data of the emulsion at various labs showed major differences between wave tank/flume tank simulations and the bench-top scale experiments. This was related to the water content after different time period: though SL Ross and NRCan also used flume tank setups (recirculating or closed), their observed water content differed a lot from those at NJIT as NJIT removed sheen materials continuously – and this difference happened after 6 hrs. Herein, the mesocosm-scale investigation was conducted while simulating multiple natural processes (i.e., artificial solar lamps, hite-cap mixing, Langmuir cell, removing sheen, etc.) during the oil emulsification process, aiming to provide more data points to discern the role of photo-oxidation.

METHODS

The flume tank is located in an indoor lab at the Center for Natural Resources at NJIT (NJ, US). The flume is 8 m long × 0.6 m wide × 0.3 m deep (Figure 1), and a disposable white polyethylene plastic sheet covers the inside wall of the flume tank for easy setup and cleaning between experiments. The tank was filled with artificial seawater (salinity at 33%°, maintained at 15°C by a chiller, using instant ocean sea salt (Pentair Aquatic Eco Systems, Inc., NC, US)) to a water depth of 0.25 m for each experiment. The flume tank is housed in a flame-retardant tent (Item 1) with air filtration via two 0.18 m³/s fans (Item 2, Vivosun, CA, US) to two 140 lbs activated carbon drums (Item 3, Carbtrol Corporation, CT, US) to prevent odor and contaminant escape. An automated control system (LabVIEW by National Instruments) is designed to control the oil movement, sheening, and white-caps through a control camera (Item 4).

View Figure

• Download as Powerpoint
• Download Figure

The major components of the mesoscale testing system (illustrated in Fig. 1) are described below:

Sunlight:

Artificial sunlight (Item 5) is generated by two Philips metal halide bulbs (BT56, 1500 W; Grainger, US), placed above the flume to illuminate the whole area of the thick oil slick. The spectrum measured by Flame Spectrometers (Ocean Insight, FL, US) is given in Fig. A1 (the probe was placed on the water surface under one of the bulbs), where the UV range intensity (W/m²) is comparable to the intensity of natural sunlight (20 – 40 W/m² in summer time, and 10 to 30 W/m² in winter time (Apell and McNeill 2019; Rezaei et al. 2017; Sichel et al. 2007).

View Figure

• Download as Powerpoint
• Download Figure

Analysis:

The analyses conducted for both bulk oil samples and sheen samples are listed in Table 1.

Table 1. Oil properties measured at various time steps

View Fullscreen

The oil volume is of importance as samples were collected 5 times (at least 80 mL bulk oil in total) during the 72 h, and with evaporation and sheen-removal, the total oil loss could be significant, so the sampling was designed to be of small volumes to avoid affecting the bulk oil behaviors through the experiment time. The initial oil loading volume could not be large, as otherwise the surface oil would be too thick and would block the penetration of light (Katz et al. 2022).

RESULTS

An experiment with Alaska North Slope oil is shown in Fig 2. The oil slick in the normal spreading period was around 6 m long, and slightly narrower than the tank width (0.6 m), showing a fair spreading and thickness. Fig. 3 shows the bulk oil at various times. After 1 h, the oil was broken up into droplets and separated floating on the surface, with water bubbles entrained in most oil droplets (Fig 3b). After 9 h, water bubbles in the oil phase were of diverse sizes, with obvious large bubbles (around 2 cm) as shown in Fig. 3c. After 24 h, the water bubbles in the oil slick were smaller at less than 1 cm and were more homogeneous (Fig 3d). After 48 h (Fig. 3e), the oil slick was thicker and the entrained water bubbles were smaller and less visible to the naked eye. After 72 h (Fig. 3f), the water bubbles were evenly distributed within the oil slick, causing the oil slick to be brownish and form ‘chocolate mousses’ (Bridie et al. 1980; Clark et al. 2010).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The physical properties of the bulk oil at various time steps are reported in Fig. 4. Fig. 4a shows the density of the ANS oil increased from 0.83 g/mL (fresh) to 0.96 g/ml after 72 hours emulsification simulation, and the viscosity (Fig. 4b) increased from 13 cP (fresh) to >1800 cP. The oil/water interfacial tension (Fig. 4c) increased from 6 mN/m (fresh) to >20 mN/m after 72 h. Fig. 4d shows that the entrained water content increased from 5% to >54% after 72 h. The changes of these physical properties of the bulk oil, as well as visual observations from the CCTV camera system, clearly indicated that the crude oil took up more water bubbles as time passed, and the viscosity, water content, density and interfacial tension all increased.

View Figure

• Download as Powerpoint
• Download Figure

Spreading tests were run in a stainless-steel pan covered with a thin PTFE film. The photos in Fig. 5a show that the fresh oil was very dispersive, and after only 20 s there was no obvious oil slick on the water surface. At 9 h, the sample mostly dispersed by 60 s (Fig. 5b), but a few emulsified oil droplets persisted. After 24 h (Fig. 5c) the spreading took longer and there were major oil droplets floating on the surface at 60 s. These droplets diminished after 300 s. At 72 h (Fig. 5d), the emulsified oil droplets there were still droplets entraining water bubbles on the surface at 300 s.

View Figure

• Download as Powerpoint
• Download Figure

The water bubble size distribution results are given in Figs. 6 and A2. After 1 h, the median diameter (d50) was ~35 μm (Fig. 6) and it dropped to ~10 μm after 6 h and 9 h emulsification simulations. After 48 h and 72 h, the d50 increased to 30 μm and 50 μm, respectively. Figure A2 shows that this was likely due to the formation of compound droplets (Muriel and Katz 2021), where crude oil was entrained in water bubbles which were in turn entrained in the oil slick.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The FTIR results are presented in Fig. 7. The photo-oxidation generated typical chemical groups such as the OH stretching band at around 3000 – 3500 cm^-1 (Boukir et al. 1998; Colthup 2012) and carbonyl groups (diarylketone or quinone structures) at wavenumber 1600 – 1700 cm^-1 (Boukir et al. 2001)) within 24 hand. These increased by 72 h photo-oxidation and perhaps explain the more stable water-crude oil emulsion at these times; such compounds are more hydrophilic and might act as surfactants (Brandvik and Daling 1991; Lee 1999) to decreased the interfacial tension between crude oil and seawater.

View Figure

• Download as Powerpoint
• Download Figure

CONCLUSIONS

The emulsification of Alaskan North Slope (ANS) was studied in a mesocosm scale flume tank equipped with artificial solar lamps, white-cap mixing water jets, Langmuir cell mixing, and two weirs to remove diffused sheening components to simulate the oil emulsification process where spreading is designed to be ‘infinite’ and the oil weathering process is in a mostly ‘natural’ environment. The oil became progressively emulsified over 3 days, becoming denser and more viscous as water was incorporated. The water content was over 50% after 3 days under sunlight. The size of entrained water bubbles likely contributed to the increased viscosity, and compound droplets formed with a crude oil in water in crude oil structure. The FTIR spectra showed the addition of oxidized groups, which intensified after longer light exposure. These compounds apparently acted like surfactants to accelerate the emulsification process and stabilize the formed emulsions.

This investigation provided new data points in a mesocosm scale vessel that adopted key abiotic parameters to mimic field conditions, especially including the quantification of photo-oxidation. Further work is needed with additional oils to understand how different oil components (perhaps obtained with SARA) affect the oil emulsification process. In addition, the role of photo-14 oxidation and sheening need to be further explored through additional control experiments (dark environment, closed tanks, etc.).