Scanning Electron Microscopy for Quantitative Small and Large Deformation Measurements Part I: SEM Imaging at Magnifications from 200 to 10,000

A series of baseline displacement measurements have been obtained using 2D Digital Image Correlation (2D-DIC) and images from Scanning Electron Microscopes (SEM). Direct correlation of subsets from a reference image to subsets in a series of uncorrected images is used to identify the presence of non-stationary step-changes in the measured displacements. Using image time integration and recently developed approaches to correct residual drift and spatial distortions in recorded images, results clearly indicate that the corrected SEM images can be used to extract deformations with displacement accuracy of ±0.02 pixels (1 nm at magnification of 10,000) and mean value strain measurements that are consistent with independent estimates and have point-to-point strain variability of ±1.5 × 10⁻⁴.

Variational optical flow estimation for particle image velocimetry

We introduce a novel class of algorithms for evaluating PIV image pairs. The mathematical basis is a continuous variational formulation for globally estimating the optical flow vector fields over the whole image. This class of approaches has been known in the field of image processing and computer vision for more than two decades but apparently has not been applied to PIV image pairs so far. We pay particular attention to a multi-scale representation of the image data so as to cope with the quite specific signal structure of particle image pairs. The experimental evaluation shows that a prototypical variational approach competes in noisy real-world scenarios with three alternative approaches especially designed for PIV-sequence evaluation. We outline the potential of the variational method for further developments.The publications of the CVGPR Group are listed under .

---

Microscale Synthetic Schlieren

We develop the axisymmetric Synthetic Schlieren technique to study the wake of a microscale sphere settling through a density stratification. A video-microscope was used to magnify and image apparent displacements of a micron-sized random-dot pattern. Due to the nature of the wake, density gradient perturbations in the horizontal greatly exceed those in the vertical, requiring modification of previously developed axisymmetric techniques. We present results for 780 and 383 μm spheres, and describe the limiting role of noise in the system for a 157 μm sphere. This technique can be instrumental in understanding a range of ecological and environmental oceanic processes on the microscale.

---

In vivo blood flow and wall shear stress measurements in the vitelline network

The wall shear stress plays a key role in the interaction between blood flow and the surrounding tissue. To obtain quantitative information about this parameter, velocity measurements are required with sufficient spatial (and temporal) resolution. We present a methodology for the determination of the wall shear stress in vivo in the vitelline network of a chick embryo. Velocity data is obtained by microscopic particle image velocimetry using correlation ensemble averaging; the latter is used to increase the signal-to-noise ratio of the measurements. The temporal evolution of the pulsatile flow is reconstructed by sorting the image pairs based on a phase estimate. From these flow measurements, the wall shear stress can be derived either directly from the magnitude of the gradients or from fits to velocity profiles. Both methods give results that are in good agreement with each other, while the former method is significantly easier to implement. For more accurate studies, the full three-dimensional velocity field may be required. It is demonstrated how this velocity field can be obtained by scanning the measurement volume.

---

Recent experimental probes of shear banding

Recent experimental techniques used to investigate shear banding are reviewed. After recalling the rheological signature of shear-banded flows, we summarize the various tools for measuring locally the microstructure and the velocity field under shear. Local velocity measurements using dynamic light scattering and ultrasound are emphasized. A few results are extracted from current works to illustrate open questions and directions for future research.

Limitations of accuracy in PIV due to individual variations of particle image intensities

The effect of independent variations of the intensity of individual tracer particles between consecutive images on the accuracy of common displacement estimation methods in particle image velocimetry (PIV) is investigated. Such variations can be observed, e.g., in flows with components perpendicular to the illumination sheet, leading to out-of-plane displacements of the tracer particles. The achievable accuracy of PIV measurements is shown to be limited by this effect alone to be of the order of 0.1 pixel, yielding a basic limitation of the PIV technique.

---

Analysis of velocity interpolation schemes for image deformation methods in PIV

The spatial resolution of PIV can be increased significantly by using an image deformation method (IDM) and very small grid distance (i.e. the final distance between vectors), therefore, also increasing the processing time. By using an interpolation scheme with a good spectral response, in the dense predictor step of the algorithm, it is possible to increase the grid distance without decreasing the spatial resolution therefore decreasing the total processing time.

---

Turbulence measurements in a swirling pipe flow

This paper reports laser-Doppler measurements of the mean flow and turbulence stresses in a swirling pipe flow. Experiments were carried out under well-controlled laboratory conditions in a refractive index-matched pipe flow facility. The results show pronounced asymmetry in mean and fluctuating quantities during the downstream decay of the swirl. Experimental data reveal that the swirl significantly modifies the anisotropy of turbulence and that it can induce explosive growth of the turbulent kinetic energy during its decay. Anisotropy invariant mapping of the turbulent stresses shows that the additional flow deformation imposed by initially strong swirling motion forces turbulence in the core region to tend towards the isotropic two-component state. When turbulence reaches this limiting state it induces rapid production of turbulent kinetic energy during the swirl decay.

