The spatial resolution of dual-tracer fluorescence thermometry in volumetrically illuminated channels

This study estimates the spatial resolution and accuracy of dual-tracer fluorescence thermometry (DFT) for measuring water temperature fields in channels where the entire channel is illuminated as is typical in microfluidics. Temperature fields are measured in heated laminar Poiseuille flow through a 1 mm2 channel. The working fluid, an aqueous solution of the temperature-sensitive fluorophores fluorescein (Fl) and sulforhodamine B (SrB), is volumetrically illuminated over the entire channel cross-section at a wavelength of 514 nm, and the temperature of the solution is estimated from images of the longer-wavelength fluorescence from Fl and SrB. These temperature data are compared with numerical simulations of the same flow where the heat transferred to the water is estimated from independent measurements of wall surface temperature to determine the accuracy and the spatial resolution of the DFT results. The results suggest that temperature measurements in the volumetrically illuminated channel are significantly corrupted by the fluorescence emissions from beyond the focal plane. A model based on the point spread function for an aberration-free lens is employed to estimate the effect of the background “noise,” i.e., the signal from beyond the object plane, on the accuracy of these DFT measurements. The results show that this background is about 30 times the signal from the focal plane. Further experiments where the channel is illuminated by a light sheet over about 40 % of the channel cross-section give estimates of the water temperature field that are on average within about 0.3 °C of the numerical predictions at an in-plane spatial resolution of 50 μm. The model is used to estimate the signal-to-background ratio for this case, as well as for a variety of commercially available microscope objectives.