Article pubs.acs.org/est
Effects of Oil and Dispersant on Formation of Marine Oil Snow and Transport of Oil Hydrocarbons Jie Fu,† Yanyan Gong,† Xiao Zhao,† S. E. O’Reilly,‡ and Dongye Zhao*,† †
Environmental Engineering Program, Department of Civil Engineering, Auburn University, Auburn, Alabama 36849, United States U.S. Department of the Interior, Gulf of Mexico OCS, Office of Environment, New Orleans, Louisiana 70123, United States
‡
S Supporting Information *
ABSTRACT: This work explored the formation mechanism of marine oil snow (MOS) and the associated transport of oil hydrocarbons in the presence of a stereotype oil dispersant, Corexit EC9500A. Roller table experiments were carried out to simulate natural marine processes that lead to formation of marine snow. We found that both oil and the dispersant greatly promoted the formation of MOS, and MOS flocs as large as 1.6−2.1 mm (mean diameter) were developed within 3−6 days. Natural suspended solids and indigenous microorganisms play critical roles in the MOS formation. The addition of oil and the dispersant greatly enhanced the bacterial growth and extracellular polymeric substance (EPS) content, resulting in increased flocculation and formation of MOS. The dispersant not only enhanced dissolution of n-alkanes (C9−C40) from oil slicks into the aqueous phase, but facilitated sorption of more oil components onto MOS. The incorporation of oil droplets in MOS resulted in a two-way (rising and sinking) transport of the MOS particles. More lower-molecular-weight (LMW) n-alkanes (C9− C18) were partitioned in MOS than in the aqueous phase in the presence of the dispersant. The information can aid in our understanding of dispersant effects on MOS formation and oil transport following an oil spill event.
■
INTRODUCTION Marine snow (MS) refers to a class of organic and inorganic particles or aggregates (≥0.5 mm) naturally formed in the ocean, consisting of minerals, detritus, bacteria mucus, phytoplankton, and zooplankton feces.1,2 Marine snow is ubiquitous in the ocean, and plays an important role in downward transport of materials and energy by gravitational settling.3 Marine snow can be an important food source for organisms living in the aphotic zone.4 While the exact formation mechanism of MS is quite complex, it is believed that both physicochemical processes (e.g., coagulation and flocculation) and microbial actions are involved.2 The 2010 Deepwater Horizon (DwH) oil spill incident gushed approximately 4.9 million barrels (779 million liters) of South Louisiana Sweet Crude oil into the Gulf of Mexico.5,6 To mitigate the environmental impacts, ∼2.1 million gallons (7.9 million liters) of chemical dispersants (Corexit EC9500A and Corexit 9527A) were applied at the sea surface and near the 1500 m deep wellhead.7 In May 2010, shortly after the onset of the DwH oil spill, oil-associated marine snow (hereafter referred to as marine oil snow, MOS) of very large floc size (>1 cm) was observed at the surface in the vicinity of the spilled oil slicks near the oil platform.1 During the spill, a continuous oil plume was reported about 10 miles long, 3 miles wide (16.9 × 4.83 km2), and up to 400 ft (122 m) in depth, and the plume persisted for months without appreciable biodegradation.8 While the mechanisms of plume formation are complex and largely unknown, the following factors have been cited to have a role: the interplay of gas and oil in multiphase flow, preferential solubility of each oil constituent, and potential gas hydrate formation.8,9 However, the © XXXX American Chemical Society
role of oil, the dispersants and dispersed oil in the oil plume formation or formation of MOS has not been addressed. We hypothesize that oil and oil dispersants will interact with suspended particulate matter (SPM), and thus impact the formation of MOS. Furthermore, MOS will play an important role in transporting oil components in different environmental compartments.10 Many investigations have indicated that accumulation of oil aggregates, dispersant components, minerals, and biologically derived solids on the seafloor is associated with the DwH oil spill.11,12 Researchers13 observed that some brown flocculent (likely MOS) induced widespread signs of stress for seafloor coral colonies following the DwH incident. Yet, the role of MOS in the material transport process remains unexplored. The most important factors that can affect formation of MOS include the hydrodynamic conditions, collision rate of suspended mineral particles, biologically derived mucus and biopolymer, interactions of oil components with SPM, particle coagulation/ flocculation characteristics, and interactions of oil components with microorganisms.1 In addition, based on our recent work,5,6 oil dispersants can significantly facilitate uptake of oil components on sediment particles, and dispersed oil is likely more adsorbable to SPM. However, detailed information remains lacking on how oil and dispersants impact the formation rate and characteristics of MOS. Moreover, it is not clear whether Received: August 27, 2014 Revised: November 16, 2014 Accepted: November 24, 2014
