Blood may indeed be thicker than water, but the old saying isn’t exact enough for manufacturers of medical diagnostic systems and other devices based on fluid mixing. Critical to having a device work the same way every time is knowing how much thicker one fluid is over another and being able to quantify fluid flow and behavior.This is particularly true when there’s no forced flow. Then only gravity and capillary forces drive the action, explained Tom Inglese, business development manager at San Diego-based Cyth Systems LLC, an integrator of test and manufacturing automation solutions. Such a fluid-flow-related problem confronted one of Cyth’s clients — a company that was beginning production on a new device. “They didn’t have any reliable way of tracking this fluid, this blood, through their medical device.”In a fluid-based diagnostic device, a liquid such as blood or urine travels down a channel and interacts with reagents along the way. The chemistry involved could lead to a visible color change, as is the case in a pregnancy test, or to fluorescence that must be excited into emission, as can be the case in blood analysis. If the fluorescence is in the infrared, information about the fluid emission is captured, but that of the surroundings is lost. Such information is important because the response can depend on the concentration of the reactants and on how long the reaction is sustained.Two channels yield the complete picture needed to characterize a diagnostic device. Fluorescent fluid flow (red) highlights a chemistry change in blood or another fluid. This is overlaid on a visible image in green that pinpoints the reaction in the physical device — important information for characterization. Courtesy of Cyth Systems.In the case of the Cyth project, two components were needed for this determination: a visible image related to fluid mixing, and infrared fluorescence. To accomplish this, the company used photonics, motion control and image processing to develop an imaging subsystem in conjunction with Edmund Optics of Barrington, N.J.Gregory Hollows, director of machine vision solutions at Edmund Optics, noted that microscopy was needed to record the fluorescence. The group used a red laser as a source and detected fluorescence in the 780- to 800-nm band with a standard CCD camera and an infrared filter removed for imaging the fluorescence. Hollows characterized this part of the solution as straightforward and fairly easy to design and implement.Detecting flowTracking fluid flow was more difficult because there was little contrast between the liquid and the container. “We were looking at this material that has a plastic window on top of it, and then it has a clear liquid with a white backing,” Hollows explained. The trick, he said, was careful placement of the light source, a linear array of visible LEDs set at a low angle to the device under inspection. This resulted in an image where the fluid was clearly discernible. Detection was performed with a camera similar to the one used to capture the fluorescence. This second camera retained its infrared filter, and its optical path included another filter to remove stray light from the excitation laser. A glancing angle light makes visible a clear fluid droplet on top of a white background, thanks to high contrast. Courtesy of Cyth Systems.A beamsplitter sent the visible and infrared signals to the appropriate cameras, with the two precisely placed in the right position thanks to software developed by Cyth. Telecentric lenses helped to ensure that the images on the two channels were well aligned, even at the extreme edges of the image. This allowed the visible information about fluid location and mixing to be overlaid with the fluorescence data in real-time video about chemical activity.This microscopic picture of what was happening was limited by a field of view 10 or so times smaller than the fluid movement through the device, Inglese noted. The company’s engineers solved this problem by moving the stage automatically as needed. When the software detected that the fluid features of interest were approaching the edge of the field of view, the stage moved so that the region of interest was close to the center again. By doing so in 50 to 100 steps, the system allowed the flow to be tracked throughout the entire device with micron-scale precision. This big picture proved crucial to completing device development.Although intended for a specific project, Inglese said that the same techniques could be applied in a number of areas. An example might be an investigation of how cells behave in a fluid. In that case, the cells might be marked fluorescently or by other means. Such labeling allows investigators and product developers to see what the cells are doing but doesn’t always allow their location to be known.“This technology allows us to find the cells and track them in a flow,” Inglese said.Contact: Gregory Hollows, Edmund Optics; e-mail: email@example.com.