Recent Advances in Micro-PIV - 2009 PDF

ANRV400-FL42-23
ARI
13 November 2009
18:24
Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

by University of California - Santa Barbara on 07/06/12. For personal use only.
Recent Advances
in Micro-Particle
Image Velocimetry
Steven T. Wereley1 and Carl D. Meinhart2
1
School of Mechanical Engineering, Purdue University, West Lafayette, Indiana 47907;

email: Wereley@purdue.edu
Department of Mechanical Engineering, University of California, Santa Barbara,

California 93106; email: meinhart@engineering.ucsb.edu
Annu. Rev. Fluid Mech. 2010. 42:55776
Key Words
First published online as a Review in Advance on

September 21, 2009
PIV, microuidics, Brownian motion, microscope, optical diagnostics
The Annual Review of Fluid Mechanics is online at

uid.annualreviews.org
Abstract
This articles doi:

10.1146/annurev-uid-121108-145427
c 2010 by Annual Reviews.
Copyright
All rights reserved
0066-4189/10/0115-0557$20.00
Microuidic devices are becoming increasingly common and are seen in applications ranging from biology to nanotechnology and manufacturing. Flow
behavior in these small domains can often be counterintuitive because of the
low Reynolds number or the relative importance of surface forces. Microparticle image velocimetry (PIV) is a quantitative method that can be used
to characterize the performance of such microuidic systems with spatial
resolutions better than one micron. Illustrating the impact of this measurement technique, more than 100 journal articles are published per year that
feature PIV velocity measurements. This article discusses the fundamentals
of the technique, its theoretical background, and several applications.
557
ANRV400-FL42-23
ARI
13 November 2009
18:24
1. INTRODUCTION

Micron-resolution particle image velocimetry (PIV, also known as micro-PIV) refers to the
measurement technique wherein uid motion is measured in a spatially resolved manner with
resolved length scales ranging from 104 m to 107 m. Flow-tracing particles, either articially
added or naturally occurring, are used to make the motion of the uid observable. Two or more
images of the moving particles are captured and analyzed using spatial correlation methods to
infer the uids velocity eld from the particle motion. This measurement technique borrows
from the macroscopic particle image velocimetry (PIV) methodextensive reviews of the PIV
technique are given by Adrian (1991, 2005) and Raffel et al. (2007)but it has considerably
different optical and mechanical constraints and is thus regarded as a separate technique. Before
the development of spatially resolved, microscopic ow measurement techniques, such as PIV,
most experiments conducted in microscale geometries consisted of bulk ow measurements, such
as ow rate, pressure drop, and thrust (for a review, see Gad-el-Hak 1999 and Ho & Tai 1998).
Efforts toward developing PIV can be traced to the particle-streak velocimetry experiments
described by Taylor & Yeung (1993) and Brody et al. (1996), among others. They used epiuorescent microscopy and long exposure times to record streak images of moving submicron
uorescent particles, from which the uid velocity eld was inferred, albeit in an irregularly
spaced and lowspatial resolution manner. The PIV technique commonly used in microuidics
today was rst introduced by Santiago et al. (1998). They used epi-uorescent illumination to
record discrete particle images (i.e., not streaked) of uorescent particles and correlation methods
to calculate the speed of the particles, and thus the uid, with spatial resolutions of less than
10 m. Meinhart et al. (1999) followed up this work by using a similar technique to measure the
ow in a microchannel with spatial resolutions less than a single micron and a total measurement
volume size of 13.6 0.9 2.2 m (27 m3 or 27 femtoliters). Since this early work, there have
been many variations and extensions of the technique. As a result, the original PIV technique is
now commonly referred to as classical PIV. There are currently more than 100 journal articles
published per year in which PIV is the primary measurement technique, and as of November
2008, there have been cumulatively approximately 450 journal papers published (see Figure 1).
Comprehensive reviews on the subject have been published by Sinton (2004), Wereley & Meinhart
450
Cumulative citations
400
350
300
250
200
150
100
50
20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
19
99
Year
Figure 1
Cumulative number of articles, data as of November 2008.
558
Wereley
Meinhart
ANRV400-FL42-23
ARI
13 November 2009
18:24
(2005, 2007), and more recently by Meinhart & Wereley (2009). The technique has also been
described in books by Nguyen & Wereley (2006) and Raffel et al. (2007). The purpose of this
Annual Review article is to extend the previous reviews on PIV and to update the reader on
recent developments that have occurred over the past several years.
Commonly in PIV, two successive images of ow-tracing particles are recorded using a specied time delay, t. Typically, the two particle image elds are subdivided into uniformly spaced
regions, known as interrogation spots, which are cross-correlated to determine the most probable
local displacement. The local velocity of the uid, u, is then obtained by dividing the measured
displacement, x, by the time delay:
x
.
(1)
t
High spatial resolution can be obtained by recording the images of ow-tracing particles with
sufciently small diameters, dp , so that they faithfully follow the ow in microuidic devices,
which often exhibit high velocity gradients near ow boundaries. Spatial resolution can be further
improved by imaging the particles with highnumerical aperture, diffraction-limited optics and
with sufciently high magnication so that the particles are resolved with at least 34 pixels
per particle diameter. Microscope objective lenses that are used commonly in PIV range from
oil-immersion lenses (M = 60, NA = 1.4) to air-immersion lenses (M = 10, NA = 0.1).

u=
2. VOLUME ILLUMINATION
The small length scales associated with measuring uid motion at the microscale create several
challenges. To resolve these issues, one must make judicious choices involving seed particles,
particle illumination and recording, and particle-image interrogation/post-processing algorithms.
In macroscale PIV, ow-tracing particles are usually illuminated by a sheet of light, which denes
the measurement plane (Adrian 1991). At the microscale, however, it is common to use volume
illumination (Santiago et al. 1998) or evanescent illumination (Zettner & Yoda 2003).
Figure 2 is a schematic drawing of a typical epi-uorescent PIV system. A double-pulsed
laser is used to generate monochromatic Nd:YAG laser light (NewWave Research, Inc., 47613
Warm Springs Blvd., Fremont, CA). These types of laser systems are specically designed for PIV
applications and consist of two Nd:YAG laser cavities, beam-combining optics, and frequencydoubling crystals. The laser emits two pulses of light at 532 nm. The duration of each pulse is
on the order of 10 ns, and the time delay between light pulses can vary from tens of nanoseconds
to a few seconds. The illumination light is delivered to the microscope to broadly illuminate the
microuidic device.
A number of investigators have used continuouschromatic light sources (instead of pulsedlight sources), such as mercury arc lamps or halogen lamps, to provide illumination (see Santiago
et al. 1998). In this situation, an excitation lter must be used such that only a narrow wavelength
band of light illuminates the test section. If a continuous-light source is used, then a mechanical
or electro-optical shutter must be inserted into the optical path to gate the light before it reaches
the image recording device, or the image recording device must gate the light itself.
The microuidic test section must be designed with at least one optically transparent wall,
so that it can be viewed with the objective lens. A barrier lter must be positioned between the
mirror and the relay lens that passes the uorescent light from the seed particles, while ltering
out illumination light that is reected by the surface of the test section.
Typically, a sensitive large-format interline-transfer CCD (charge-coupled device) camera is
used to record the particle-image elds. A large-format CCD array is desirable because it facilitates
www.annualreviews.org Recent Advances in PIV
559
ANRV400-FL42-23
ARI
13 November 2009
18:24
Syringe pump
Fluorescent
microparticles
01.0
Flow chamber
Double-pulsed laser
Inverted microscope
Dichroic
mirror