A
dx.doi.org/10.1021/es5042157 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Table 1. Observations of Marine Snow Formation during Roller Table Experiments case
formation of MSa
aggregates type
I: untreated seawater only
no
/
II: untreated seawater + oil
yes (day 2b)
III: untreated seawater + dispersant IV: untreated seawater + oil and dispersant V: microfiltered seawater + oil VI: sterilized seawater only
yes (day 1) yes (day 2)
flocs, strips (day 6) flocs flocs
no no
/ /
VII: sterilized seawater + oil
no
/
a
maximum Dm
aggregation ratec 24.9 d−1
floating behavior
sinking behavior
0.48 mm (day 25) 2.10 mm (day 6)
31.8 d−1
7.68 mm/sd (day 28) 15.76 mm/s (day 9)
1.65 mm (day 3) 1.55 mm (day 4)
32.2 d−1 26.3 d−1
14.81 mm/s (day 3) 5.90 mm/s (day 2)
/ 0.34 mm (day 26) 0.39 mm (day 12)
/ 9.6 d−1
/ 3.61 mm/s (day 28)
day 6e, 9.34 mm/s (day 14) no day 3, 4.98 mm/s (day 19) / no
21.8 d−1
3.14 mm/s (day 28)
no
no
A Dm of >0.5 mm indicates formation of MS bAppearance time cAggregation rate = (particle number at Day 28 − initial particle number)/28 Maximum Vms or Vmf eTime of first appearance
d
± 1 °C). The rotation speed was set at 20 rpm to keep MS/MOS suspended. On a daily basis, the bottles were removed from the roller tables and placed up-right to allow the MS/MOS to sink/ rise to the bottom/top. Then, the bottles were gently turned over to initiate resinking/rising of the flocs. The morphology and motion of the MS/MOS flocs were then recorded through photographic imaging and video recording with a Canon EOS 600D camera and a Digital IXUS 80 IS camera (Tokyo, Japan). All tests were carried out in duplicate. Characterization of Marine Snow Particles. Four parameters were used to characterize the MS/MOS particles, that is, particle number (N), mean diameter (Dm, mm), total volume (TV, mm3), and mean sinking/rising velocity (Vms or Vmf, mm/s). Dm is defined as the average particle size measured at 2 degree intervals and passing through the particle’s centroid. TV is calculated from the equivalent spherical volume. A smaller particle size or higher number density favors incorporation of more oil into the particles; and a larger TV is associated with a looser and less stable structure of the flocs. Details on the characterization of MS/MOS particles are described in Section S1 of the Supporting Information (SI). Total Bacteria Counting and Determination of Extracellular Polymeric Substance. The total bacteria numbers (TBN) in cases I−IV (with untreated seawater) after predetermined incubation times (0, 2, 14, and 28 days) were estimated following the fluorochrome staining method.15 In addition, the extracellular polymeric substance (EPS) in Cases I−IV incubated for 0, 2, 14, and 28 days were extracted and determined through the gravimetric method. Details on the methods for quantifying TBN and EPS are described in Section S2 of the SI. Analytical Methods. Section S3 in the SI gives methods for the seawater analysis. n-Alkanes (C9−C40) in the aqueous phase and in MS/MOS were determined using gas chromatography− mass spectrometry (GC-MS) by modifying the method by Liu et al.12 The seawater or MS/MOS samples were first extracted with dichloromethane/n-hexane (1:1, v/v), concentrated and then analyzed by GC-MS. Details on the extraction method and GCMS programing are described in Section S4 of the SI.
or how MOS facilitates uptake and transport of important oil components. The overall goal of this work was to investigate the effects of oil and oil dispersants on the formation of MOS. The specific objectives were to (1) investigate the formation of artificial marine snow through roller table experiments in the presence of oil and a prototype oil dispersant (Corexit EC9500A), (2) evaluate the role of natural suspended solids, especially indigenous microorganisms, in the formation of MOS, (3) characterize physical properties (number, size, and sinking/rising velocity) of MOS, and (4) examine the influence of MOS on the fate and transport of spilled oil.
