A numerical study of turbulence seeded with light particles is presented. We analyze the statistical properties of coherent, small-scale structures by looking at the trapping events of light particles inside vortex filaments. We study the properties of particles attracting set, measuring its fractal dimension and the probability that the separation between two particles remains within the dissipative scale, even for time lapses as long as the large-scale correlation time, T L . We show how to estimate the vortex lifetime by studying the moment of inertia of bunches of particles, showing the presence of an exponential lifetime distribution, with events up to T L . © 2010 American Institute of Physics. ͓doi:10.1063/1.3431660͔ Vorticity dynamics, in general, and vortex filaments, in particular, have been the subject of many theoretical, phenomenological and experimental studies. [1][2][3][4] According to their inertia properties particles respond differently to fluctuations of the advecting-Eulerian-velocity field producing locally nonhomogeneous concentration, a phenomenon dubbed as "preferential concentration." 5,6 Thanks to their capability of strongly concentrating in high vorticity regions, very light particles ͑i.e., small bubbles in water͒ have been used to visualize small scale vortex filaments, 7,8 and to measure pressure statistics. 9 Similar phenomena, based on complex response of microscopic hydrogen particles in quantum fluids, have also been exploited recently to visualize quantized vortices. 10 In the present work, we study the statistical correlation between light-particle dynamics and the one of small scale vortex filaments. Previous numerical 11 and experimental 12 works have assessed in great detail the spatial distribution and correlation of intense vortex filaments. Here we want to focus on different properties: their temporal evolution. Thanks to the correlation with the trajectories of light particles, we measure the vortex filament lifetime distribution.Data come from a direct numerical simulation ͑DNS͒ of 3D fully periodic Navier-Stokes equations plus particles. Together with the Eulerian field, we integrated the Lagrangian evolution of particles by mean of a model of dilute, passively advected, suspensions of spherical particles 13,14In the above equations, x͑t͒ and v͑t͒ denote the particle position and velocity, respectively, p = a 2 / ͑3 ͒ is the particle response time, a is the particle radius, St= p / is the Stokes number of the particle, is the dissipative time, and  =3 f / ͑ f +2 p ͒ is related to the contrast between the density of the particle, p , and that of the fluid, f . Let us stress that in the model particles are not defined in terms of their material physical properties ͑size and density͒ but in terms of their dynamical properties ͑response time and density contrast͒ and that the formula given for p is a physical interpretation connecting the two points of view. For instance  =0 ͑limit of very heavy particles͒ with finite Stokes is then only valid assuming vanishing part...