Back Issues

Mar/Apr 2012  

Measuring flow in microdevices

FtrIssueGroup

September/October 2010 Volume 3 Issue 5

FtrAuthorGroup

By Susan Woods

Contributing Editor

(224) 225-6120

Making microstructures such as microfluidic, MEMS and biomedical devices is a challenge. So is accurately measuring fluid flow through these devices, which often incorporate channels measured in microns.

That’s where particle image velocimetry (PIV) comes in. The technology measures flow velocity in a range of applications—including microscale ones.

“We look at fluid phenomena across a range of different scales, from relatively big ones, such as in naval engineering, to tiny ones in microfluidic or pharmaceutical apps,” said Callum Gray, CEO of LaVision Inc., Ypsilanti, Mich.

Before microPIV, there was just plain PIV for larger-scale flow-velocity applications. Both processes operate in a similar manner.

PIV measures flow velocity by recording images of tracer particles in the flow. A light sheet, formed by a laser beam, is used to illuminate the particles. The light sheet pulses twice and a special CCD camera captures images of the tracer particles in two consecutive frames, providing a record of the flow velocity. Then, the distance particles travel during the time between the first and second pulse is measured, which allows flow velocity to be calculated.

“We divide the image up into a grid of interrogation spots,” Gray said. “These local spots are a small region of the image. We track the motion of local groups of particles, not individual particles.”

A typical PIV setup consists of a fast interline-transfer CCD camera, double-pulsed laser, tracer particles, a synchronizer to control the timing of the laser and camera, and software for data capture, analysis and display. A PC provides the user interface.

Example of a microPIV system from TSI Inc. Illustration courtesy TSI.

Because the flow can be fast, double-pulsed Nd-YAG lasers are typically used as the illumination source to avoid blurred images. “When you are measuring higher-speed flows—say, meters-per-second flows such as you might get with an inkjet printer head—you need a laser,” said Steve Wereley, professor of mechanical engineering at Birck Nanotechnology Center, Purdue University. “The laser is basically a flashbulb, and you need to have that flashbulb to stop or freeze that particle motion.”

According to Wereley, the laser beam is formed into a light sheet by passing it through a cylindrical lens that expands the beam in only one direction. Only particles in the sheet of light are illuminated.

Tracer particles are made of a variety of materials and range from 55nm to 50µm in size, but they can be larger or smaller. They can be hollow glass spheres, plastic spheres or fluorescent spheres. The type of particles used varies according to the type of liquid being examined. The user purchases the particles and adds them to the liquid or the particles come suspended in the liquid.

The number of particles in the flow does affect the process. “In a fluid that has a lot of scattering particles, you are going to get a lot of detail,” Gray said. “If you don’t have enough of them, flow detail is proportionally less. Fluid flow consists of a range of scales of motion. The ability to resolve these scales depends on the density of particles.”

Conventional PIV systems can be used for measuring flows in small channels, and light sheets can be as thin as 100µm. “However, for applications with a field of view less than or equal to a couple hundred microns, you would use microPIV,” Gray said.

The light sheet is one of the main differences between conventional and microPIV. “In microPIV, you have to use volume illumination,” Wereley said. “You shine the laser into your flow, but because the dimensions are so small, it is virtually impossible to form a useful light sheet. So you end up illuminating the whole flow and not just the plane at which you are looking.”

The other difference is the addition of an epifluorescent microscope equipped with a filter for particle imaging. The fluid is seeded with fluorescent tracer particles, most often fluorescent polystyrene latex spheres, which are illuminated by the light source.

The microscope lens collects the particle signal, which has a longer wavelength than the illuminating light. This signal is separated from the light by the filter and recorded by the CCD camera. “You synchronize the laser pulse with the image capture, but you are looking at a much smaller field of view,” Gray explained.

The filter excludes all the light that is scattered at the same wavelength as the laser. The raw laser beam reflections are blocked by the filter while the particles are visible and imaged to track the flow. Only the light flourescing from the particles is allowed past.

Wereley pointed out that the illumination source for microPIV does not have to be a laser. “Quite often, people who use microPIV are looking at relatively slow flows, in the millimeter-per-second range, and then you can get away with not using the laser,” he said. “Using a laser is expensive and can be dangerous, so if you can just use white light, it is much easier.”

Because the whole flow and not just the plane being looked at is illuminated, particles in front of or behind the focal plane can still be seen, but they are blurry.

Users can still attain accurate measurements by correlating observed with expected flow, Wereley said. “If you know what the flow should look like, you can measure it and look for differences between your results and what you should see theoretically,” he explained. “The bottom line is, you can do a good job with microPIV, but it takes some work to make it a very good measurement.”

MicroPIV usually is used to measure flow velocities in microchannels, typically around 100µm wide, according to Wing T. Lai, product specialist for fluid mechanics at TSI Inc., Shoreview, Minn. Applications include measuring flows in microfluidic, MEMS and biomedical devices.

Lai said many applications TSI’s products are used for relate to mixing, such as mixing one fluid coming from a 100µm channel and another coming from a 50µm channel. His company has also provided microPIV systems to a manufacturer that uses microfluidic devices to cool semiconductor chips.

Another application is measuring flows through lab-on-a-chip devices used for DNA sequencing, according to Gray of LaVision. “You have a couple of glass slides with a series of etched channels on either side,” he said. “Different samples are pumped in and different diagnostics are performed, such as illumination with UV light, which can indicate a whole range of interactions.” MicroPIV can be performed on those channels.

MicroPIV is suited for flow measurement when physical probes are impractical. “With something as small as a human hair, it is difficult to put anything inside to measure flow,” Lai said. “MicroPIV, and PIV in general, is a noninvasive technique.”

Another advantage of microPIV is it provides a spatially resolved view of the flow. “It can show you that the flow is fast in one place and slow in another,” said Wereley.

However, microPIV is not cheap—systems range in price from about $80,000 to $200,000. “You can certainly get by with just some approximation of what the flow is doing in a device,” Wereley said. “But if you are building a device and it is not working right, you may need to get a more accurate picture of what the flow looks like.” µ

About the author: Susan Woods is a contributing editor to MICROmanufacturing magazine. E-mail: susanw@jwr.com.


Contributors

LaVision Inc.
(734) 485-0913
www.lavision.de/en

TSI Inc.
(800) 874-2811
www.tsi.com

Steve Wereley Purdue University
(765) 494-5624

FtrAuthorAbout

Telephone: (224) 225-6120. E-mail:  Susan Woods.