■
MATERIALS AND METHODS Materials. Seawater samples were collected from the top water column (30 cm) at Grand Bay, AL (30°37′45″ N, 88°18′25″ W) in October 2012. Key physicochemical properties of the seawater include pH = 7.6 ± 0.1, dissolved organic carbon (DOC) = 0.70 ± 0.2 mg/L, total dissolved solid (TDS) = 30 ± 1 g/L, suspended solids (SS) = 34.2 ± 0.8 mg/L, Cl− = 16.79 g/L, Na+ = 9.87 g/L, SO42− = 2.48 g/L, Mg2+ = 1.13 g/L and ionic strength (IS) = 0.60 M. A surrogate Louisiana Sweet Crude oil (LSC) was acquired through the courtesy of BP in America (Houston, TX), which mimics the BP Macondo well oil.14 Corexit EC9500A was acquired through the courtesy of Nalco Company (Naperville, IL). All chemicals used in this study were of analytical or higher grade and obtained from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Fair lawn, NJ). Roller Table Experiments. Roller table experiments were conducted using 755 RMV Jar Mills (45.0 cm × 28.6 cm × 34.9 cm, U.S. Stoneware, East Palestine, OH) and 250 mL Boston round glass bottles (6 cm × 13.6 cm). Experiments were designed to test MS/MOS formation under seven scenarios: (I) untreated seawater only; (II) untreated seawater + oil; (III) untreated seawater + dispersant; (IV) untreated seawater + oil and dispersant; (V) microfiltered seawater + oil; (VI) sterilized seawater only; and (VII) sterilized seawater + oil. The doses of oil and dispersant were set at 0.06% (v/v, oil/seawater) and 1:20 (dispersant/oil), respectively. Oil was injected into the upper water column and there was no headspace in the bottles. In selected cases, microfiltration was exercised using 0.45 μm cellulose nitrate membranes to remove the SPM from the seawater, and seawater sterilization was carried out by autoclaving the seawater at 121 °C for 35 min. In each scenario, the bottles were incubated for 28 days at room temperature (22
■
RESULTS AND DISCUSSION Formation of MS/MOS. Table 1 summarizes the observations on the MS/MOS formation during the roller table experiments for the aforementioned seven cases. In Case I, the suspended particles in the raw water gradually aggregated over
B
dx.doi.org/10.1021/es5042157 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
Figure 1. Morphological change of particle aggregates in Cases II−IV during 28-day roller table experiments. Case II: seawater + LSC oil (0.06%, v/v); Case III: seawater + Corexit EC9500A (0.003%, v/v); and Case IV: sweater + LSC oil (0.06%, v/v) + Corexit EC9500A (0.003%, v/v). All bottles (250 mL) were incubated at room temperature (22 ± 1 °C) at a rotation speed of 20 rpm. Scale bar = 10 mm.
Figure 2. Changes of marine snow characteristics in Cases II−IV during 28-day roller table experiments: (a) particle number (N), (b) mean diameter (Dm), (c) total volume (TV), and (d) mean sinking and rising velocity (Vms and Vmr).
the experimental period (N reduced from 956 ± 32 to 258 ± 25). While aggregates were observed with a maximum Dm of 0.48 mm, which is slightly smaller than the threshold of 0.5 mm.2,16 As such, no typical MS was formed in Case I.
