Tracking microparticle motions in three-dimensional flow around a microcubic array fabricated on the wall surface
- 709 Downloads
This paper describes the application of a microscopic defocusing image system to track 3D particle motions in microflow around a microcubic array near a microchannel wall surface. The measurement principle and calibration method were evaluated to provide accurate 3D microparticle location. Particle trajectories were measured in two microchannels. Measured velocity profiles of Hele–Shaw flows for two Reynolds numbers agreed well with theoretical profiles. Three-dimensional particle tracking in fluid flow around a microcubic array exhibited quasi-periodic and non-periodic trajectory modes. The experimental results indicated that 3D particle motions are spatially and time dependent even when the flow rate is constant, and microparticle trajectories may deviate from steady flow streamlines.
KeywordsDefocusing particle image technique Microcubic array Three-dimensional particle trajectory Quasi-periodic and non-periodic modes
Great progress in fluid diagnosing methods at micro- and nanoscales has been made in recent decades promoting developments in physics, chemistry, biology, medicine, and microelectromechanical systems (MEMS). Most previous studies have focused upon flow structure measurements in microdevices (Cierpka and Kähler 2012). Transportation, separation, and collection of solid particles, biological cells, liquid droplets, and gas bubbles have become very important in the study of microfluidic internal flows (Stone et al. 2004).
The influence of complex three-dimensional (3D) wall surface structures on internal flows at micro- or nanoscales should also be considered, and many studies have investigated the influence of wall roughness on flow field and friction characteristics (Natrajan and Christensen 2010; Gamrat et al. 2008; Wang and Wang 2007). Recent studies have indicated that hierarchical micro- or nanostructures on a solid surface can have significant effects on hydrophobic and hydrophilic characteristics, changing flow structure and friction characteristics significantly (Nosonovsk and Bhushan 2008; Ou and Rothstein 2005; Choi and Kim 2006; Lu et al. 2010). However, there are few experimental studies of wall surface roughness flow field influence (Gamrat et al. 2008); particularly, the motion of microparticles influenced by 3D wall surface microstructure has been seldom reported.
Various optical measurement techniques related to particle motions on the microscopic scale have been recently developed. Microscopic particle image velocimetry (micro-PIV) and microscopic particle tracking velocimetry (micro-PTV) have allowed detailed detection of full field microparticle motion. The micro-PIV technique, first introduced by Santiago et al. (1998), captures two or more images of moving particles and analyzes their motion using spatial correlation to infer the fluid velocity field (Wereley and Meinhart 2010). On the other hand, micro-PTV is based on single-particle position recognition to detect individual particle trajectories. Particle motion behaviors and local flow characteristics can be investigated by tracking 3D particle trajectories. To investigate the microparticle motions influenced by wall surface microstructure, 3D microparticle tracking techniques need to be employed. Up till now, several 3D microparticle tracking techniques have been developed following different principles.
Stereoscopic micro-PTV techniques measure all three velocity components of particles using two cameras. Out-of-plane particle velocity is derived from two different perspectives, i.e., at least two different views are required. Both cameras record the illuminated particles, and in-plane velocity vectors can be estimated using PTV algorithms (Bown et al. 2006). However, micro-PIV technique microscope objectives have a large numerical aperture, which limits off-axis viewing angle (Wereley and Meinhart 2010). In addition, the technique requires complex calibration (Cierpka and Kähler 2012).
Digital holographic micro-PTV applies digital holography to the microscopic domain. Particle spatial positions can be reconstructed from holographic images using the image sensor to record scattered and non-scattered light interference patterns as a hologram. Digital holography for microflows was first introduced by Yang and Chuang (2005), and Satake et al. (2006) subsequently developed a holographic particle tracking method. However, the technique requires significant computational overhead for reconstruction. In-line holography requires coherent light from opposite the camera; hence, a second optical access is necessary and particle concentration is limited (Cierpka and Kähler 2012).
Astigmatism micro-PTV breaks the axial symmetry of an optical system, allowing particle depth coding in 2D images. A cylindrical lens replaces the optical field lens, creating a deformation imaging system (Hain et al. 2009; Cierpka et al. 2011). The main advantage of this method is relatively low cost and easy application that enables full 3D and three-component (3D3C) measurement with standard microscope equipment. The influence of unwanted image aberrations can be excluded by calibration (Cierpka and Kähler 2012).
Defocusing particle image techniques use different particle image patterns to obtain the particle depth using diffraction, or fix a mask with pinholes in the optical system (Park and Kihm 2006; Yoon and Kim 2006; Pereira et al. 2007; Peterson et al. 2008). 3D particle tracking can be relatively easily realized using defocusing particle image techniques, which have been widely applied to measure particle trajectories in 3D flow structures.