Computer
CCD camera
Figure 2
Schematic view of a PIV apparatus using pulsed laser-light illumination into an epi-uorescent inverted microscope. The illumination
light is reected by a dichroic mirror into a owing microchannel being charged by a syringe pump. Fluorescently labeled seed particles
(commonly polystyrene) are excited by the illumination light and emit uorescent light that passes through a barrier lter and dichroic
mirror, recorded by a CCD or CMOS camera, and then digitally processed.
a large spatial dynamic range. The interline-transfer feature is desirable for two particle image
frames to be recorded within 200-ns time delays.
In volume illumination, the entire depth of the test section is illuminated by a volumetric cone
of light, and the measurement plane must then be dened by the depth of eld of the recording
lens (as shown in Figure 3). Although this is not feasible at the macroscale, it is feasible at the
microscale. Microscope objective lenses have large apertures with sharply dened objective planes,
thereby allowing particles to transition quickly from being in focus to being out of focus.
Equations describing the measurement depth for volume illumination are presented by
Meinhart et al. (2000) and Olsen & Adrian (2000). The theory for correlation depth was originally
developed by Olsen & Adrian (2000) and veried experimentally by Bourdon et al. (2004).
The f number (commonly used in macroscale imaging) can be related to NA for an innitycorrected lens following Meinhart & Wereley (2003),

1/2
1 n 0 2
f# =
1
,
(2)
2
NA
where n0 is the index of refraction of the immersion medium, and NA is the numerical aperture.
By applying the paraxial approximation, Equation 2 can be reduced to
n0
.
(3)
f p#
2NA
Bourdon et al. (2006) use the paraxial approximation to derive an expression for depth of correlation, zcorr , as

1/2
2 2
n0 d p
1
5.95 (M + 1)2 2 n40
+
.
(4)
zcorr = 2
4 NA2
16 M2 NA4
Erratum
The small parameter 0.01 is a customary value for the threshold weighting function, and
M is magnication. Equation 4 gives good agreement with experimental results for M = 20,
560
Wereley
Meinhart
ANRV400-FL42-23
ARI
13 November 2009
18:24
M = 60, NA = 1.4
Oil-immersion lens

Oil
z-scanning
Microchannel
30 m
Focal plane
2.5 m
Red blood
cell
Figure 3
Example of volume illumination, in which epi-illumination is used such that the light is transmitted through
the objective lens and an immersion medium (air or oil), thereby illuminating a volumetric cone of light in
the microchannel. The measurement plane is dened as the object focal plane. This particular schematic
shows a red blood cell adhered to the bottom of the microchannel. The object plane can be scanned in the
out-of-plane direction through movement of the objective lens.
NA = 0.7; M = 40, NA = 1.25; and M = 63, NA = 1.4 imaging systems (using 1.0-mdiameter uorescent particles) (see Bourdon et al. 2006). Surprisingly, the depth of correlation
agrees more closely to experiments when using the paraxial approximation relation between the
f number and NA (Equation 3) than when using the exact relationship given by Equation 2.
Table 1 shows typical values of the depth of correlation for common lenses.
3. EVANESCENT WAVE IMAGING

Zettner & Yoda (2003) were the rst to report the use of an evanescent eld to illuminate owtracing particles in a microchannel. An evanescent eld refers to the light that leaks through an
Table 1
Thickness of the measurement plane for typical experimental parameters, zcorr (m)
Microscope objective lens characteristics
M = 60
NA = 1.4
M = 40
NA = 0.75
M = 40
NA = 0.6
M = 20
NA = 0.5
M = 10
NA = 0.25
0.01 m
2.1
2.1
2.1
2.2
2.3
0.10 m
2.1
2.2
2.2
2.3
2.9
0.20 m
2.2
2.4
2.6
2.8
4.3
0.30 m
2.3
2.8
3.1
3.5
5.9
0.50 m
2.6
3.7
4.3
5.0
0.70 m
3.1
4.7
5.7
6.7
13
9.3
18
Particle size dp
1.00 m
3.00 m
3.9
10
6.4
18
7.9
23
27
9.4
55
561
ARI
13 November 2009
18:24
interface when a light beam in a high-refractive-index material is totally internally reected by a

low-refractive-index medium. The evanescent eld decays exponentially with distance from the
interface and is typically on the order of 200 nm thick (depending upon indices of refraction and
incident angle). This mode of illumination can be used to illuminate only those particles very near
the wall and effectively reduce the out-of-plane thickness of the measurement plane to the order
of several hundred nanometers. Typical implementations of evanescent wave illumination have
required relatively large in-plane interrogation regions ( Jin et al. 2004; Li & Yoda 2008; Sadr et al.
2004, 2005; Zettner & Yoda 2003). This is because of the relatively small number of particles in the
immediate vicinity of the wall and the high occurrence of unpaired particle images. Zettner & Yoda
(2003) initially reported a measuring volume of 40 40 0.38 m (608 m3 or 608 femtoliters),
whereas Li & Yoda (2008) reported a measuring volume of 154 24 0.2 m (740 m3 or 740
femtoliters). When designing a PIV experiment, one must decide where spatial resolution is most
important. If in-plane spatial resolution is the most important consideration, traditional volume
illumination PIV may be the better choice. However, if out-of-plane resolution near a wall is the
most important consideration, then evanescent wave illumination may be the better option.

ANRV400-FL42-23
4. FLOW-TRACING PARTICLES
The diameters of ow-tracing particles for PIV typically range from 200 nm to 2 m.
These particles are usually made out of uorescently labeled polystyrene latex (PSL). Fluorescent PSL particles are available from manufacturers such as Duke Scientic (http://www.
dukescientific.com), Bangs Laboratories (http://www.bangslabs.com), and Invitrogen (http://
probes.invitrogen.com). These particles provide good visibility for diameters down to 200 nm,
when using highnumerical aperture (NA = 1.4) oil- or water-immersion objective lenses.
Pouya et al. (2005) introduced quantum dots (QDs) as a new type of ow-tracing particle
for PIV. QDs have several advantages over uorescent polystyrene particles. They are quite
small with hydrodynamic diameters of 18 nm, do not suffer from photobleaching, and exhibit
a relatively large Stokes shift. However, the small size of QDs causes them to exhibit a high
degree of Brownian motion, i.e., have a high diffusivity. The high levels of Brownian motion
associated with QDs can be reduced by conjugating 60 QDs to a single polystyrene nanoparticle
(Freudenthal et al. 2007). This has the additional advantage of allowing independent control of
the particle diffusivity and emission color. Beyond PSL particles and QDs, there are many other
possible choices for particle typestoo many to be fully enumerated here.
Regardless of the choice of particle type, particle size must be optimized carefully such that it
is sufciently small compared to the length scale of the ow, so that it follows the ow faithfully
and does not clog the device (Sharp & Adrian 2005). The particle must also be sufciently large
so that it can be readily imaged and sufciently dampen Brownian noise.
Random thermal noise in the velocity elds can result from Brownian motion. This source of
error B is quantied by Santiago et al. (1998), relative to a displacement in the x direction, x,
as