In Case II, the presence of oil greatly speeded up particle aggregation, and large MS flocs were formed in 2 days (Table 1). At Day 6, the smallest N (19 ± 3), and the largest Dm (2.10 mm) and TV (92.17 mm3) were obtained (Figures 1, 2a−c). The C
dx.doi.org/10.1021/es5042157 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
formed MS flocs appeared relatively elongated, though largely irregular, and showed a high Vms (15.76 mm/s) (Table 1 and Figure 2d). Afterward, the flocs number remained stable at approximately 22. From Day 6 on, some oil droplets were formed (due to mechanical dispersion and bacterial activities) and became incorporated into the MS flocs and smaller SPM seeds, resulting in more fine strip-like particles (N > 70) with a relatively bigger head of oil droplet (approximately 0.5 mm in diameter) and a mineral tail (1−2 mm in length) (Figure 1). The MS/MOS particles formed during the initial stage (≤6 d) tended to settle by gravity, whereas those strips with sufficient oil incorporated were lighter than water, and thus, ascended at a relatively high Vmr (9.34 mm/s) (Figure 2d). From day 16 to 18, an interesting phenomenon was observeda large oil droplet of approximately 2 mm in diameter was incorporated into a floc and further became interwoven with several flocs to form a large star-like MOS (Figure 1). After day 19, with the dissipation of the oil droplet, this large MOS gradually disassembled. The addition of the dispersant (Case III) resulted in rapid formation of MS in 1 day (Table 1). The resulting flocs were generally smaller with a lower Vms (14.81 mm/s) than in Case II (Figures 1, 2b, 2d). On Day 3, the largest Dm (1.65 mm) and lowest N (19 ± 2) were observed. The flocs were then split into smaller ones gradually. On Day 12, the particle number grew to 75 ± 2, and subsequently, the MS flocs slowly reaggregated (Figures 1, 2a, 2b). In the presence of both dispersant and oil (Case IV), the natural suspended particles aggregated rapidly to form large flocs within 2 days (Table 1). At Day 4, the largest flocs (Dm = 1.55 mm) were formed and the liquid appeared translucent with a sandy color (Figures 1, 2b). After Day 4, these large flocs were broken into many smaller ones with the incorporation of oil droplets. The formed oil droplets in Case IV were mostly less than 0.1 mm, which were much smaller than those in Case II. It is noteworthy that from Day 3 on, the previously sinking flocs gradually began to rise due to the incorporation of oil droplets or oil components of lower density. On Day 28, only a few small particles sank (Figure 2), while most MOS either floated at the surface or suspended in the water column. Compared with Cases II and III, more and smaller flocs (N = 190 ± 10, Dm = 0.58−0.98 mm) were formed in the presence of the dispersed oil at the end of incubation with a slower Vms or Vmr (2.14 and 2.93 mm/s, respectively), and the TV of these flocs was much higher (90.9 mm3) (Figure 2c). The sinking velocity of collected MS in the Gulf of Mexico ranged from 68 to 553 m/d.1 In our study, the sinking velocities in Cases II−IV were 369−1361, 332−1280, and 164−510 m/d, respectively. The faster velocities from this can be attributed to the relatively higher turbulence in the experiments, leading to formation of denser MS/MOS. A sample video is provided in SI (Video S1) to demonstrate the motion of MS flocs in Cases II−IV. No MS was generated in the microfiltrated seawater even in the presence of oil (Case V) (Table 1). The particles aggregation in sterilized seawater (Cases VI and VII) was quite slow compared with Cases I−IV and no MS was formed (Table 1). Formation Mechanism of MOS and Effects of Oil and Dispersant. Proper turbulence can induce suspension of the particles and facilitate particle collisions.17 Based on our preliminary work and work by others, the roller table approach provided sufficient rotational turbulence, which kept the particles in suspension, increased the collision rate between suspended particles and facilitated the particle aggregation,18 while simulating conditions near the sea surface.
