Energy Harvesting Using Piezoelectric Nanowires-A Correspondence on "Energy Harvesting Using Nanowires?" by Alexe et al

We have demonstrated an innovative approach to convert mechanical energy into electrical energy by piezoelectric zinc oxide nanowire (NW) arrays. The mechanism of the nanogenerator (NG) relies on the coupling of piezoelectric and semiconducting dual properties of ZnO as well as the elegant rectifying function of the Schottky barrier formed between the metal tip and the NW. Alexe et al. have recently reported their assessment of the mechanism of the NG and they have raised the following three concerns. First, the piezoelectric charges in the ZnONWwas suggested to be completely cancelled out by existing free charge carriers in a ZnO NW within a very short period of time, thus, a piezoelectric potential would not be observed. Second, the detected output voltage from a single NW in the order of 10mV was insufficient to drive the Schottky barrier formed between the Pt electrode and the ZnO NW, thus no rectifying effect would be expected. Lastly, an output voltage was observed by them using their equipment for Si NWs, which are non-piezoelectric, thus, it was suggested that the received output from a NG might not be a result of a piezoelectric effect. This paper is set to fully analyze the questions raised and data presented by Alexe et al. and give a full comment. Based on a series of systematic experiments that we have carried out over the past three years, my conclusions are as follows. Alexe et al. overestimated the carrier density in the ZnO NW by up to two orders of magnitude. A UV-tuned conductivity experiment using a ZnO NW showed that the carrier density does affect the performance of the NG, but the conductivity of our as-synthesized NW is just right for one to observe the piezoelectricpotential-driven flow of external electrons. The piezoelectric potential remains in the NW for an extensive period of time, which allows direct detection of a piezoelectric-induced effect. The role played by the piezoelectric potential is to overcome the threshold voltage at the Pt–ZnO junction, while the observed output signal of 10mV is the difference in Fermi levels between the two electrodes connected to the Pt tip and the ZnO NW. Finally, the observation of potential generation by Si NWs by Alexe et al. is a result of system artifacts in their experiments. Their measurement system had a strong 25Hz interference background from the environment, an output noise of 10mV, and a huge equivalent capacitance of 320 pF. They used a bipolar amplifier with a bias current of 3.29 nA or even larger and an offset voltage of 32.5mV to amplify the signal, which produced an RC (resistor–capacitor) discharge signal at a magnitude of 800mV and shadowed any true signal ( 10mV) one tried to measure. As a result, their output signal was independent of the type, the size, and aspect ratio of the NWs. Consequently, they suggested without solid evidence that a piezoelectric effect was not responsible for the NG. More importantly, they mistakenly assumed that our measurement system was similar/identical to theirs, thus they used the artifacts ( 800mVpeaks) received from their system to explain the results ( –10mV) we have obtained using a highly sensitive measurement system (noise 0.5mV, capacitance 1.2 pf, amplifier introduced voltage offset of 1mV, and undetectable RC discharge effect (<1mV)). Finally, Alexe et al.’s model cannot explain 11 key experimental facts we have determined.