1 2D
s 2 1/2
B =
=
,
(5)
x
u t
where the random-mean-square particle displacement is denoted by S2 and u is the local uid
velocity. The diffusivity of spherical particles, D, is estimated by the Stokes-Einstein equation
D=
562
Wereley
Meinhart
kB T
,
3 d p
(6)

ANRV400-FL42-23
ARI
13 November 2009
18:24
where kB is Boltzmanns constant, T is the absolute uid temperature, and is the dynamic
viscosity of the working uid.
For a given characteristic velocity, the error due to Brownian motion (Equation 5) establishes a
lower limit on the measurement time interval t and particle diameter dp . Importantly, the relative
Brownian intensity error decreases as the time delay between exposures increases. Larger time
intervals produce ow displacements proportional to t, whereas the root mean square of the
Brownian particle displacements grows as t1/2 . In practice, Brownian motion is an important
consideration for 50- to 500-nm particles with characteristic ow velocities of less than approximately 1 mm s1 . For lower velocities, on the order of 0.5 mm s1 , and seed particles of 500 nm
diameter, the lower limit for the time delay is approximately 100 s (assuming a maximum of 20%
error due to Brownian motion). The Brownian error can be reduced further by both averaging
over several particles in a single interrogation spot and by ensemble averaging over multiple re
alizations. The diffusive uncertainty decreases as 1/ N, where N is the total number of particles
used in the average (Bendat & Piersol 1986).
5. IMAGE INTERROGATION METHODS FOR PIV

By denition, microuidics refers to ows with relevant length scales of, say, 100 nm to 100 m.
These small length scales generally ensure that the relevant Reynolds numbers are relatively low,
and in many cases of order one or smaller. This simplies the uid mechanics analysis and renders
many problems to be quasi-steady or periodic. As a result, novel types of processing algorithms
can be used, such as ensemble correlationalso commonly referred to as correlation averaging.
Correlation averaging was originally developed to reduce the inuence of Brownian motion, low
seed particle concentration, and low-quality images.
In correlation averaging, an ensemble of particle-image pairs is collected over a specied sample
period. In the case of periodic ows, the ensemble is collected at a particular phase. Each set of
image pairs is spatially correlated. The resulting correlation functions for coincident interrogation
regions are then ensemble averaged (or phase averaged for periodic ows) to obtain an averaged
correlation function. This averaged correlation function provides a more reliable estimate of the
mean particle displacement than standard correlation functions from single image pairs. This is
especially true in situations of low particle density and high Brownian motion (Devasenathipathy
et al. 2003, Meinhart et al. 2000, Wereley & Meinhart 2005).
The main drawback of this technique is that all instantaneous information is lost. Therefore it
is not well-suited for many macroscale ows, such as turbulence studies where one is interested
in analyzing coherent structures, turbulence intensities, Reynolds stresses, etc.
The correlation averaging technique was originally demonstrated by Meinhart et al. (2000).
They compared three different averaging algorithms applied to a series of images acquired from
a steady Stokes ow of water through a 30 m 300 m glass microchannel. The signal-tonoise ratio for measurements obtained from a single pair of images was relatively low, because
there were on average only 2.5 particle images located in each 16 64 pixel interrogation window. The velocity measurements were noisy, and approximately 20% appear to be erroneous,
and it was therefore a good candidate for correlation averaging. The three different types of
averaging schemes compared were (a) image averaging, in which the images are averaged to
produce an average rst image and an average second image that were then correlated; (b) correlation eld averaging, in which the correlation function for each interrogation region is averaged across all image pairs; and (c) velocity eld averaging, in which a velocity measurement is
calculated for each interrogation region in each image pair and then averaged across all image
pairs.
563
ARI
13 November 2009
18:24
The relative performance of the three averaging algorithms was compared by varying the
number of image pairs used in each averaging technique from 1 to 20. The fraction of valid
measurements for each technique was determined by identifying the number of velocity measurements in which the streamwise velocity component deviated by more than 10% from the
known solution at each point. The fraction of valid measurements for each of the three averaging
algorithms was determined as a function of the number of realizations used in the average. The
average correlation method outperformed the other two methods and produced fewer than 0.5 to
1% erroneous measurements after averaging eight realizations (Meinhart et al. 2000). The image
averaging method produced approximately 95% reliable velocity measurements and reached a
maximum at four averages. Further increases in the number of realizations used to average the
images actually decreased the signal-to-noise ratio and produced noise in the correlation plane due
to random correlation between nonpaired particle images. The average velocity method reached
a maximum of 88% reliable measurements using only two velocity averages. Further increases
in the number of averages decreased the fraction of reliable measurements, due to the increasing
probability of encountering an erroneous velocity estimate. It should be noted that these particular
numbers will vary from measurement to measurement, but the trends are fairly universal.
Correlation averaging increases the effective particle-image density and allows the corresponding interrogation region to be reduced in size, thereby increasing spatial resolution. In principle,
this effect can be extended down to interrogation regions consisting of only a single pixel. Standard cross-correlation requires two multipixel interrogation windows to calculate a single valid
displacement vector. The rst interrogation window denes the in-plane spatial resolution of the
velocity measurement. The second interrogation window should be sufciently large and offset
appropriately from the rst window so that it contains nearly all the information contained in the
rst window. Furthermore, the windows should be sufciently small so that effects due to velocity
gradients are negligible. The spatially averaged two-dimensional (2D) correlation can be written as