No MS was formed under the particle-free conditions even in the presence of oil (Case V, Table 1), indicating that the particulate matters were essential for the formation of MS. The lack of sufficient sticky particulate matters may account for the reported observation that there was no MOS formed in situ in the deepwater−oil plumes following the 2010 DWH oil spill,1 i.e., the plumes are likely composed of mostly dispersed oil droplets. As most particles in marine system carry net negative charges, the classical Derjaguin−Landau−Verwey−Overbeek (DLVO) theory can be used to interpret the collision and adhesion of these like-charged particles.19 According to this theory, the forces governing the particle interactions are the van der Waals attraction (WA) and the electrostatic repulsion (ER).20 When the collision energy overcomes the primary maximum energy barrier to reach the primary minimum, aggregation will occur; alternatively, when it falls in the secondary minimum domain, flocculation takes place. In this work, the breakup of the initially formed large MS/MOS flocs suggested that the flocculation was at least partially reversible, that is, the secondary minimum was an important mechanism in the formation of MS/MOS. The total volume of MS/MOS in cases II−IV increased from 9.5 to 27.6 mm3, 10.0 to 91.7 mm3, and 10.6 to 50.8 mm3, respectively, during the 28 day incubation (Figure 2c). The TV of sinking flocs in case IV fluctuated remarkably in the initial 7 days, indicating the instability of the young MOS structure. Due to the presence of the organic substances (i.e., oil hydrocarbons, dispersants and EPS), the non-DLVO forces (e.g., hydrophobic force and Lewis acid−base interactions) also play important roles in the formation of MS/MOS. The organic substances increase the surface friction or stickiness, leading to sticky collisions, which are more effective than elastic collisions. Consequently, not only more MS/MOS flocs are formed, but the flocs are likely bulkier because of the heat effect due to the transfer of more kinetic energy into the internal energy. The experimental results also revealed that the indigenous microbes play a critical role in the particle aggregation and formation of MS/MOS. Of the seven cases, the lowest particle aggregation rate was found in Cases VI and VII (9.6 and 21.8 d−1, respectively, Table 1), where the seawater was sterilized. This observation reveals that the bacteria activity plays a critical role in the formation of MS. In fact, bacteria are known to produce sticky matters, that is, EPS, which can be incorporated in the suspended particles to form mucus matrices that facilitate particle aggregation and MS formation.1 The main contents of EPS are carbohydrate, protein and nucleic acid, and these macromolecules can cause aggregation of particles through sweep flocculation and bridging effect.21 The functional groups of EPS include carboxylic, hydroxyl, and amino ions,22 which can mediate particle aggregation through electrostatic and Lewis acid−base interactions.23 Figure 3 shows that both TBN and EPS in Cases I−IV rapidly increased with the particle aggregation from day 1 to 14, indicating the important role of EPS in the MS/ MOS formation process. The much enhanced MS/MOS formation in Cases II−IV (Table 1) indicated the important role of oil and the dispersant. Dissolved/dispersed oil hydrocarbons in aqueous solutions can compress the diffuse layer surrounding the particles and reduce the repulsive energy, and thus, promoting the aggregation of particles.24 In addition, the incorporation of the oil/dispersant hydrocarbons can enhance the hydrophobic interactions between SPMs, leading to the accelerated aggregation rates.25 Such a snowballing effect continues until a dynamic steady state D
dx.doi.org/10.1021/es5042157 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
nonionic and anionic surfactants, and organic solvent (including kerosene).27 Therefore, the aforementioned oil-facilitated aggregation mechanisms remain operative due to the presence of hydrophobic tails and kerosene. Moreover, the uptake of the surfactants renders the particle surface more hydrophobic, promoting the attractive hydrophobic interactions between particles.28 Surfactants also promote the adsorption of hydrocarbons and EPS by suspended particles.6,29 Furthermore, the dispersant components are biodegradable and may serve as carbon and energy sources that stimulate the microbial activities.30 Figure 3 shows that the dispersant greatly elevated the TBN and especially EPS levels. At Day 14, TBN in Case III increased by 15.3-fold compared with the initial level, while TBN in Case I increased by only 4.3-fold; At Day 14, the EPS content in Cases III increased by 115.0-fold, compared to only 4.8- and 11.0-fold respectively for Cases I and II. Interestingly, the combination of oil and the dispersant (Case IV) did not show an additional or synergistic effect in terms of aggregation rate. In the first 3 days, N in Case IV was reduced from 925 ± 126 to 23 ± 4 with an aggregation rate of 301 d−1 (Figure 2a), which is between those in Cases II (oil only) and III (dispersant only). This is possible because the interactions between oil and the dispersant prevent either oil or the dispersant from fully exerting its individual role. Compared with Cases II and III, the flocs in Case IV were more abundant in number, larger in volume, but relatively smaller in size. After the 28-day incubation, most of the flocs in the presence of the dispersed oil showed positive or neutral buoyancy while those formed in the presence of oil alone or dispersant alone all sank (Figure 2). The different characteristics of flocs resulted from the incorporation of different types and amounts of oil components. The application of a chemical dispersant to an oil slick increases the formation of much smaller oil droplets and disperses much more oil in the water column, compared to naturally/mechanically dispersed oil.31 These fine oil droplets not only facilitate formation of more MOS flocs, but resulted in much higher oil content in the resulting MOS. On Day 28, approximately 200 small oil droplets (