---

Full-field Thermal Deformation Measurements in a Scanning Electron Microscope by 2D Digital Image Correlation

Using recently developed methods for application of a nano-scale random pattern having high contrast during SEM imaging, baseline full-field thermal deformation experiments have been performed successfully in an FEI Quanta SEM using 2D-DIC methods. Employing a specially redesigned commercial heating plate and control system, with modified specimen attachment procedures to minimize unwanted image motions, recently developed distortion correction procedures were shown to be effective in removing both drift and spatial distortion fields under thermal heating. 2D-DIC results from heating experiments up to 125°C on an aluminum specimen indicate that (a) the fully corrected displacement components have nearly random variability and a standard deviation of 0.02 pixels (≈25 nm at 200× and ≈0.5 nm at 10,000×) in each displacement component and (b) the unbiased measured strain fields have a standard deviation ≈150 × 10⁻⁶ and a mean value that is in good agreement with independent measurements, confirming that the SEM-DIC based method can be used for both micro-scale and nano-scale thermal strain measurements.

An application of Gappy POD

An iterative procedure, based on the proper orthogonal decomposition (POD), first proposed by Everson and Sirovich (J Opt Soc Am A 12(8):1657–1664, 1995) is applied to marred particle image velocimetry (PIV) data of shallow rectangular cavity flow at Mach 0.19, 0.28, 0.38, and 0.55. The procedure estimates the POD modes while simultaneously estimating the missing vectors in the PIV data. The results demonstrate that the absolute difference between the repaired vectors and the original PIV data approaches the experimental uncertainty as the number of included POD modes is increased. The estimation of the dominant POD modes is also shown to converge by examining the subspace spanned by the POD eigenfunctions.

---

Two- and three-dimensional flows in nearly rectangular cavities driven by collinear motion of two facing walls

Two- and three-dimensional flows in nearly cuboidal cavities are investigated experimentally. A tight cavity is formed in the gap between two long and parallel cylinders of large radii by adding rigid top, bottom, and end walls. The cross-section perpendicular to the axes of the cylinders is nearly rectangular with aspect ratio Γ. The axial aspect ratio Λ > 10 is large to suppress end-wall effects. The fluid motion is driven by independent and steady rotation of the cylinders about their axes which defines two Reynolds numbers Re _1,2. Stability boundaries of the nearly two-dimensional steady flow have been determined as functions of Re _1,2 for Γ = 0.76 and Γ = 1. Up to six different three-dimensional supercritical modes have been identified. The critical thresholds for the onset of most of the three-dimensional modes, three of which have been observed for the first time, agree well with corresponding linear-stability calculations. Particular attention is paid to the flow for Γ = 1 under symmetric and parallel wall motion. In that case the basic flow consists of two mirror symmetric counter-rotating parallel vortices. They become modulated in span-wise direction as the driving increases. Detailed LDV measurements of the supercritical three-dimensional velocity field and the bifurcation show an excellent agreement with numerical simulations.

---

Time-resolved reconstruction of the full velocity field around a dynamically-scaled flapping wing

The understanding of the physics of flapping flight has long been limited due to the obvious experimental difficulties in studying the flow field around real insects. In this study the time-dependent three-dimensional velocity field around a flapping wing was measured quantitatively for the first time. This was done using a dynamically-scaled wing moving in mineral oil in a pattern based on the kinematics obtained from real insects. The periodic flow is very reproducible, due to the relatively low Reynolds number and precise control of the wing. This repeatability was used to reconstruct the full evolving flow field around the wing from separate stereoscopic particle image velocimetry measurements for a number of spanwise planes and time steps. Typical results for two cases (an impulsive start and a simplified flapping pattern) are reported. Visualizations of the obtained data confirm the general picture of the leading-edge vortex that has been reported in recent publications, but allow a refinement of the detailed structure: rather than a single strand of vorticity, we find a stable pair of counter-rotating structures. We show that the data can also be used for quantitative studies, such as lift and drag prediction.

---

A distributed parameter integral model for predicting the thermal performance characteristics of single- and multi-wall radiation heat shield systems

Heat shielding has become an increasingly necessary means for protecting temperature-sensitive components from direct exposure to thermal radiation from high temperature sources. A simple but comprehensive distributed parameter integral model has been developed for predicting the temperature distribution of the shield and the protected component for a variety of heat shield systems. The integral model presented here is seen to be more accurate than lumped models, and can be computed with much greater speed than that required for numerical models.

