Session Q20: Experimental Techniques: Microscale and Multiphase Flows

Suppressing interfacial instability with nanoparticles

Fabrication of flexible mirrors requires deposition of relatively rigid thin films on soft compliant substrates. Due to film instability induced by substrate compression beyond a critical strain, direct deposition of a rigid film on soft substrates often suffers from wrinkling. We report a novel technique to fabricate wrinkle free rigid thin films in polymer by suppressing such instabilities with a nm layer of “jamming” nanoparticles. The thickness of the jamming layer can be used to control the wrinkle periodicity and amplitude. For demonstration, we selected Poly(dimethylsiloxane) as the compliant substrate, poly(monochloro-p-xylylene) as jammer, and various metals (e.g., Cu, Al, Ti) as film. An atomic force microscopy reviews that as the jammer thickness increases, the wrinkle wavelength increases accordingly, and when the thickness exceeds a certain threshold, wrinkles are completely suppressed. Nanoindentation measurements show that the interfacial stiffness increases by two orders of magnitude as the jammer thickness increases. On-going experiments using various elastomeric hydrogels will examine the effect of bulk substrate properties on interfacial jamming by NPs and assess its applications in distrubted wall stress (shear/normal) sensor for near wall flow investigations.   

Axial and temporal resolution of structured illumination microscopy particle tracking velocimetry

In microscale particle velocimetry, particles outside the focal plane can add noise and increase measurement error. Structured-illumination microscopy particle tracking velocimetry (SIM PTV) uses spatially varying illumination to enhance the signal from the tracer particles in the focal plane of the imaging system.  A pair of “raw” images are acquired of the flow with a phase shift in the sinusoidally varying illumination. A single PTV image is then reconstructed from these raw images via a Hilbert transform, eliminating most of the signal outside the focal plane.  Experimental SIM PTV results are presented for microchannel Poiseuille flow at flow speeds of O(0.1 mm/s), or Re = 0.1, using 3 μm tracer particles.  Improved reconstruction enables sub-pixel location of isotropic particle images. The effect of illumination spatial frequency upon axial spatial resolution (i.e., the thickness of the imaged slice) and the number of velocity samples is characterized and compared with previous theory.  Results are presented that illustrate how the axial spatial resolution of SIM depends upon the object characteristics.  Finally, the double-exposure SIM approach used here is compared with other SIM techniques using more raw images.   

Characterizing a Live Shear-Resistant (SR-) Biofilm with Realtime Particle Tracking Micro-rheology in eChip microcosm

Environmental insults play important roles in biofilm development. Contrary to conventions, recent studies reveal that biofilm can develop under severe flow shear (e.g. >10,000 s^-1) and eventually becomes resistant to shear erosion. To understand the underlying adaptive mechanism, rheology measurement of a live biofilm must be performed as the film develops in its natural environment. Lack of tools has hindered past efforts. Here, we present the development of experimental techniques that combines long-term ecology-on-a-chip (eChip) microfluidic platform and digital holographic 3D particle tracking micro-rheology (DH-PTM) to measure the rheological evolution of a live SR-biofilm. The eChip platform provides long-term exquisitely-controlled environment to SR-biofilms, while DH-PTM tracks microbes in films and subseqently estimating viscoelasticity from 3D trajectories. Concurrent PTM measurement using newly developed 1nm quantum dot (CaO⁺) as tracers will also be performed. Techniques is applied to study SR-biofilms of P. aeruginosa and its 12 mutants.   

Fast and accurate approach 3D-micro PTV localization in densely seeded flows by deep convolutional neural network

In 3D-microPTV the depth of the particle is inferred through changes in its image. Algorithms developed to determine the depth of a particle from its image are both computationally expensive and limited in the density of particles in an image which they can analyze. This, ultimately limits the temporal resolution with which 3D fluid flow can be visualized. Here we developed an end-to-end convolutional neural network to analyze densely seeded microflows. Convolutional Neural Network (CNN) has proven to be a powerful image-processing tool in computer vision applications such as pattern recognition, object localization, and object tracking. In this study, we took the advantage of CNN and presented an end-to-end model to localize sub-10nm objects in 3 dimensional. The performance of the model compared against the MLE algorithm as the gold standard using the experimental results of BBM system. As a result of the comparison, the lateral and axial localization precisions using CNN are convincingly close to the MLE algorithm results and insensitive to the background noise. Meanwhile, the computation runtime of the CNN localization model is greatly reduced and the number of particles capable of being simultaneously localized is increased.   

3D Tracking of Weak Phase Objects by Digital Holographic Microscopy

In this work, we improve the measurement accuracy of Digital Holographic Microscopy (DHM) for tracking the three-dimensional (3D) motions of Weak Phase Objects (WPO). Particles with a small size and a refractive index similar to the surrounding medium, such as bacteria and transparent particles, are classified as WPO. First, to maximize the signal-to-noise ratio of the hologram, we place the hologram plane in the middle of the sample volume. As a result, both real and virtual images are reconstructed from the hologram. We distinguish between the two images based on the axial intensity profiles of WPO, and recover the original 3D particle distribution. Second, we address the low axial localization accuracy in DHM. We find a shorter axial elongation of WPO image in the reconstructed field consisting of both scattered and incident waves, and utilize them to reduce the uncertainty of the particle position. Finally, we test and demonstrate our methods by tracking the 3D Brownian motions of micro-particles, as well as the swimming of marine bacteria. The particle concentration exceeds 5000 particles/mm³. Our method can be applied to understand bacterial movement, bacteria-wall interactions, and the mechanism of biofilm formation.   