ANRV400-FL42-23
c c (m, n) =
Q
P

f (x, y) g(x + m, y + n),
(7)
x=1 y=1
where P and Q denote the size of the interrogation window in the x and y direction, respectively. The image intensity functions in the rst and second windows are denoted by f and g,
respectively. The integers m and n are the displacement components in the correlation domain.
The parameters that inuence the accuracy and reliability of the measurements stem from the particle density, image quality, and interrogation window size. Enlarging the window size can increase
the number of particle images used in the correlation. However, larger interrogation regions
result in lower spatial resolutions and may produce signicant bias errors in the presence of large
velocity gradients. To mitigate this problem, researchers proposed a novel algorithm denoted as
single-pixel evaluation (SPE) (Westerweel et al. 2004) to drastically increase the spatial resolution
without decreasing SNR. The cross-correlation function from Equation 7 can be rewritten for
an interrogation region consisting of a single pixel averaged over k time steps such that
SP E,F DI (m, n) =
fk (x, y) g k (x + m, y + n),
(8)
k=1
where N is the number of image pairs used in the correlation (Gui & Wereley 2002). The
idea is to shrink the interrogation window to one pixel and collect spatial information by varying
the second pixel location within a specied radius. Based on the correlation averaging technique
described by Meinhart et al. (2000), the SNR can be increased by increasing the number of image
pairs in a given sample. Assuming a particle density of, say, 10 particles over a 32 32 pixel
window (i.e., 9.77 103 particles per pixel), at least 1024 image pairs are required in SPE to
564
Wereley
Meinhart
ANRV400-FL42-23
ARI
13 November 2009
18:24
reach the same SNR (i.e., particles per pixel) as that obtained by standard cross-correlation of a
32 32 pixel window.
The accuracy of the SPE technique can be improved further by using a more accurate nitedifference approximation. Equation 8 is a rst-order accurate forward difference approximation
(hence the FDI subscript on SPE,FDI above). This can be improved by using a second-order
accurate central difference approximation (denoted SPE,CDI ) such that
SPE,CDI (m, n) =
[ fk (x, y) g k (x + m, y + n) + g k (x, y) fk (x m, y n)],
(9)

k=1
which contains an extra term to account for the backward-time correlation. In addition to
improving the measurement accuracy, this approach provides rapid elimination of the background
noise, which results from effectively doubling the number of the image pairs. Theoretically, an
ultimate in-plane resolution of 60 nm would be attained when a 60-nm particle is imaged with an
M = 100, NA = 1.4 objective lens and a CCD camera with a 6-m pixel size.
6. 2D VELOCIMETRY
6.1. Flow Through a Rectangular Capillary
The accuracy of PIV was initially established by measuring the velocity eld in a 30 m
300 m 25 mm rectangular glass microchannel (Wilmad Industries), which has a known
analytical solution (Meinhart et al. 1999). The ow in the glass microchannel was imaged using
an inverted epi-uorescent microscope and a Nikon Plan Apochromat oil-immersion objective
lens with a magnication M = 60 and a numerical aperture NA = 1.4 (as shown in Figure 2).
The object plane was placed at approximately 7.5 1 m from the bottom of the 30-m-high
microchannel. Because deionized water (refractive index nw = 1.33) was used as the working uid,
the effective numerical aperture of the objective lens was limited to approximately NA 1.23. A
Harvard Apparatus syringe pump was used to maintain an average 200 l h1 ow rate through the
microchannel. The particle images were analyzed using a custom-written PIV package described
by Wereley et al. (1998). The following processing steps were used: (a) correlation averaging,
(b) background removal, and (c) central difference interrogation (Wereley & Meinhart 2001).
For this experiment, 20 realizations were correlation averaged to obtain a high-quality signalto-noise ratio. The rst interrogation window was chosen to be 120 8 pixels. The high aspect
ratio of the interrogation region was chosen to provide high spatial resolution in the wall-normal
direction, where there existed a large velocity gradient. The signal-to-noise ratio resulting from
these interrogation techniques was sufciently high, so that there were no erroneous velocity
measurements.
Figure 4 shows the resulting correlation-averaged velocity-vector eld. The spatial resolution,
dened by the size of the rst interrogation window, was 120 40 pixels in the region far from the
wall and 120 8 pixels near the wall. In physical units, this corresponds to a spatial resolution of
13.6 m 4.4 m and 13.6 m 0.9 m, respectively. The agreement between the experimental
results and the analytical solution is within 2% of the full-scale measurement, conrming for the
rst time that PIV can be used to measure accurately 2D velocity-vector elds in microchannels
(as reported in Meinhart et al. 1999).
6.2. Flow Around an Adherent Blood Cell

Figure 3 depicts the epi-uorescent imaging conguration used to measure ow around a hemispherical red blood cell made adherent to the bottom wall of a 30 300 m microchannel seeded
565
ANRV400-FL42-23
ARI
13 November 2009
18:24
10 mm s1
30

Spanwise position (m)
25
20
15
10
40
45
50
55
60
65
70
Streamwise position (m)

Figure 4
Velocity-vector eld measured in the near-wall region of a nominally 30 m 300 m channel. The spatial
resolution is 0.98 m per interrogation region. Figure taken from Meinhart et al. 1999.
with 200-nm uorescent polystyrene particles. The experiment is described in detail by Pommer
(2007). The x, y, and z coordinates are in the streamwise, cross-stream, and wall-normal directions, respectively. A pseudosteady ow with a mean velocity of 1.95 mm s1 (measured using
the PIV technique) was actuated using a syringe pump. The Reynolds number of the ow was
Re 0.15, making the ow well within the Stokes ow regime. A nominal 200 200 m region
of interest was imaged using an oil-impression M = 60, NA = 1.4 lens. A total of 100 image
pairs were recorded at each of four vertical z locations from the bottom microchannel wall. The
time delay, t, between image pairs was adjusted at each z location to maintain characteristic
particle displacements of approximately eight pixels.
The interrogation region size was chosen to be 24 24 pixels, with 50% overlap. The resulting
in-plane spatial resolution was 3 m and the velocity-vector spacing 1.5 m. Figure 5 shows
the averaged 2D velocity-vector eld centered at z 1.8 m from the bottom wall. From Table 1,
we estimate the characteristic thickness of the measurement plane to be z 2 m. The missing
data near the center of the vector eld result from the red blood cell occupying that region. The
uncertainty of PIV velocity vectors is estimated to be approximately 1% FS. The velocity eld is
resolved sufciently such that information about the shear stress on the cell wall can be extracted.
7. 3D VELOCIMETRY
The extension of PIV into 3D space has been achieved by a number of techniques, including
scanning multiple 2D planes, digital holography, and confocal uorescent microscopy.
7.1. Multiple 2D Planes

For steady and periodic ows, volumetric PIV information can be obtained by scanning the
objective lens in the out-of-plane direction. Using this technique, one can extend the 2D realization
566
Wereley
Meinhart
ANRV400-FL42-23
ARI
13 November 2009
18:24
Flow direction
10
y (m)

10
10
10
x (m)
Figure 5
Measurements of ow around an adherent red cell in a microchannel as shown in Figure 3. The velocityvector spacing is 0.88 m. The velocity eld was taken at z = 1.76 m above the microchannel oor (see
Figure 3). The red cell is drawn (not to scale) to show its location. Figure taken from Pommer 2007.
of ow over an adherent cell shown in Figure 5 into a series of 2D realizations at multiple outof-plane z locations, which are shown in the left-hand column of Figure 6. Figure 6 illustrates
z locations centered at approximately 1.8, 2.6, 3.5, and 4.4 m. These experimentally measured
vector elds can be directly compared to numerical simulation results, which are shown in the
center column of Figure 6 (for details, see Pommer 2007). The in-plane velocity vectors measured
experimentally agree with numerical simulation results to within a few percent.
The out-of-plane component of velocity can be estimated by integrating the continuity equation for incompressible ow ux + v
+ w
= 0, with respect to the z coordinate, such that
y
z

z2
v
u
w =
+
z + f (x, y)
x
y
z1
(10)

u
v
(z1 z2 ) + f (x, y) ,
w=
+
x
y
where f (x, y) is an arbitrary function of integration, which must be known a priori, and can be a wall
boundary condition w = 0, or a plane of symmetry. Here, we choose the wall condition w = 0 at
z = 0. Figure 6 compares the out-of-plane w component of velocity estimated by PIV measurements to numerical simulation results. The velocity proles agree to within 3% FS (Pommer 2007).
7.2. Digital Holography

