Bacteria forming drag-increasing streamers on a drop implicates complementary fates of rising deep-sea oil droplets

Competing time scales involved in rapid rising micro-droplets in comparison to substantially slower biodegradation processes at oil-water interfaces highlights a perplexing question: how do biotic processes occur and alter the fates of oil micro-droplets (<500 μm) in the 400 m thick Deepwater Horizon deep-sea plume? For instance, a 200 μm droplet traverses the plume in ~48 h, while known biodegradation processes require weeks to complete. Using a microfluidic platform allowing microcosm observations of a droplet passing through a bacterial suspension at ecologically relevant length and time scales, we discover that within minutes bacteria attach onto an oil droplet and extrude polymeric streamers that rapidly bundle into an elongated aggregate, drastically increasing drag that consequently slows droplet rising velocity. Results provide a key mechanism bridging competing scales and establish a potential pathway to biodegradation and sedimentations as well as substantially alter physical transport of droplets during a deep-sea oil spill with dispersant.

where 〈 〉 denotes ensemble averaging performed over trajectories of the same species. The asymptotic value of , as → ∞, yields the Fickian diffusion coefficient.
Chemotaxis. Pseudomonas P62 chemotaxis (or lack thereof) near a crude oil-water was studied using a straight 200 m deep and 1 cm wide microchannel wherein half of the channel (500 m thick) is crude oil and the other is bacteria culture. This produces a quiescent rectilinear oil-water interface at which DHM is used to measure bacterial swimming trajectories. Figure S2 shows the probability density functions (PDF) of swimming velocities within 300 m of the oil-water interface. Results at 0, 1 and 4 h after exposure to the oil interface are shown. The data indicate that the mean velocities in the culture near the crude oil show no bias towards or away from the interface. Therefore, we conclude that Pseudomonas P62 exhibits no chemotactic behaviors towards crude oil.

S.2.1 Estimation of momentum budgets.
To estimate the drag on the droplet, we estimate the 2D momentum balance around a droplet. The 2-D momentum balance at the imaging plane (the mid-plane) of the microfluidic channel can be expressed as where ⃗ is mean flow velocity, ∇ ⃗ is the gradient operator, and are the density and kinematic viscosity of the surrounding fluids respectively. For the flow around a micro droplet with Re << 1 and a steady rising velocity, the unsteady term (1 st term) n Eq. S1 can be neglected. The steady momentum balance can be expressed as the following: where and are x-(streamwise) and y-(spanwise) components of mean fluid velocity around the droplet. To scale our results later to other droplet size, we express the momentum equation in their dimensionless forms: * * * *

( S 6 )
It is worth to point out that the relative pressure field can be obtained by integrating the above pressure gradient fields: * / * and * / * , but the absolute pressure depends on pressure along the boundary. The magnitude of pressure gradients, * / * * / * , for the first 100 minutes in our kernel experiment (E7 in   Figure S3C shows the contours of pressure gradient magnitude contours over the micrograph. It is clear that regions with elevated pressure gradient magnitude indicate unequivocally the locations of those two streamers.

S.2.2 Control volume analysis for estimating drag over a droplet
To estimate the drag force on the droplet from mean velocity measurement, a control volume analysis over x-momentum is performed. A control volume is drawn around the periphery of the velocity field as shown in Fig. S3A. In this 2D schematic, the stationary drop (gray circle) is enclosed by a control region ABCD (Fig. S3A) with x-, y-axis and normal vector, ⃗ ⃗ ⃗ where ⃗ , is the unit direction vectors for x-and y-axis. Using the control volume defined in Fig. S3A, the dimensionless steady x-momentum balance across the control region is (for brevity, Einstein index notation is used hereinafter):

S4
where "•" denotes dot product operator, ⃗ ⃗ * is a dimensionless viscous stress tensor defined by ⃗ ⃗ * ∇ ⃗ * ⃑ * ∇ ⃗ * ⃑ * , * is area of the local control surface and * the total dimensionless drag force on the drop per unit length into the paper. Evaluating every term in Eqn. S7 except for drag over the control surface, and integrating over it, the total drag force can be expressed as * , Although the pressure gradients are well resolved, absolute pressure distribution over the entire control volume is difficult to determine and depends on pressure on the boundary. However, since the relative pressure is only needed along the boundary, pressure gradients can then be integrated along the boundary, ABCD. Integration is performed using first order forward/backward finite differences. From two corners of the control volume, the resolved pressure gradient field is  another, it will become neutrally or even negatively buoyant.

S.3.2 Implications of increased residence times of oil droplets
Streamers by bacteria on a rising oil droplet will reduce its velocity and subsequently increase its residence time in the deep-sea plume. As a result processes such as Marine Oil Snow (MOS) formation can readily occur in situ in the deep-sea plume where a microbial bloom has been reported [5][6][7][8][9][10][11][12][13][14][15] . This MOS can trap oil in the plume region for extended times leading to enhanced S8 biodegradation as well as cause the oil to sink to the sea floor as sediment which is corroborated by field data [16][17][18][19][20][21] . If oil droplets from the deep-sea plume are the source of sedimenting oil and are subject to enhanced degradation, it would be expected that indicators of biodegradation such as n-C17:pristane ratios 22 should indicate enhanced degradation below the plume.
To test this hypothesis we use the BP Gulf Science Data inventory of water chemistry data from These low values suggest plume droplets have resided in the water column for relatively short times. Less than 5 km from the wellhead, larger degradation ( 0.3) is seen both above and below the plume. Here droplet transport may be from the plume upwards, from the plume downwards, and/or from the surface oil slicks downwards. However, >5 km from the wellhead, larger are primarily within or below the plume (shaded oval in Fig. S5). This suggests droplet S9 transport is primarily from the plume downward i.e. plume droplets >5 km from the wellhead underwent enhanced degradation and sedimentation. This anecdotal evidence from the field data provides additional plausibility for the mechanism of bacteria forming drag-increasing streamers directly on oil droplets in the deep-sea plume in the aftermath of the Deepwater Horizon oil spill, causing them to drastically increase their residence times in the water column and even sink to the sea floor as sediment.

Video Legends
Video S1. A video of the oil drop from Fig. 2 (Table 1)  Marker sizes additional increase with increasing degree of degradation. S18