Park and Kihm (2006) developed microdefocusing from its macroscopic counterpart (Willert and Gharib 1992) to track the velocity of 500-nm-diameter fluorescent particles in a 100 × 100 μm microchannel. This method tracks line-of-sight flow vectors by correlating diffraction pattern ring size variations with defocusing distances of small particle locations. Yoon and Kim (2006) used a high-speed camera (1000 fps) and a pinhole mask to track particle trajectories in 768 × 388 × 50 μm microchannels. Depth calibration was performed in a microvolume, and compensation functions were obtained. Effects of pinhole masks with different pattern sizes were also described. Pereira et al. (2007) applied this technique to measure 2-μm-diameter fluorescent particles in evaporating liquid. A microvolume 400 × 300 μm2 and 150 μm depth was mapped using an inverted microscope equipped with a 20× objective lens. Peterson et al. (2008) measured the velocity profile in a 50-μm-deep channel by manually identifying the particles and calculating the size of the outer diffraction ring. They found particles with 3 μm diameter were sufficient for later data processing. However, large uncertainty can be generated due to the short illumination time of the small particles (100 ms) and the consequential low signal-to-noise ratio (SNR) (Cierpka and Kähler 2012). Lu et al. (2008) applied the same technique to measure zebra fish embryonic heartbeat patterns. Fluorescent tracer particles (1 μm) were injected into the blood stream, and velocity fields of cardiovascular blood flow and trajectories of heart-wall motions were obtained. Lin et al. (2008) proposed an annular-aperture-based defocusing technique for 3D particle metrology from a single camera view. They showed that depth uncertainty of 23 μm could be achieved over a range of 10 mm for macroscopic systems. Nasarek (2010) measured toroidal flow in a microchannel with a depth of 50 μm. Measurement accuracy was estimated to be ± 2 pixels, which was ± 2 μm for the optical setup employed. Deviation in the z direction was determined from the mean standard deviation of the measured radii from the calibration procedure. Tien et al. (2014) developed color-coded 3D microparticle tracking velocimetry based on Tien et al. (2008) and applied it to microbackward-facing step flows. Experimental location uncertainties were less than 0.10 and 0.08 μm for in-plane and 0.82 μm for out-of-plane components, respectively. Displacement uncertainties were 0.62 and 0.63 μm for in-plane and 0.77 μm for out-of-plane components, respectively. Winer et al. (2014) developed a defocusing image technique to determine 3D locations of cell-sized particles in microscale flows, which was the first implementation of this technique for particle focusing applications.
In this study, defocusing particle image technique was employed to investigate 3D microparticle motions in the microchannels. The current paper describes the construction of a high-speed defocusing particle image system and related calibration based on Yoon and Kim (2006), and particle trajectories in two different microchannels were measured. The primary contribution of this study is to obtain 3D trajectories of particles in the flow around a microcubic array near the wall surface, identifying two distinct trajectory modes.
2 Experimental system and methods
2.1 Defocusing image technique
2.2 Experimental system
2.3 Experimental materials and methods
2.3.1 Hele–Shaw flows
Drawing the particle trajectories was time-consuming because particle density was low. Since the velocity profiles in the z direction may be expected to be symmetrical parabolic curves (Poiseuille flows), measurements were mainly implemented within the top half of the microchannel region to decrease data processing. Up to 25,000 particle images were acquired, and individual particles at different depths of the microchannel (Fig. 4) were tracked to obtain their 2D trajectories and experimental velocity profiles, which were compared with the theoretical 2D streamlines and velocity profiles for flows between parallel plates.
2.3.2 Flow around microcubic arrays
Microchannel and microcube parameters
Parameter (see Fig. 5)
For the particle tracking experiments, we employed Wiener filtering to remove image noise, and Video Spot Tracker (VST, v07.02) software to track the defocused particle image centroid. VST tracked the three points of defocused particle images on the horizontal plane (xy-plane) and recorded their coordinates simultaneously. Particle trajectories were subsequently synthetized using Eqs. (A.1)–(A.3). Particle depth was derived from the depth calibration function, Eqs. (A.4) and (A.5). Accurate horizontal coordinates were obtained by correcting the centroid using Eqs. (A.6) and (A.7).
3 Experimental results and discussion
3.1 Particle trajectories in Hele–Shaw flow
To validate the experimental particle trajectories, particle Stokes numbers, St, at low Reynolds number, are calculated as (Zhang et al. 2016)
Equation (2) indicates that St is proportional to Re (for fixed d p /d), so St ≪ 1 (Re ≪ 1). The Stokes number is the ratio of particle response time to flow response time. A particle with high St will require longer time to adjust to the flow, whereas one with low St will adjust faster (Zhang et al. 2016). Since St ≪ 1, particle flow response will be very quick, i.e., particle trajectories should agree well with streamlines.