Digital holography for microows was rst introduced by Yang & Chuang (2005) who used a
photopolymer plate to record stereoscopic images of ow over a backward facing step. Satake
et al. (2005, 2006) developed a holographic particle-tracking method and applied it to a 92-mdiameter micropipe. A typical reconstructed holographic image of 1-m-diameter particles is
shown in Figure 7. From these particle-image elds, 104 3D velocity vectors were recovered.
The holographic technique has been extended recently by Kim & Lee (2007).
567
ANRV400-FL42-23
ARI
13 November 2009
18:24
Uavg
z = 4.4 m
Uavg
b
w (mm s1)
y (m)
10
5
0
5
10
y = 0 m
y = 1.4 m
y = 3 m
0.1
0
0.1
10 5
x (m)
10
10 5
z = 3.5 m
x (m)
10
10
10
10
10
10
10
10
10
x (m)
w (mm s1)
y (m)

10
5
0
5
10
0.1
0
0.1
10 5
x (m)
10
10 5
x (m)
10
x (m)
z = 2.6 m
5
w (mm s1)
y (m)
10
0
5
10
0.1
0
0.1
10 5
x (m)
10
10 5
x (m)
10
x (m)
z = 1.8 m
w (mm s1)
y (m)
10
5
0
5
10
0.1
0
0.1
10 5
x (m)
10
10 5
x (m)
10
x (m)
Figure 6
A collection of 2D 2-C velocity-vector elds at various planes in the out-of-plane direction. (a, left column)
Experimental results resolved using the PIV technique. (a, right column) Velocity elds simulated
numerically. (b) Experimental and simulated out-of-plane velocity at each focal plane. Experimental results
sampled from y = 0 m (triangles), y = 1.4 m (squares), and y = 3 m (diamonds). Simulated results are
illustrated using a dashed line. Error is estimated to be = 3% of free stream velocity (1.95 mm s1 ) at
z = 4.4 m from the bottom microchannel surface (see Figure 3). Figure taken from Pommer 2007.
568
Wereley
Meinhart
ANRV400-FL42-23
ARI
13 November 2009
18:24
d
a
b
c

Figure 7
Typical reconstructed holographic images of 1-m-diameter particles. Several particle images are
highlighted by circles. Figure taken from Satake et al. 2006.
A major drawback to holographic imaging is the relatively low particle density that can be
recorded. In an effort to obtain relatively high image density, Sheng et al. (2006) were able to
achieve approximately 1000 particle images using 0.75-m- and 3.2-m-diameter particles in
measurement volumes having depths of 110 mm.
An indirect method of using holographic imaging to obtain velocity information was developed
by di Leonardo et al. (2006). In this approach, they used holographic techniques to create multiple
optical traps in a owing microchannel. These optical traps were actuated on and off to trap and
release polystyrene beads to probe the velocity at multiple locations in a microchannel ow.
7.3. Stereo Imaging

Stereo PIV allows one to measure all three components of velocity in a single 2D plane. The
stereo-imaging technique has been developed and used successfully in macroscale PIV (Arroyo &
Greated 1991, Prasad 2000, Willert 1997). Stereo-PIV hardware and software are also available
commercially.
The application of stereo imaging to PIV is considerably more complicated than its application
to macroscale PIV due to conicting requirements between the spatial resolution and accuracy of
the out-of-plane velocity component. High spatial resolution necessitates a large imaging aperture
(i.e., largenumerical aperture objective lens). However, a large numerical aperture limits the offaxis viewing angle, which is required for accurate stereo imaging.
The rst successful stereo PIV measurements were reported by Lindken et al. (2006).
They investigated the 3D ow in a T-shaped micromixer with cross-sectional dimensions of
800 200 m2 , at a Reynolds number Re = 120. A typical three-component velocity eld is
shown in Figure 8. To achieve off-axis viewing, they used low numerical apertures of NA = 0.14
and 0.28. Their reported spatial resolution was 44 44 15 m3 , with an out-of-plane distance
between successive measurement planes of 22 m.
Bown et al. (2006) compared the performance of stereo PIV to computational uid dynamics
results of ow over a micron-scale backward facing step. They reported spatial resolutions of
10 10 10 m3 . The results indicate that the accuracy of correlation-based measurements is
limited by the degree of overlap between the two stereo-imaged object elds. They suggested that
particle tracking is more accurate than correlation-based methods and that the accuracy of stereo
PIV is not affected by the slight misalignment between the two object planes.
569
ANRV400-FL42-23
ARI
13 November 2009
18:24
200
z (m)
Vz (m s1)
0.6
100
0.4
0
0.2
200 m
0
0.2

0.4
0.6
800 m
Figure 8
Stereoscopic velocity eld of a T-mixer operating at a Reynolds number Re = 120. Figure taken from Lindken et al. 2006.
7.4. Particle-Image Defocusing

The complexity of stereo imaging optics can be mitigated by employing defocusing digital particle image velocimetry (DDPIV). With DDPIV, particle images are recorded through a single collection lens and a multihole aperture (Willert & Gharib 1992). The DDPIV technique
was applied to a micron-scale backward facing step by Yoon & Kim (2006). By averaging over
2000 vector volumes, they achieved an average spatial resolution of 5 m (in plane) and 1 m (out of
plane).
Pereira et al. (2007) obtained 3D velocity measurements of an evaporating water droplet using a
400 300 150 m volume with 2-m-diameter seed particles. Figure 9 shows a reconstructed
3D trajectory (and the respective planar projections) of a single particle in the evaporating droplet.
The relatively complex uid motion in the droplet is clear.
The concept of particle-image defocusing can also be applied to single-aperture imaging systems. Extending the technique of deconvolution microscopy, Park & Kihm (2006) used a single
aperture to record images of 500-nm uorescent particles owing over a spherical obstacle. They
estimated the out-of-plane location of individual particles by comparing the particle-image diameter to the point-spread function of the recording optics.
7.5. Confocal Imaging

Confocal imaging occurs when the illumination and recording optics are focused on a single
coincidental spot. Figure 10 shows the optical path through a confocal imaging system. Confocal
imaging can decrease signicantly the depth of eld, while increasing the signal-to-noise ratio of
the image. Because confocal imaging can only image a single point at any one instance in time,
it must scan over a 2D plane to obtain a single image. The application of uorescent confocal
microscopy to PIV was rst demonstrated by Park et al. (2004). The time duration required
to scan a 2D plane places restrictions on the applicability of confocal to high-speed microuidic
570
Wereley
Meinhart
ANRV400-FL42-23
ARI
13 November 2009
18:24
yz projection
xz projection
U (mm s1)
1.0
0.8
0.6
0.10
0.4
0.2
z (mm) 0.05
0.20

