I'm not sure if I'm posting in the right forum, please excuse if misposting. I have the following situation: pieces of conductive nanostructured material are attached to different substrates and heated resistively. We can only measure voltage and current through the piece of material (and respective contacts). The temperature of the nanostructured layer is derived using the temperature coefficient of resistivity, assuming linearity and that the piece of material behaves as a classical conductor. The same type of temperature derivation is then used to plot temperature against "on-off" cycles, where several voltage pulses are applied to the material. That is, 2 volts are applied for a few minutes, then a lower voltage for another few minutes. The current variation is used to calculate resistance and then temperature. The plot shows the square voltage pulses in time, overlapped with the corresponding, derived temperature pulses. This is done for each substrate. The cycling plots are meant to show that the material can heat up repeatedly and consistently, and to discuss differences in its heating speed given by the different substrates. My question: is this is a valid discussion, given that temperature is not measured directly?