Fig. 3: Electro-optical characterization.

a I–V curves of a RIET device (with AgNRs) and its control sample (without AgNRs). b Voltage dependence of LSP emission power. c Voltage dependence of the EQE. The error bars were obtained from multiple experimental measurements. d REET process at peak-1 in a. The REET current reaches its peak when the conduction band of ITO is aligned with the resonant state |6〉 of the MQW. The resonant state |5〉 is now below the Fermi level of the right electrode, so that the corresponding RIET process is forbidden. e Image of far-field photon emission at \({V}_{b}\) = 2.2 V. f RIET process at peak-3 in a. The REET process of the electron via the resonant state |6〉 is suppressed, while its RIET process by the resonant state |5〉 is active to excite the LSP. These two factors together lead to the high EQE obtained in c. Note that the RIET via the resonant state |5〉 is also possible; however, it occurs at the mid-infrared wavelength which is out of our current plasmonic design and experimental window, so the effect of this RIET process can be ignored.