Abstract
Temperature sensitive quantum dots (QDs) (CdTe and CdSe/ZnS) are investigated as internal temperature sensors for the growing field of 3D printed microfluidic devices. Two devices were created, one for using CdTe as the temperature sensor and another for using CdSe/ZnS. The QDs were mixed with poly (ethelyne glycol) diacrylate (PEGDA) resin and a thermal curing initiator, inserted into their devices, and cured in place. The fluorescence-to-temperature correlation was calibrated across the span of 30-90°C and then tested using a 3(rd) order fit of photoluminescence peak intensity (PLPI) to temperature and a feed-forward neural network (FFNN) combining multiple features of the fluorescence into a single temperature. This results in an improved temperature reconstruction accuracy of ±0.13°C for a FFNN of 27 inputs to one output, compared to ±0.29°C for PLPI as the single input.