Thickness Characterization of Crude Oil Slicks on Water

Knowledge of crude oil slick thickness is essential in determining the oil spill response, e.g., recovery, dispersant application, or in-situ burning, and in assessing the fate of the oil. In this study, we examine several methods to measure the thickness of slicks ranging from submicron to several mm. There is no single method that covers this entire thickness range. For thin films, natural fluorescence of oil slicks excited by UV light is measured by an EMCCD camera through a narrow bandwidth (5 nm) filter centered at 515 nm. Once calibration tests determine the absorption coefficient of the oil, imaging of the slick is used for measuring the spatial variation in slick thickness from the distribution of fluorescence intensity. For weathered HOOPS oil, this approach is effective for slick thicknesses up to 40 μm. For intermediate thicknesses, in the 40 to 1000 μm range, the thickness assessment is based on attenuation of a red (635 nm) laser light sheet transmitted through the oil slick and observed by the same camera. Both optical methods become ineffective for thicker slicks of opaque crude oil. Hence, above 1 mm, the slick thickness is assessed based on range-gating using a 7.5 MHz ultrasonic imaging system with a linear transducer.   

3-D Measurements of Optically Opaque Multiphase Flows Using Limited Angle X-ray Tomography

Measurement of 3-D multiphase flow using optical methods poses a significant challenge due to their opacity. Application of radiation-based methods such as X-ray tomography is challenging, mainly for high-speed multiphase flows. In addition, geometrical constraints such as placement of the X-ray source and detectors to obtain a complete set of projections makes the application of the method hard. In this study, we present a limited angle X-ray tomography system that can measure optically opaque multiphase flows. The configuration consists of an almost semi-circular stationary high speed detector array located concentrically around a pipe. The X-ray source position is varied between 0-140 degrees, around the pipe, by deflecting an electron beam onto a tungsten target. A state-of-the-art statistical algorithm capable of reconstructing limited-angle projection data with reduced artifacts is used to produce cross-sectional images of flows. The system is validated using a static and moving phantom of known material distribution. Finally, measurements of instantaneous material fraction for a rising cap bubble in glycerin is presented.   

Surface modification of particles causes anomalous spouting in a tapered spouted bed

A spouted bed is an alternative to a fluidized bed to handle coarser particles (Geldart B & D) with an added advantage of a sudden reduction in operating bed pressure drop and improving particle circulation rate. It has three different regimes as defined: spout, fountain & annulus. The spout is the dilute zone forming inside the center of the bed, above the bed surface a fountain is formed and the annulus region is defined in between the spout and vessel wall.

In a usual spouted bed, when fluid is injected at the bottom of the bed, at a particular gas velocity a cavity is formed near the gas entry point. Then with increasing gas velocity internal jet and elongated jets are formed. But in the present study, when we use color coating on the particle surface, a completely different spouting pattern is observed. At a particular gas flow rate, instead of the formation of a cavity, particles from the top of the bed are fluidized and crack propagates downwards. After that, there is a formation of a big gas bubble from the bottom, which subsequently undergoes bursting with the release of pressure. We report such phenomena for a shallow granular bed for the first time. Instantaneous particle dynamics have been characterized by high-speed imaging.   

Particle/droplet sizing by a modified miniLDV for two-phase flows

A new optical particle sizing (OPS) sensor for measurement of the speed in multi-phase flows is introduced. The OPS sensor is based on the miniLDV sensor technology. Modifications in the optics and the data analysis allows for the sensor to measure both the speed and the size of the particles, droplets, and/or bubbles. The data is sorted according to the measured size and the mean and rms velocities for the carrier fluid and for the discrete phases are calculated. The OPS does not require calibration, alignment or modifications of the probe by the user and works equally well in water with bubbles, and in air with droplets and solid particles. Data is presented demonstrating the accuracy and the efficiency of the OPS in two-phase flows.   

Not Participating

The recent application of machine learning methods has provided new insights for particle/bubble detection and segmentation in fluid mechanics research. However, the training of artificial neural networks usually requires a formidable amount of high-quality labeled data which are tedious to acquire. In this work, we propose an unsupervised approach based on Cycle-Consistent Generative Adversarial Networks (CycleGAN to generate realistic bubbly flow images from given mask images. Namely, a mapping based on a group of unpaired bubble images of interest and masks is trained to translate a given mask to a bubbly flow image with the image closely following the geometric information of the mask. Compared to previous GAN-based methods, the proposed architecture can generate more realistic bubbly flow images by considering the differences in bubble images due to the change in the experimental setting. We test the accuracy of the method using both BubGAN data and labelled experimental data. For both cases, realistic bubbly flow images closely following the geometric properties of mask images are generated. The root-mean-square geometric error for single bubble generation using BubGAN data is about 0.85%. Our method can greatly reduce the cost of labeling images and the generated images can be used as benchmark or training data for other experimental data processing algorithms. Moreover, the method is not limited to bubbly flows and can be easily adapted to other particle-laden flow applications.