0
0.1
0.15
0.05
0.10
0
y (mm)
xy projection
0.05
0.10
x (mm)
0.05
Figure 9
3D particle trajectories of ow in an evaporating droplet. The measurement volume is 400 300
150 m. Figure taken from Pereira et al. 2007.
applications. A Nipkow disk is commonly used to increase the scanning rate of confocal systems
(see Figure 10).
A comparison between standard epi-uorescence imaging and confocal imaging was presented
by Park et al. (2004) and is shown in Figure 11. Here, 200-nm-diameter uorescent particles are
imaged through a 100-m ID pipe. The image shown in Figure 11a was taken using a standard
40 objective lens. The background light decreases the quality of the in-focus particle-image
eld. By comparison, the image in Figure 11b was taken using confocal imaging. The confocal
images clearly have higher signal-to-noise ratios and provide stronger signals for the correlation
operation, as compared to the standard epi-uorescence images shown in Figure 11a.
In this particular situation, improvement in particle-image quality could also have been achieved
in the epi-uorescent imaging mode by using a highernumerical aperture objective lens, say an
M = 60, NA = 1.4 oil-immersion lens. This would have provided high-quality images, without requiring the expensive hardware associated with confocal microscopy. Furthermore, epiuorescence imaging does not suffer from the restriction of measuring relatively low-speed velocity elds, which results from the time required for confocal scanning a 2D plane.
The primary advantages of confocal microscopy are the sharp extinction of background light
from out-of-focus images and the increased spatial resolution in the out-of-plane direction. These
advantages come at the cost of the requirement that the image plane be scanned point by point.
The ability of confocal imaging to optically section static and dynamic images is a valuable
tool for microuidic diagnostics (Chao et al. 2005). The technique has been applied to study ow
of blood cells in a physiological uid by Lima et al. (2006) and used to measure simultaneously
velocity elds and pH elds by Ichiyanagi et al. (2007).
Digital microuidics is a growing eld of microuidics, in which discreet oil/water droplets are
formed and manipulated in a microuidic device. Kinoshita et al. (2007) applied confocal PIV
571
ANRV400-FL42-23
ARI
13 November 2009
18:24
Suspension
6 0x
1.25
Objective

Piezo
Nipkow disk
Dichroic
Band pass
DPSS laser
CMOS
Figure 10
Ultra-high-speed confocal imaging system with a piezo objective positioned. The system incorporates a
specially balanced Nipkow disk to achieve 2D imaging rates of 5000 frames per second. CMOS,
complementary metal oxide semiconductor. Figure courtesy of J. Posner, ASU, 2009.
to investigate a digital microuidic device in which water/glycerol droplets were surrounded by

silicone oil owed through a 100 58 m PDMS microchannel. Figure 12 shows a typical
velocity eld inside a water/glycerol droplet at 12 m. The ow is viewed in a frame of reference
traveling with the droplet. After scanning multiple planes using confocal imaging, investigators
applied the continuity equation (Equation 10) to estimate the out-of-plane velocity components,
thereby measuring the full 3D 3-C velocity eld with a spatial resolution of 9.1 9.1 2 m.
8. FUTURE DIRECTIONS AND CONCLUDING REMARKS

Although it is difcult to predict the exact future directions of PIV, it is likely that imaging
technologies will continue to improve. Illumination sources, such as lasers, will become more
reliable, obtain higher efciencies, and be further reduced in cost. In addition, CCD and CMOS
cameras will become more sensitive, and have higher framing rates and lower readout noise.
These somewhat predictable improvements in technology will provide new opportunities for
PIV measurements that would have been considered impractical just a few years ago by classical
572
Wereley
Meinhart
ANRV400-FL42-23
ARI
13 November 2009
18:24

Figure 11
Typical images of 200-nm-diameter ow-tracing particles imaged in a 100-m ID pipe using (a) standard
epi-uorescent imaging with a 40 objective lens and using (b) confocal microscopy. Figure taken from Park
et al. 2004.
PIV standards. One such example is currently being developed at Arizona State University by
Klein & Posner (2009) (see Figure 10). The system uses a specially balanced ultra-high-speed
Nipkow disk in a confocal imaging conguration. The objective lens can be scanned in the out-ofplane direction using a piezo positioner, which is capable of scanning 100-m channel depths in
1 ms. A diode pumped solid-state laser provides continuous excitation light. A high-speed CMOS
camera (Phantom V12.1) is used to record uorescent images and is capable of 6242 frames per
second in full resolution mode (1280 800 pixels) with 40% quantum efciency. This confocal
imaging system has been used to obtain 2D confocal images with frame rates approaching 5000
frames per second (Klein & Posner 2009). 3D volumes can be scanned in real time with repetition
rates of 30 Hz.
As reported above, there are several methods in the literature for measuring all three components of velocity: multiple 2D image slices, holographic imaging, stereo imaging, and particle
defocusing. Technological improvements in high-speed cameras have facilitated the development
of real-time PIV. Coupling this with improvements in objective-lens-positioning technology has
allowed researchers to obtain 2D imaging slices in real time (see Erkan et al. 2008, Klein & Posner
Moving direction
0.1 mm s1
Y
X
50 m
Figure 12
2D velocity measurements of a water droplet surrounded by oil owing through a microchannel. The
particles were imaged using confocal microscopy. Figure taken from Kinoshita et al. 2007.
573
ANRV400-FL42-23
ARI
13 November 2009
18:24
2009). Once a 3D image eld is obtained, the out-of-plane component of velocity can be estimated
using a variety of techniques: (a) integrating the continuity equation, (b) analyzing the peak in the
cross-correlation functions (Erkan et al. 2008), or (c) tracking 3D reconstructed particle images.
If technology trends continue on their current trajectory, we predict that the preferred method
for measuring three-component velocity elds within the next 5 years will be the reconstruction
of multiple 2D image slices. This technique does not require off-axis viewing, is compatible with
the highnumerical aperture imaging, and avoids the complex alignment and registration steps
associated with stereo imaging. Furthermore, the method of reconstructing multiple 2D images
is compatible with both standard 2D epi-uorescence imaging and confocal imaging.
DISCLOSURE STATEMENT
S.T.W. and C.D.M. hold two patents (6,653,651 and 7,057,198) on PIV.
ACKNOWLEDGMENTS
This work was supported in part by the Institute for Collaborative Biotechnologies through
contract number W911NF-09-D-0001 from the U.S. Army Research Ofce. The content of
the information herein does not necessarily reect the position or policy of the government and
no ofcial endorsement should be inferred.
LITERATURE CITED
Adrian RJ. 1991. Particle-imaging techniques for experimental uid mechanics. Annu. Rev. Fluid Mech. 23:261
304
Adrian RJ. 2005. 20 years of particle image velocimetry. Exp. Fluids 39:15969
Arroyo M, Greated C. 1991. Stereoscopic particle image velocimetry. Meas. Sci. Technol. 2:118186
Bendat JS, Piersol JG. 1986. Random Data: Analysis and Measurement Procedures. New York: Wiley
Bourdon CJ, Olsen MG, Gorby AD. 2004. Validation of an analytical solution for depth of correlation in
microscopic particle image velocimetry. Meas. Sci. Technol. 15:31827
Bourdon CJ, Olsen MG, Gorby AD. 2006. The depth of correlation in micro-PIV for high numerical aperture
and immersion objectives. J. Fluids Eng. 128:88386
Bown MR, Macinnes JM, Allen RWK. 2006. Three-component micro-PIV using the continuity equation and
a comparison of the performance with that of stereoscopic measurements. Exp. Fluids 42:197205
Brody JP, Yager P, Goldstein RE, Austin RH. 1996. Biotechnology at low Reynolds numbers. Biophys. J.
71:343041
Chao SH, Holl MR, Koschwanez JH, Carlson RH, Jang L-S, Meldrum DR. 2005. Velocity measurement
in microchannels with a laser confocal microscope and particle linear image velocimetry. Microuid.
Nanouid. 1(2):15560
Devasenathipathy S, Santiago JG, Wereley ST, Meinhart CD. 2003. Particle tracking techniques for microfabricated uidic systems. Exp. Fluids 34:50413
di Leonardo R, Leach J, Mushque H, Cooper J, Ruocoo G, Padgett M. 2006. Multipoint holographic optical
velocimetry in microuidic systems. Phys. Rev. Lett. 96:134502
Erkan N, Shinohara K, Someya S, Okamoto K. 2008. Three-component velocity measurement in microscale
ows using time-resolved PIV. Meas. Sci. Technol. 19:057003
Freudenthal P, Pommer M, Meinhart CD. 2007. Quantum nanospheres for sub-micron velocimetry. Exp.
Fluids 43:52533
Gad-el-Hak M. 1999. The uid mechanics of microdevices: the Freeman scholar lecture. J. Fluids Eng. 121:5
33
Gui L, Wereley ST. 2002. A correlation-based continuous window-shift technique to reduce the peak-locking
effect in digital PIV image evaluation. Exp. Fluids 32:50617
574
Wereley
Meinhart

