Skip navigation

Stanford University

Stanford Microfluidics Laboratory

Micron-Resolution Particle Tracking Velocimetry (mPTV)

For various particle dynamics applications, micron-resolution particle tracking velocimetry (mPTV) offers several advantages over mPIV.  For example, PTV measurements allow the study of dynamics of Brownian particles in the presence of velocity gradients, the quantification of particle-particle interactions in a flow,  and the (direct) study of electrophoretic (drift) particle motion.  In the latter case, PTV can be used to quantify particle-to-particle differences in electrophoretic drift mobilities and to quantify channels with varying electroosmotic wall mobility (Devasenathipathy et al. 2002).

PTV is the general term for image-based tracking of individual seed particles in a flow.  The technique uses custom pattern recognition, particle image analysis, and particle image tracking algorithms.  In our lab, we typically use the so-called super-resolution PIV algorithm in which we first obtain PIV measurements of the flow, then use this preliminary (PIV) data to predict probably displacements of individual particles, and finally search images for the displacement of individual particles.  We leverage a Kalman filtering algorithm for the prediction of the particle trajectory and use c2 testing for validation of particle identification (Takehara et al. 2000).  After the PIV step, the velocity of each particle is interpolated from the measured velocity field, and this velocity is used as a predictor for the particle information in the next image via Kalman filtering. The c2 statistical validation uses information of particle location, previous particle velocity, particle image size, and particle image intensity to validate particle image pairs.  An iterative procedure is implemented until the number of particle pairs matched is maximized.  The goal is a robust individual particle tracking algorithm with a minimal number of user inputs.  The technique is not as robust as PIV but offers unique information in some cases.   
 

Optical setup
The particle tracking system uses epifluorescent microscopy, pulsed-laser illumination, and CCD imaging to capture particle images.  The imaging system is identical to the system we use for micron-resolution particle image velocimetry (mPIV).   mPTV is achieved using specialized image interrogation algorithms to provide micro-scale displacement measurements for individual particles.

Figure 1 Optical setup for mPIV or mPTV (a) Optics including (i) dual-head, doubled Nd:YAG laser,  (ii) syringe pump used to pump liquids, (iii) timing electronics for control of laser and camera,  (iv) oil immersion objective, (v) interline-transfer CCD camera, (vi) liquid-filled optical fiber, (vii) fluorescence excitation filter, (viii) dichroic beamsplitter, (ix) emission color filter, (x) computer with frame grabber, (xi) microfluidic test section, (xii) demagnifying lens (for larger field of view).


Sample particle velocity field quantification

Below is an example application of mPTV.  As described above, we used the super-resolution algorithm to apply  mPIV or mPTV respectively to a pressure driven flow through a microchannel intersection. 

 

 

(a)

(b)

Figure 2.  Schematic of intersection region of a cross-channel system   (a) Schematic of intersection showing the field of view for the experiments shown below, and  (b) a three-dimensional representation of the geometry for a numerical model. Drawing shows the D-shaped cross-section characteristic of isotropic etching.  The channel top width is 120mm and a 50 mm depth at the centerline.

 

(a)

MATLAB Handle Graphics

(b)

Figure 3  Particle tracking velocimetry measurements for 500 nm diameter particles in a pressure driven flow through a cross-channel interaction.  (a) Velocity field obtained by applying a combined PIV and PTV analysis to particle images, (b) individual velocity vectors in an exploded view of sub-region shown in part a. The maximum particle velocity is approximately 500 um/s.

 

Figure 3 shows the result for the data images described above using the combined PIV/PTV analysis.  The velocity field obtained with the PIV/PTV analysis is unstructured, given the random position of validated particle displacements inside the flow field. The number of vectors obtained by PIV is 480 in the test section while the number of vectors identified by the super-resolution-Kalman filtering c2 method is 5200. Note that the vectors from the PTV analysis are individual particle displacements, randomly positioned in the flow field, and include the Brownian motion component. 


References

1.) Devasenathipathy, S., Santiago, J.G., Meinhart, C.D., and Wereley, S.T., and Takehara, K., "Particle Tracking Techniques for Microfabricated Fluidic Systems," Experiments in Fluids; April 2003; vol.34, no.4, p.504-14

2.) Keane RD; Adrian RJ; Zhang Y. (1995) Super-Resolution Particle Imaging        Velocimetry. Measurement Science Technology 6: pp. 754-768.

3.) Takehara K; Adrian RJ; Etoh GT; Christensen KT (2000) A Kalman tracker for super-resolution PIV. Exp Fluids 29: pp. S34-S41.