Electronic Supplementary Material (ESI) for Soft Matter. This journal is The Royal Society of Chemistry 2018 Electronic Supplementary Information for Cargo carrying bacteria at interfaces Liana Vaccari, a Mehdi Molaei, a Robert L. Leheny, b and Kathleen J. Stebe a a Chemical and Biomolecular Engineering, University of Pennsylvania, Philadelphia, PA, 19104, USA; E-mail: kstebe@seas.upenn.edu b Department of Physics and Astronomy, Johns Hopkins University, Baltimore, Maryland, 21218, USA. S1 Bacterial trajectories Bacteria at the interface were tracked to obtain the eMSD. 1 out of 10 bacterial trajectories are shown in Fig. S1. A detailed study of bacteria trajectories is the focus of ongoing work. Fig. S1 Trajectories of bacteria trapped on the interface shown as small grey sports in Fig. 1. b. One out of ten trajectories are shown. S2 Velocity distribution Fig. S2 shows the probability distribution of the speed of the mi- crobes and the particles in different populations. The data were fit to Rayleigh distribution PD(v)= 2v/v 2 0 [e (−v 2 /v 2 0 ) ]. S3 Simulated curly trajectories To guide intuition about the behavior of curly particles, con- sider a particle that follows a perfectly circular path, with n(τ )= cos θ (τ )e x + sin θ (τ )e y . For a particle moving at constant speed, n(τ ) · n(τ + t )= cos θ (t ) cos θ (t + τ )+ sin θ (t ) sin θ (t + τ )= cos θ (t ). The direction autocorrelation function can be evaluated: φ n (t )= 1 T  T 0 n(τ ) · n(t + τ )dτ = cos θ (t ). If τ osc is the time for one full rotation then θ (t )= 2π τ osc t and φ n (t )= cos 2π τ osc t . This qualitatively captures the oscillatory form of 0 5 10 15 20 25 30 v inst [μm/s] 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Probability denisty Diffusive Persistent Mixed Curly Bacteria Fig. S2 Probability distribution of the instantaneous velocity of the parti- cles in different populations, symbols, and fit using Rayleigh distribution, solid lines. the autocorrelation function for the individual curly trajectories. Applying this in equation 4 in the main text, one can determine MSD  Δr 2 (t )  = 2v 2  t 0 dτ ′  τ ′ 0 cos( 2π τ osc τ )dτ , which leads to  Δr 2 (t )  =( vτ osc π sin π τ osc t ) 2 . The curly trajectories observed in the experiment, however, are heterogeneous, each rotating at different speed and radius. Furthermore, randomizing interactions introduce noise to their circular paths. Therefore, to model the eMSD of the curly trajectories we used Monte-Carlo simulation. Using the Monte-Carlo method, we generated 10,000 simulated curly trajectories with different oscillation periods, T , and radii of curvatures, R = TV /2π , where V is the velocity of the particles along the circular paths. The probability distribution of the speed and the oscillation period are chosen to be similar to the experimental data. Each trajectory lasts for 1000 s with time step of dt = 1/60 s. The displacement at each time step, dx is the sum of the rotational displacement dx r = R cos(2π t/T ) and ran- dom displacement, dx n , with a Gaussian probability distribution Electronic Supplementary Material (ESI) for Soft Matter. This journal is © The Royal Society of Chemistry 2018