ANRV400-FL42-23
ARI
13 November 2009
18:24
Ho CM, Tai YC. 1998. Micro-electro-mechanical systems and uid ows. Annu. Rev. Fluid Mech. 30:579612
Ichiyanagi M, Sato Y, Hishida K. 2007. Optically sliced measurement of velocity and pH distribution in
microchannel. Exp. Fluids 43:42535
Jin S, Huang P, Park J, Yoo JY, Breuer KS. 2004. Near-wall surface velocity using evanescent wave illumination.
Exp. Fluids 37:82533
Kim S, Lee SJ. 2007. Measurement of 3-D laminar ow inside a micro tube using digital holographic particle
tracking velocimetry. J. Micromech. Microeng. 17:215762
Kinoshita H, Kaneda S, Fujii T, Oshima M. 2007. Three-dimensional measurement and visualization of
internal ow of a moving droplet using confocal micro-PIV. Lab Chip 7:33846
Klein S, Posner JD. 2009. Temporally resolved velocimetry of unsteady microscale ows using confocal uPIV.
Exp. Fluids. Manuscript under review
Li HF, Yoda M. 2008. Multilayer nano-particle image velocimetry (MnPIV) in microscale Poiseuille ows.
Meas. Sci. Technol. 19:075402
Lima R, Wada S, Tsubota K, Yamaguchi T. 2006. Confocal micro-PIV measurements of three-dimensional
proles of cell suspension ow in a square microchannel. Meas. Sci. Technol. 17:797808
Lindken R, Westerweel J, Wieneke B. 2006. Stereoscopic micro particle image velocimetry. Exp. Fluids
41:16171
Meinhart CD, Wereley ST. 2003. The theory of diffraction-limited resolution in microparticle image velocimetry. Meas. Sci. Technol. 14:104753
Meinhart CD, Wereley ST. 2009. Micro particle velocimetry. In Micro/Nano Technology Systems for Bio-Medical
Applications: Microuidics, Optics and Surface Chemistry, ed. C-M Ho. New York: Oxford Univ. Press. In
press
Meinhart CD, Wereley ST, Santiago JG. 1999. PIV measurements of a microchannel ow. Exp. Fluids 27:414
19
Meinhart CD, Wereley ST, Santiago JG. 2000. A PIV algorithm for estimating time-averaged velocity elds.
J. Fluids Eng. 122:28589
Nguyen NT, Wereley ST. 2006. Fundamentals and Applications of Microuidics. Boston: Artech House. 2nd ed.
Olsen MG, Adrian RJ. 2000. Out-of-focus effects on particle image visibility and correlation in microscopic
particle image velocimetry. Exp. Fluids 29(Suppl.): S16674
Park JS, Choi CK, Kihm KD. 2004. Optically sliced micro-PIV using confocal laser scanning microscopy
(CLSM). Exp. Fluids 37:10519
Park JS, Kihm KD. 2006. Three-dimensional micro-PTV using deconvolution microscopy. Exp. Fluids
40:49199
Pereira F, Lu J, Castano-Graff E, Gharib M. 2007. Microscale 3D ow mapping with mDDPIV. Exp. Fluids
42:58999
Pommer MS. 2007. Fluid mechanic manipulations on cells. PhD diss., Univ. Calif., Santa Barbara
Pouya S, Koochesfahani M, Snee P, Bawendi M, Nocera D. 2005. Single quantum dot (QD) imaging of uid
ow near surfaces. Exp. Fluids 39:78486
Prasad A. 2000. Stereoscopic particle image velocimetry. Exp. Fluids 29:10316
Raffel M, Willert C, Wereley S, Kompenhans J. 2007. Particle Image Velocimetry: A Practical Guide. New York:
Springer
Sadr R, Li H, Yoda M. 2005. Impact of hindered Brownian diffusion on the accuracy of particle-image
velocimetry using evanescent-wave illumination. Exp. Fluids 38:9098
Sadr R, Yoda M, Zheng Z, Conlisk AT. 2004. An experimental study of electro-osmotic ow in rectangular
microchannels. J. Fluid Mech. 506:35767
Santiago JG, Wereley ST, Meinhart CD, Beebe DJ, Adrian RJ. 1998. A micro particle image velocimetry
system. Exp. Fluids 25:31619
Satake S, Kunugi T, Sato K, Ito T, Kanamori H, Taniguchi J. 2006. Measurements of 3D ow in a micropipe
via micro digital holographic particle tracking velocimetry. Meas. Sci. Technol. 17:164751
Satake S, Kunugi T, Sato K, Ito T, Taniguchi J. 2005. Three-dimensional ow tracking in a micro channel with
high time resolution using micro digital-holographic particle-tracking velocimetry. Opt. Rev. 12:44244
Sharp KV, Adrian RJ. 2005. On ow-blocking particle structures in microtubes. Microuid. Nanouids 1:37680
575
ARI
13 November 2009
18:24
Sheng J, Malkiel E, Katz J. 2006. Digital holographic microscope for measuring three-dimensional particle
distributions and motions. Appl. Opt. 45:3893901
Sinton D. 2004. Microscale ow visualization. Microuid. Nanouid. 1:221
Taylor JA, Yeung ES. 1993. Imaging of hydrodynamic and electrokinetic ow proles for capillaries. Anal.
Chem. 65:292832
Wereley ST, Meinhart CD. 2001. Adaptive second-order accurate particle image velocimetry. Exp. Fluids
31:25868
Wereley ST, Meinhart CD. 2005. Micron-resolution particle image velocimetry. In Micro- and Nano-Scale
Diagnostic Techniques, ed. K Breuer, pp. 51115. New York: Springer-Verlag
Wereley ST, Meinhart CD. 2007. Micro-PIV. In Particle Image Velocimetry: A Practical Guide, ed. M Raffel, C
Willert, S Wereley, J Kompenhans, pp. 36878. New York: Springer
Wereley ST, Santiago JG, Meinhart CD, Adrian RJ. 1998. Velocimetry for MEMS applications. Proc.
ASME/DSC Micro-Fluid. Symp. Anaheim, CA, Vol. 66
Westerweel J, Geelhoed PF, Lindken R. 2004. Single-pixel resolution ensemble correlation for mPIV applications. Exp. Fluids 37:37584
Willert C. 1997. Stereoscopic digital particle image velocimetry for application in wind tunnel ows. Meas.
Sci. Technol. 8:146579
Willert CE, Gharib M. 1992. Three-dimensional particle imaging with a single camera. Exp. Fluids 12:35358
Yang CT, Chuang HS. 2005. Measurement of a microchamber ow by using a hybrid multiplexing holographic
velocimetry. Exp. Fluids 39:38596
Yoon SY, Kim KC. 2006. 3D particle position and 3D velocity measurement in a microvolume via the defocusing concept. Meas. Sci. Technol. 17:2897905
Zettner C, Yoda M. 2003. Particle velocity eld measurements in a near-wall ow using evanescent wave
illumination. Exp. Fluids 34:11521

