Energy Band Alignment of a Monolayer MoS₂ with SiO₂ and Al₂O₃ Insulators from Internal Photoemission

Abstract

Internal photoemission of electrons (IPE) from large area one monolayer 2H-MoS 2 films synthesized on top of amorphous (aÀ) SiO 2 or Al 2 O 3 is used to determine the energy of the semiconductor valence band (VB) relative to the reference level of the insulator conduction band (CB). This allows us to compare the VB top energy in MoS 2 to that of the (100)Si substrate crystal at the interface with the same insulator. Despite the CB in a-Al 2 O 3 is found to be %1 eV below that in SiO 2 as measured relative to the Si VB edge, the authors observe nearly no shift of the spectral threshold in the case of IPE from the MoS 2 VB. This observation indicates violation of electroneutrality at the MoS 2 /a-Al 2 O 3 interface causing an increase in barrier by %1 eV. This conclusion is supported by the much weaker field dependence of the IPE threshold at the MoS 2 /a-Al 2 O 3 interface compared to the MoS 2 /a-SiO 2 one, suggesting the presence of negative charges and/or interface dipoles. Therefore, the commonly accepted electron affinity rule (EAR) appears to be not appropriate to describe the band alignment at 2D/insulator interfaces.

Band alignment at interfaces of two-dimensional materials: internal photoemission analysis

“…For example, the energy distribution of electrons emitted from a one ML 2D semiconductor such as MoS2 [83] and WS2 [84] into vacuum indicates that the DOS in the semiconductor VB increases linearly with energy within a ≈1 eV interval below the VB top similarly to the case of photoemission from a bulk Si crystal [85]. According to the Powell model [82], such DOS would correspond to the dependence, above the spectral threshold e, of the electron IPE quantum yield Y on photon energy h as given by the power function Y(h) ~ (h-e) 3 in agreement with experimental observations [57,86,87]. Apparently then, the analysis of IPE spectra for 2D photoemitters can be done in a similar way as carried out for the conventional 3D material interfaces.…”

“…Then, the presence of islands of the 2 nd and 3 rd MLs on top of the single ML MoS2 [98] will lead to lateral variations in the VB top energy position and, consequently, impair the hole transport properties of the TMD semiconductor film. Furthermore, violation of the Schottky-Mott rule is clearly revealed when comparing electron IPE from MoS2 into SiO2 with IPE into Al2O3 using the VB top of the substrate Si crystal as the common energy reference [87]: The thresholds for electron IPE from the Si VB into SiO2 (4.25 eV) and into ALD-grown amorphous Al2O3 (3.25 eV) differ, pointing to a ≈1 eV difference in the electron affinities of these oxides. However, the IPE threshold from a single-ML MoS2 appears to be barely influenced when changing the oxide from SiO2 to Al2O3, indicating that the intrinsic differences in the bulk electron structure of the distinct insulating oxide are "compensated" by a dipole (or charges) at the van der Waals bonded MoS2/oxide interface.…”

“…In turn, the field-dependent image force barrier model [82] in many cases describes the IPE threshold behavior reasonably well even for 2D photoemitters [86][87][88]. Therefore, it is still logical to use this model to account for the image-force barrier lowering and determine the true barrier height by extrapolating the field-dependent IPE threshold to zero field strength.…”

Energy Band Alignment of Few-Monolayer WS2 and WSe2 with SiO2 Using Internal Photoemission Spectroscopy

“…13 In particular, deposition of MoS 2 on top of Al 2 O 3 appears to induce a ≈1 eV downshift of the VB top energy as compared to SiO 2 when referencing it to the same energy level of the Si substrate VB top. 12 These results clearly show the importance of comparing the impact of the synthesis routes on the band alignment at the WS 2 /oxide interface, while this factor seems to be of less importance for WSe 2 as discussed below.…”

“…9,10 Among possible sources of this discrepancy one may mention not only the limited accuracy of theoretical methods in determining relative energies of bands at interfaces but, also several experimental aspects. The latter include the impact of the substrate which may introduce additional interface-specific electrostatic potentials, 11,12 the influence of material synthesis route as already known for MoS 2 13 and the 2D layer transfer processing. 14 All these factors potentially leading to a different band edge position at an interface compared to that measured under vacuum conditions.…”