What determines the performance of lanthanide-based ratiometric nanothermometers?

Luminescence intensity ratio (LIR) nanothermometers are ideally suited for noninvasive temperature detection of microelectronic devices and living cells, and the painstaking pursuit of new nanothermometers with higher absolute temperature sensitivity (S-a) or relative temperature sensitivity (S-r) has dominated recent research. However, whether higher S-a and S-r values can intrinsically improve the performance of LIR nanothermometers and what factors essentially determine their accuracy have rarely been considered; these considerations are instructive for their design and application while reducing time and costs. Here, we clarify that the accuracy of lanthanide-based LIR nanothermometers is essentially determined by S-r and the relative error of the luminescence intensity (sigma I/I) but not S-a based on lanthanide-doped NaYF4, YPO4, YVO4, CaF2, YF3, Y2O3, BaTiO3, LaAlO3 and Y3Al5O12 temperature sensors, meaning that our previous pursuit of higher S-a does not contribute to the accuracy of lanthanide-based LIR nanothermometers. Further research reveals that sigma I/I is primarily influenced by energy level splitting, which can deteriorate the temperature uncertainty. For actual temperature detection of biological tissues, in addition to the above intrinsic factors, we shed light on the effects of probe self-heating, excitation power density, emission intensity and penetration depth on temperature readouts via a polyethyleneimine-modified NaYF4:Er3+/Yb3+@NaYF4-PEI aqueous solution, implying that we will continue to optimize nanothermometers and calibrate readouts according to the local environment. This work unifies the metrics of lanthanide-based LIR nanothermometers, corrects the previous misunderstanding of S-a to mitigate invalid work, and provides careful guidance for their development.