Nanoparticle flow velocimetry with image phase correlation for confocal laser scanning microscopy

Abstract

We present a new particle image correlation technique for resolving nanoparticle flow velocity using confocal laser scanning microscopy (CLSM). The two primary issues that complicate nanoparticle scanning laser image correlation (SLIC)–based velocimetry are (1) the use of diffusion-dominated nanoparticles as flow tracers, which introduce a random decorrelating error into the velocity estimate, and (2) the effects of the scanning laser image acquisition, which introduces a bias error. To date, no study has quantified these errors or demonstrated a means to deal with them in SLIC velocimetry. In this work, we build upon the robust phase correlation (RPC) and existing methods of SLIC to quantify and mitigate these errors. First, we implement an ensemble RPC instead of using an ensemble standard cross-correlation, and develop a SLIC optimal filter that maximizes the correlation strength in order to reliably and accurately detect the correlation peak representing the most probable average displacement of the nanoparticles. Secondly, we developed an analytical model of the SLIC measurement bias error due to image scanning of diffusion-dominated tracer particles. We show that the bias error depends only on the ratio of the mean velocity of the tracer particles to that of the laser scanner and we use this model to correct the induced errors. We validated our technique using synthetic images and experimentally obtained SLIC images of nanoparticle flow through a micro-channel. Our technique reduced the error by up to a factor of ten compared to other SLIC algorithms for the images tested in this study. Moreover, our optimized RPC filter reduces the number of image pairs required for the convergence of the ensemble correlation by two orders of magnitude compared to the standard cross correlation. This feature has broader implications to ensemble correlation methods and should be further explored in depth in the future.