Many biological analytical processes would benefit from miniaturization, but some —such as DNA amplification through polymerase chain reaction (PCR) — depend on homogeneous temperature gradients to optimize the chemical or molecular reactions. Putting PCR on a microfluidic chip, for example, should improve its efficiency, response and cost; however, achieving consistent temperatures through a liquid sample in a chip requires a finely tuned system for determining what the temperatures are at any given point in space and time. Existing temperature-monitoring technology — including thermocouples and various spectroscopic techniques — do not have sufficient sensitivity for the task. The time-integrated fluorescence intensity of fluorophores within a liquid can be measured, but this technique suffers from significant problems relating to calibration and artifacts resulting from excitation and detection efficiencies, said Andrew J. deMello, chairman of chemical nanosciences at Imperial College London. A team led by deMello and Paul M.W. French, chairman of the college’s Photonics Group, has used fluorescence lifetime imaging to create three-dimensional thermal maps of fluorophore-bearing fluids moving through microchannels on a chip. According to deMello, using fluorescence lifetime measurements means that the technique is independent of the concentration of the fluorescent dye within the liquid sample and of the efficiency of the detection system. Data in the flow The investigators report in the March 4 ASAP edition of Analytical Chemistry that they made 1-mm-thick glass microfluidic chips etched with channels 139 μm wide and ~38 μm deep. The chips were injected with solutions of rhodamine B in methanol. They chose rhodamine B because its fluorescence would not appreciably change as a result of the pH or the viscosity of the solution. They used two-photon excitation to create optical sections of the liquid, which flowed through the channels at rates between 0.1 and 10 μl/min. Varying the rate enabled them to alter how the liquid was warmed by the heating elements attached to the chip substrate. Using 840-nm, 100-fs pulses from a Ti:sapphire laser made by Spectra-Physics of Mountain View, Calif., that was coupled with a microscope from LaVision BioTec GmbH of Bielefeld, Germany, the scientists scanned the flowing liquid. Short pulses of infrared radiation helped control the amount of heat added to the fluid as it was being scanned. Two-photon excitation is less efficient than single-photon, deMello said, but because their system scans up to 64 beams in parallel, they achieved wide-field detection and good optical sectioning of the fluid flow. Furthermore, no laser-induced excitation or heating of the fluid occurred outside of the femtoliter-size excitation volume in the microchannel. They used the photocathode of a time-gated intensifier from Kentech Instruments Ltd. of Didcot, UK, to detect the rhodamine B emission. The fluorescence images were relayed to a CCD camera from LaVision GmbH of Göttingen, Germany, providing them with a series of 20 fluorescence intensity images after each excitation pulse. Acquisition times typically were 10 to 30 s per image, which they noted was compatible with the steady-state flow and heating conditions of their setup. The average of each image series provided a fluorescence decay profile for each pixel in the microscope’s field of view. The scientists converted the decay data into temperature readings by comparing it with calibration data from the literature and from measurements of bulk solutions of rhodamine maintained at various temperatures. Within the three-dimensional map of fluorescence decay that was created, they achieved a spatial resolution of 1 to 5 μm between pixels in each image section about 100 μm long and a precision of ±1 °C. Testing temperature zones They found that the fluorescence lifetime decreased by a factor of about two when the temperature of the fluid was increased from 66 to 93 °C — temperature zones commonly used during PCR. Multiple temperatures are required to perform the basic steps of amplifying DNA with PCR techniques, and miniaturizing a PCR system would require careful control of the temperature in each zone to promote the efficiency of the amplification steps. Researchers used fluorescence lifetime imaging to measure the temperature fluctuations inside a liquid flowing through a microchannel. Two-photon excitation of rhodamine B in solution with methanol enabled a 130 x 40 x 100-μm optical section showing the decay rates of the rhodamine after excitation (top). The data can be mathematically converted to temperature and smoothed (bottom). Courtesy of Andrew J. deMello. In PCR, deMello said, if the temperature at a specific location varies more than 1 °C, efficiencies are drastically reduced. “Our imaging approach will allow us to detect and correct for these variations.” To test whether the technique would work in PCR and other applications that require multiple temperature zones, the investigators acquired time-integrated fluorescence intensity images and fluorescence lifetime images at points along an S-shaped microchannel to which a temperature gradient was applied. They found that the intensity data was insufficient for recording temperature, but that the lifetime data was very much suitable. They also found that optimal fluid temperatures were achieved only within 5 mm of each heating element attached to the chip. The researchers believe that this knowledge will be important for designers of microchannels and other aspects of chips for PCR use. They are looking to improve the sensitivity of the system by using a fluorophore that exhibits a larger decrease in fluorescence lifetime as a function of temperature than the rhodamine B does. They also hope to test a chip substrate less than 1 mm thick that would permit them to work with lower working distances and to use objectives with a higher numerical aperture, which they anticipate would increase sensitivity. “The ability to image fluid mixing and flow should allow us to separate out these effects from observed chemical kinetics — for example, in continuous-flow experiments,” deMello said.