3.2 Particle trajectories around a microcubic array fabricated on the wall
3.2.1 Particle tracking
Figure 8c compares particle trajectories with numerically computed streamlines for 3D Stokes flow around the microcubic array (Hu et al. 2003), where the fluid converges in the wider space between two elements and diverges in the narrower path, periodically. (The top and bottom streamlines in this plot seem to intersect with the cube geometry, which may be due to drawing error.) Experimental particle tracking and numerical simulation for 3D Stokes flow both exhibit the periodic wave pattern and amplitude variations. From Eqs. (1) and (2), St ≪ 1, so particle trajectories should agree well with the streamlines. Thus, we can conclude that the observed particle trajectories were consistent with that of steady Stokes flow streamlines overall.
Particle deviation from theoretical trajectories or streamlines can be attributed to several factors. Flow inertia is an important reason for particle migration across streamlines in macroscopic pipes (Segré and Silberberg 1962; Matas et al. 2004). Di Carlo et al. (2009) and Abbas et al. (2014) showed that inertial cross-streamline migration occurred at microscales. Individual spherical particles cannot exhibit cross-streamline migration under viscosity-dominated Stokes flow (Bretherton 1962; Di Carlo et al. 2009; Abbas et al. 2014). Diameters of flow-tracing particles for micro-PIV/PTV typically range from 200 nm to 2 μm, and Brownian motion may cause random thermal noise in the velocity fields (Wereley and Meinhart 2010). Thus, the flow-tracing particles can be perturbed from the stochastic fluctuations in Stokes flows, deviating from their original trajectories, particularly for strongly disturbed flow regions. Sufficiently large particles, e.g., 2 μm diameter, can effectively dampen the influence of fluid stochastic fluctuations, reducing the probability of particle deviation from streamlines. Hence, the experimental observations indicate that non-periodic trajectory modes maybe result from stochastic fluctuations in strongly disturbed flow fields, but are not the dominant mode since the flow-tracing particles were sufficiently large. Further study will implemented to determine the exact cause for streamline crossing in this particular type of flow.
3.2.2 Three-dimensional quasi-steady trajectory structure
Figure 10e shows the 3D trajectory structure, where the vertical planes were created by mirroring actual measurements. Comparing Fig. 10a, c, two trajectory modes are apparent. The quasi-periodic mode is the main trajectory mode in the region far from microcubes (− 17 to − 26 µm), whereas a non-periodic mode occurs in the deeper region (− 25 to − 45 µm) most likely due to strong disturbance from the microcubes. Comparing Fig. 10b, d, particle trajectories projections on the vertical plane show waveforms due to the microcube periodic peak and valley structure, with increased amplitude closer to the microcubes. Figure 10e shows that particle trajectories alternately converge and diverge to form 3D spatial curves embracing the microcubes, presenting a complete picture of 3D particle migrations in Stokes flow around a microcubic array on the wall surface.
In this study, a high-speed microscope defocusing particle image system based on Yoon and Kim (2006) was constructed to track microparticle motions in the flow field, and a calibration function was applied between the configuration parameters for defocusing the particle image and the particle’s out-of-plane position.
To verify the reliability of the proposed defocusing particle image tracking system and define measurement accuracy, flows between parallel plates were measured for Re = 0.025 and 0.05. The results show that particle trajectories for both flows form a group of parallel straight lines along the depth direction from the top wall of the microchannel. The measured velocity profiles showed good agreement with theoretical velocity profiles of 2D Poiseuille flows between two parallel plates.
Particle trajectories were measured for 3D flow around a microcubic array fabricated on the wall surface. The experimental results showed that particle trajectories were consistent with 3D Stokes flow streamlines. Quasi-periodic and non-periodic particle trajectory modes were observed. The former was the main trajectory mode, with the latter occurring in or close to the microcubic array, most likely due to stronger disturbance from the 3D structure.
This study was funded by the National Natural Science Foundation of China (Grants 11472261 and 11172287) and the High Level Scholarship Foundation of Jinling Institute of Technology. Their support is gratefully acknowledged.
- Hain R, Kähler CJ, Radespiel R (2009) Principles of a volumetric velocity measurement technique based on optical aberrations. In: Nitsche W, Dobriloff C (eds) Imaging measurement methods for flow analysis, Notes on numerical fluid mechanics and multidisciplinary design, vol 106. Springer, Berlin, pp 1–10Google Scholar
- Lu S, Yao ZH, Hao PF et al (2010) Drag reduction in ultrahydrophobic channels with micro-nano structured surfaces. Sci China 53:1298–1305Google Scholar
- Matas JP, Morris JF, Guazzelli É (2004) Inertial migration of rigid spherical particles in Poiseuille flow. J Fluid Mech 515:171–195Google Scholar
- Nasarek R (2010) Temperature field measurements with high spatial and temporal resolution using liquid crystal thermography and laser induced fluorescence. PhD dissertation, Technische Universität Darmstadt, GermanyGoogle Scholar
- Zhao PJ (2015) Study on the three-dimensional particle tracking velocimetry based on the defocusing microscopic image, Master dissertation, China Jiliang University, ChinaGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.