---

Dual-plane PIV technique to determine the complete velocity gradient tensor in a turbulent boundary layer

Simultaneous dual-plane PIV experiments, which utilized three cameras to measure velocity components in two differentially separated planes, were performed in streamwise-spanwise planes in the log region of a turbulent boundary layer at a moderate Reynolds number (Re 1100). Stereoscopic data were obtained in one plane with two cameras, and standard PIV data were obtained in the other with a single camera. The scattered light from the two planes was separated onto respective cameras by using orthogonal polarizations. The acquired datasets were used in tandem with continuity to compute all 9 velocity gradients, the complete vorticity vector and other invariant quantities. These derived quantities were employed to analyze and interpret the structural characteristics and features of the boundary layer. Sample results of the vorticity vector are consistent with the presence of hairpin-shaped vortices inclined downstream along the streamwise direction. These vortices envelop low speed zones and generate Reynolds shear stress that enhances turbulence production. Computation of inclination angles of individual eddy cores using the vorticity vector suggests that the most probable inclination angle is 35° to the streamwise-spanwise plane with a resulting projected eddy inclination of 43° in the streamwise-wall-normal plane.

---

Turbulence statistics from optical whole-field measurements in particle-laden turbulence

Obtaining turbulence statistics in particle-laden flows using optical whole-field measurements is complicated due to the inevitable data loss. The effects of this data loss are first studied using synthetic data and it is shown that the interpolation of missing data leads to biased results for the turbulence spectrum and its derived quantities. It is also shown that the use of overlapping interrogation regions in images with a low image density can lead to biased results due to oversampling. The slotting method is introduced for the processing of particle image velocimetry (PIV) data fields with missing data. Next to this, it is extended to handle unstructured data. Using experimental data obtained by a dual-camera PIV/PTV (particle tracking velocimetry) system in particle-laden grid turbulence, the performance of the new approach is studied. Some preliminary two-phase results are presented to indicate the significant improvement in the statistics, as well as to demonstrate the unique capabilities of the system.An erratum to this article can be found at

---

XPIV–Multi-plane stereoscopic particle image velocimetry

We introduce the three-dimensional measurement technique (XPIV) based on a Particle Image Velocimetry (PIV) system. The technique provides three-dimensional and statistically significant velocity data. The main principle of the technique lies in the combination of defocus, stereoscopic and multi-plane illumination concepts. Preliminary results of the turbulent boundary layer in a flume are presented. The quality of the velocity data is evaluated by using the velocity profiles and relative turbulent intensity of the boundary layer. The analysis indicates that the XPIV is a reliable experimental tool for three-dimensional fluid velocity measurements.More information at:

---

A study of a 3-D double backward-facing step

An investigation of the flow over a three-dimensional (3-D) double backward-facing step is presented using a combination of both quantitative measurements from a particle image velocimetry (PIV) system and qualitative oil-flow visualizations. The arrangement of the PIV instrument allows for snap-shots of the (x, y) and (y, z) planes at various axial and spanwise positions. The measurements illustrate characteristics that are found in both two-dimensional (2-D) backward-facing steps and 3-D flows around wall mounted cubes. In particular, the development of a horseshoe vortex is found after each step alongside other vortical motions introduced by the geometry of the model. Large turbulence levels are found to be confined to a region in the center of the backstep; their mean square levels being much larger than what has been observed in 2-D backward-facing steps. The large turbulent fluctuations are attributed to a quasi-periodic shedding of the horseshoe vortex as it continuously draws energy from the spiral nodes of separation, which form to create the base of the horseshoe vortex. A combination of effects including the shedding of the first horseshoe vortex, the horizontal entrainment of air and the presence of two counter rotating vortices initiated at reattachment, are shown to cause the steering vector of the flow to jettison away from the surface in the first redeveloping region and along the center at z/h = 0. Oil-flow visualizations confirm these observations.

---

Assessment of camera models for use in planar velocimetry calibration

The performance of three implementations of pinhole-based camera models for use in common light-sheet imaging arrangements is investigated on the background of application to particle image velocimetry (PIV) and Doppler global velocimetry (DGV). Calibration data obtained from translated planar calibration targets was found to yield camera attitude within 0.1° on four different test cases with object distance varying as little as 2% depending on the choice of camera model. Camera calibration using data from a single image of coplanar points is considered a viable alternative to manual triangulation of camera positions but is restricted to off-normal viewing directions.

Comment on the Clauser chart method for determining the friction velocity

A known difficulty with using the Clauser chart method to determine the friction velocity in wall bounded flows is that it assumes, a priori, a logarithmic law for the mean velocity profile. Using both experimental and DNS data in the literature, this note explicitly shows how friction velocities obtained using the Clauser chart method can potentially mask subtle Reynolds-number-dependent behavior.

---

Attenuation technique for measuring sediment displacement levels

A technique for obtaining accurate, high (spatial) resolution measurements of sediment redeposition levels is described. In certain regimes, the method may also be employed to provide measurements of sediment layer thickness as a function of time. The method uses a uniform light source placed beneath the layer, consisting of transparent particles, so that the intensity of light at a point on the surface of the layer can be related to the depth of particles at that point. A set of experiments, using the impact of a vortex ring with a glass ballotini particle layer as the resuspension mechanism, are described to test and illustrate the technique.

---