Fig. 3: Excitonic emission and tunnelling mechanisms with few-layer InSe. | Nature Nanotechnology

Fig. 3: Excitonic emission and tunnelling mechanisms with few-layer InSe.

From: Electrical detection of the flat-band dispersion in van der Waals field-effect structures

Fig. 3

a, Gate-dependent photoluminescence (PL) spectra for the 6L device. The change in the main excitonic peaks with respect to gate voltage is related to different charge configurations (Supplementary Note 4). When the Fermi level in the semiconductor enters the valence band (grey dashed line), the highest energy peak shifts linearly with respect to the gate voltage due to BGR in the presence of a dense hole Fermi sea. b, Peak position of the highest energy excitonic species (red) and tunnelling photocurrent (yellow) in the vicinity of the VBM. The direct tunnelling (DT)-to-FNT transition (\({V}_{\mathrm{TRAN}}^{5{\mathrm{L}}}\)) and the BGR start (\({V}_{\mathrm{BGR}}^{5{\mathrm{L}}}\) are highlighted by dashed yellow and red lines, respectively. c, Power-dependent tunnelling photocurrent for the 5L device (Supplementary Note 5) when the Fermi level lies within the bandgap (blue) or has entered the valence band (red). The superlinear behaviour of the former case indicates that second-order excitonic effects play a role, which is excluded within the valence band. d, Auger and exciton–exciton recombination mechanisms can radiatively transfer their energy to resident carriers and induce hot-hole tunnelling through a thick hBN barrier in the presence of a vertical electric field (EV<EF<ED). When the Fermi level lies within the valence band (EF<EV), excitonic species are subject to interactions with the Fermi sea induced by the high DOS and directly photo-excited holes are responsible for the photocurrent by FNT through the hBN barrier (right). e, Transition voltage (VTRAN), BGR crossover (VBGR) and p-type onset (VON) for the 3L, 5L and 6L devices. The 5L sample was not equipped with two FLG contacts for lateral transport measurements, and the intensity of photoluminescence emission in the 3L sample was too low to be analysed as a function of gate voltage. f, Layer-dependent effective tunnelling barrier extracted from FNT fittings (Fig. 2b) and valence-band offsets obtained from DFT calculations (Supplementary Fig. 1). The effective barrier values are presented as mean values (dots) and corresponding standard deviations (error bars).

Back to article page