Scatterings image

Researchers at the U.S. National Institute of Standards (NIST) have developed a system that uses the photobleaching rate of laser-excited fluorophore molecules as a marker to keep tabs on ultraslow flows in microfluidic channels. [Image: S. Kelley/NIST]

The burgeoning field of microfluidics offers visions of lab-on-chip analysis platforms for minute liquid samples, and devices that dispense tiny amounts of drugs at regular intervals. Many such uses require continuous measurement and control of these tiny, flowing liquid fractions. And, as new applications call for lower and lower flow rates through microfluidic channels—as low as tens of nanoliters per minute (nL/min) or less—measuring those flows within acceptable uncertainties becomes a real problem.

A team at the U.S. National Institute of Standards and Technology (NIST) has now proposed an optical solution, which uses the photobleaching rate of flurophore dyes as a sort of speedometer for microfluidic flow rates (Anal. Chem., doi: 10.1021/acs.analchem.9b02056). The researchers report that the system allows the real-time measurement, within 5% accuracy, of flows as low as 10 nL/min through human-hair-thin microfluidic channels.

Current shortfalls

The current standard for low-flow measurements, gravimetric methods, can routinely provide precise flow measurements to 1 microliter per minute (μL/min)—and, with suitable optimization, have achieved low-uncertainty measurement of flow at rates below 10 nL/min. But these results have required integration across five minutes of measurement time. That’s far too long a time window for systems designed, for example, to mete out dose-sensitive cancer drugs at rates of a few nanoliters per minute.

Further, these and other methods depend heavily on the geometry of the channel, temperature and other parameters. And some require the use of microscopes or other bulk components, which limit the prospect of use on microchip-sized devices.

Photobleaching’s positive side

In a theoretical study published in the spring (Phys. Rev. Appl., doi: 10.1103/PhysRevApplied.11.034025) NIST scientists Paul Patrone, Gregory Cooksey and Anthony Kearsley proposed an alternative approach. In their scheme, a known concentration of fluorescent dye molecules are included as markers in the solution flowing through the microfluidic channel. The dye molecules have a known photobleaching threshold—the number of times they can absorb and re-emit photons of laser light before they “burn out” and cease to fluoresce.

Thus, the team reasoned, for fast-flowing liquids exposed to laser excitation across a small length of the microfluidic channel, relatively few dye molecules would photobleach, and a strong fluorescence signal would result. For slower and slower flows, however, the fraction of the dye molecules exposed to enough laser light to burn them out would become proportionally larger, and the corresponding fluorescence signal, integrated across the laser-illuminated region, would become proportionally weaker. A NIST video illustrates the basic concept.

[Video: S. Kelley/NIST]

Experimental testbed

The early theoretical study showed mathematically that this system should have a number of advantages—in particular, that it depended only on fluorescence efficiency, flow rate and laser power, and not on the specific microchannel geometry or flow profile. In a new study, Cooksey, Patrone and Kearsley, joined by colleagues James Hands and Stephen Meek, have now put the system to an experimental test.

They began by using photolithography and soft lithography to create optofluidic devices with channel widths on the order of 25 to 100 μm. On either side of the channel, at a right angle to it, they then connected several optical fibers to handle the excitation and fluorescence measurement of the fluorphore-doped fluid running through the channel. One such fiber was tied to a diode laser to deliver the excitation light, and two others were set up to collect re-emitted fluorescent light from the channel and route it to photodetectors for measurement. Still another fiber, on the opposite side of the channel, would collect transmitted excitation light, to provide a check on the excitation laser’s power.

Next, the researchers flowed through the channel a buffer solution containing 20 or 50 μmol/L solution of fluorescein, a fluorophore with an excitation wavelength of around 488 nm and an emission wavelength of approximately 520 nm. They used a gravity-based system to vary the flow rates through the microfluidic channel, and used the measured fluorescence level to calculate dose–response curves for different flow rates and laser powers.

Real-time results

As expected, the photodetector-measured green fluorescence signal predictably weakened with diminishing flow rates. By calibrating the system’s results against measurements with established flow meters at the lower limit of their sensitivity, the team was able to establish that its system could measure flows as slow as 10 nL/min, with only 5% uncertainty and in real time.

Moreover, the team was able to establish that the optical system, unlike others, could provide sensitive measurements of a “zero flow” state—the point at which flow becomes indistinguishable from diffusion (roughly 0.2 nL/min). That ability, coupled with the scaling relationships from measurements at faster flow rates, opens up the possibility of validating and controlling flows at rates of as little as 2 nL/min, the researchers report.

According to a press release accompanying the work, the team has applied for a patent on the work.