Microfluidics - Introduction to BioMEMS - page 140

between the inlet and the outlet. heir main pitfall is that water evaporation (in itself
diicult to control) changes the local osmolarity of the solution and raises the local
reagent concentration, which may be unsuitable for some assays.
Electro-osmotic pumps , which are devices that produce electro-osmotic low (see
Section 3.3.2).
Electrowetting has been used to control the displacement of droplets on dielectrics
(see Section 3.3.4).
3.6.3 Flow Visualization
Computation luid dynamics is a very powerful tool for planning the performance of a microlu-
idic device, but reality oten ofers us little surprises; nothing can substitute for observing low
behavior in real devices. Fortunately, there are methods for visualizing low in microdevices,
most of which, logically enough, involve microscopes . Most of the materials used for microluid-
ics are already transparent in the visible range, with the obvious exception of silicon, and even
that can be made transparent. (Silicon transmits adequately in the near-infrared.) In general,
glass or polymeric materials allow easy imaging. Bubbles are one thing that must be found, as
they afect low so strongly, and fortunately they are easy to see with many imaging techniques.
It is possible to use imaging to monitor (steady and unsteady) low rates through the channel
in several ways. he most common is particle imaging velocimetry ( PIV ), a method introduced
in 1998 by Juan Santiago and colleagues, in which small imageable particles are introduced
into the lowing luid. he rate of their movement is then measured using one or more images,
from which the rate of movement of the luid that surrounds them can be inferred. his method
works very well as long as the particles are not close to walls, which can collide with the particles
causing them to move slower than the surrounding luid. Of course, conventional microscopies
force the imaging of multiple depths at the same time, making it diicult to resolve low rates
along the imaging axis. he depth of ield of the microscope objective lens also has a large efect
on the precision with which the measurement can be made. To avoid this problem, it is possible
to illuminate speciic portions of the channel with a laser, or to use confocal microscopy to
image particles at speciic three-dimensional locations.
Imaging, particularly luorescence imaging, has been used to great advantage to monitor
mixing between lowing streams. As one component difuses into another stream, the luores-
cence spreads to ill the volume, so the two- or three-dimensional pattern can be used to deter-
mine the difusion coeicient of the imaged compound, or, if it is already known, the eiciency
of the mixing in the particular device.
Flow gauges can be implemented within the microchannels to directly obtain a measure of
the low velocity at one point in the channel. In 2007, Frank Vollmer's group from Harvard
University integrated an optical iber cantilever inside (across) a microchannel such that its
bending by the action of the low was a function of the low rate ( Figure 3.27 ). Later, we will see
examples of microfabricated low gauges (Section 3.8.5).
3.7 Droplet Microluidics
Droplets are extremely interesting containers: they allow for the coninement and manipulation
of (bio)chemical and (bio)physical processes in well-deined volumes ranging from nanoliters
to femtoliters and, for the same price , in a stream format (i.e., making one droplet costs virtu-
ally the same as making millions of them). hese two powerful characteristics of droplets has
attracted many scientists to this young ield, now dubbed droplet microluidics , which is now a
mature area of research covered by several good review articles.
here are three basic platforms for shuttling water droplets around, one based on electrowet-
ting, one based on oil-illed microluidic channels, and one based on air-illed microluidic