ANRV400-FL42-23
576
Wereley
Meinhart
AR400-FM
ARI
13 November 2009
15:33

Contents
Annual Review of
Fluid Mechanics
Volume 42, 2010
Singular Perturbation Theory: A Viscous Flow out of Gottingen

Robert E. OMalley Jr. p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 1
Dynamics of Winds and Currents Coupled to Surface Waves
Peter P. Sullivan and James C. McWilliams p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p19
Fluvial Sedimentary Patterns
G. Seminara p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p43
Shear Bands in Matter with Granularity
Peter Schall and Martin van Hecke p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p67
Slip on Superhydrophobic Surfaces
Jonathan P. Rothstein p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p89
Turbulent Dispersed Multiphase Flow
S. Balachandar and John K. Eaton p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 111
Turbidity Currents and Their Deposits
Eckart Meiburg and Ben Kneller p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 135
Measurement of the Velocity Gradient Tensor in Turbulent Flows
James M. Wallace and Petar V. Vukoslavcevic p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 157
Friction Drag Reduction of External Flows with Bubble and
Gas Injection
Steven L. Ceccio p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 183
WaveVortex Interactions in Fluids and Superuids
Oliver Buhler
p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 205
Laminar, Transitional, and Turbulent Flows in Rotor-Stator Cavities
Brian Launder, Sebastien Poncet, and Eric Serre p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 229
Scale-Dependent Models for Atmospheric Flows
Rupert Klein p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 249
Spike-Type Compressor Stall Inception, Detection, and Control
C.S. Tan, I. Day, S. Morris, and A. Wadia p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 275
vii
AR400-FM
ARI
13 November 2009
15:33
Airow and Particle Transport in the Human Respiratory System

C. Kleinstreuer and Z. Zhang p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 301
Small-Scale Properties of Turbulent Rayleigh-Benard Convection
Detlef Lohse and Ke-Qing Xia p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 335
Fluid Dynamics of Urban Atmospheres in Complex Terrain
H.J.S. Fernando p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 365
Turbulent Plumes in Nature
Andrew W. Woods p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 391
Fluid Mechanics of Microrheology

Todd M. Squires and Thomas G. Mason p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 413
Lattice-Boltzmann Method for Complex Flows
Cyrus K. Aidun and Jonathan R. Clausen p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 439
Wavelet Methods in Computational Fluid Dynamics
Kai Schneider and Oleg V. Vasilyev p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 473
Dielectric Barrier Discharge Plasma Actuators for Flow Control
Thomas C. Corke, C. Lon Enloe, and Stephen P. Wilkinson p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 505
Applications of Holography in Fluid Mechanics and Particle Dynamics
Joseph Katz and Jian Sheng p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 531
Recent Advances in Micro-Particle Image Velocimetry
Steven T. Wereley and Carl D. Meinhart p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 557
Indexes
Cumulative Index of Contributing Authors, Volumes 142 p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 577
Cumulative Index of Chapter Titles, Volumes 142 p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 585
Errata
An online log of corrections to Annual Review of Fluid Mechanics articles may be found
at http://uid.annualreviews.org/errata.shtml
viii
Contents

Recent Advances in Micro-PIV - 2009 PDF

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Recent Advances in Micro-PIV - 2009 PDF

Enviado por

Direitos autorais:

Formatos disponíveis

ANRV400-FL42-23

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

School of Mechanical Engineering, Purdue University, West Lafayette, Indiana 47907;

Department of Mechanical Engineering, University of California, Santa Barbara,

Annu. Rev. Fluid Mech. 2010. 42:55776

First published online as a Review in Advance on

PIV, microuidics, Brownian motion, microscope, optical diagnostics

The Annual Review of Fluid Mechanics is online at

This articles doi:

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

3. EVANESCENT WAVE IMAGING

www.annualreviews.org Recent Advances in PIV

interface when a light beam in a high-refractive-index material is totally internally reected by a

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

5. IMAGE INTERROGATION METHODS FOR PIV

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

f (x, y) g(x + m, y + n),

[ fk (x, y) g k (x + m, y + n) + g k (x, y) fk (x m, y n)],

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

6.2. Flow Around an Adherent Blood Cell

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Spanwise position (m)

Streamwise position (m)

7.1. Multiple 2D Planes

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

7.2. Digital Holography

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

7.3. Stereo Imaging

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

7.4. Particle-Image Defocusing

7.5. Confocal Imaging

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

to investigate a digital microuidic device in which water/glycerol droplets were surrounded by

8. FUTURE DIRECTIONS AND CONCLUDING REMARKS

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Annu. Rev. Fluid Mech. 2010.42:557-576. Downloaded from www.annualreviews.org

Singular Perturbation Theory: A Viscous Flow out of Gottingen

Airow and Particle Transport in the Human Respiratory System

Fluid Mechanics of Microrheology

Você também pode gostar