Unveiling the Influence of Surface Fermi Level Pinning on the Piezoelectric Response of Semiconducting Nanowires

Abstract

environment ever since they were first demonstrated. [4] Experimentally, it has been proved that ZnO NW based piezoelectric devices can generate a potential of a few volts. [15] This can be retrieved theoretically as long as there are no free carriers. However, experimentally prepared ZnO NWs are spontaneously n-type doped by impurities during their synthesis [16,17] and in such a case, screening by free carriers is expected to reduce drastically the piezoelectric response. There remain many such contradictions between experimental results and the present theoretical understanding of ZnO NW-based devices when realistic doping levels are considered. Chief among them are (1) decent and length-dependent performance of ZnO NWs, [18] while anticipations based on analytical and computational study showed that the output of NWs under compression were reduced to a few millivolts and presenting length-independent performance due to the screening effect; [19,20] (2) enhanced piezoelectric coefficients, [21] which were measured for ZnO NWs with diameter beyond the size effects anticipated by ab initio method; [22] and (3) dissymmetric piezoelectric response of bent nanogenerators (NGs) under tensile and compressive strain. [23] Relatively high potentials (>0.2 V) can be repeatedly obtained from NGs integrating long and large ZnO NWs [24][25][26][27] with presumably nonintentional doping. Increasing the doping concentration will reduce the generated piezopotential. [28] The nature of matrix material, as well as processing conditions, has been shown to play an important role. [24] (Table S1, Supporting Information).In this paper, we solve these contradictions by accounting for surface Fermi level pinning (SFLP). This effect commonly exists at the surface of III-V and II-VI semiconductor compounds. [29][30][31] At the surface of n-type ZnO NWs, oxygen molecules get negatively charged by capturing the free electrons from NW core. This forms a low-conductivity depletion layer near the surface, where the screening of piezoelectric potential by free carriers effect is suppressed. [30,[32][33][34][35][36] This effect has been used to neglect free carriers in most simulation studies dealing with the piezoelectric response of semiconducting nanowirebased devices. A few papers have accounted for free carriers, although without account for SFLP. [20,37] Here instead, full coupling between piezoelectric effect and semiconducting physics, including free carriers and SFLP, was considered.ZnO nanowires (NWs) are excellent candidates for the integration of energy harvesters, mechanical sensors, piezotronic and piezophototronic devices. However, ZnO NWs are usually nonintentionally n-doped during their growth. Thus, it can be expected theoretically that their piezoelectric response be degraded and mostly geometry-independent as a result of strong screening effects by free carriers, while experimentally many NW-based piezoelectric transducers demonstrate quite reasonable performance. In this paper, this apparent contradiction is explained by...

Threshold Switching in Single Metal‐Oxide Nanobelt Devices Emulating an Artificial Nociceptor

“…With Ti electrodes, the theoretical difference between the Ti work function (4.33 eV) and the TiO 2 Fermi level (4.2 eV) is small enough to lead to an Ohmic contact. Nevertheless, the sweeping performance is similar to that observed for the Pt and Au electrodes, indicating that at least, partial Fermi level pinning may be occurring due to the surface‐dominating nature of the TiO 2 nanobelts. A similar transition from a capacitive behavior to a capacitive‐coupled memristive behavior is observed for the device with Au‐Au electrodes (Figure S5, Supporting Information).…”

Dimensional Roadmap for Maximizing the Piezoelectrical Response of ZnO Nanowire-Based Transducers: Impact of Growth Method

“…Basically, it is not straighforward to predict the output potential generated from NW-based piezoelectric transducers nor their general performance, because several key parameters, such as geometrical dimensions (i.e., radius and length), doping level (Nd), and surface trap density (Nit) [ 20 ] all play a significant role, and their relative effect on general performance may not be readily decoupled. One reason for that is that these key parameters strongly depend on the growth method used to form ZnO NWs.…”

“…It has been shown recently by numerical simulation that Fermi level pinning, resulting from the presence of surface traps at the interface between ZnO and the matrix material could explain the experimentally observed length dependence [ 20 ]. The presence of surface and interface traps has been widely acknowledged in III–V and II–VI semiconductors [ 59 , 60 , 61 , 62 ].…”

“…The presence of surface and interface traps has been widely acknowledged in III–V and II–VI semiconductors [ 59 , 60 , 61 , 62 ]. Their impact on device operation has been explained as follows [ 20 ]. In the absence of surface traps, or for low trap densities, there is no surface charge, and thus no band bending at the surface of ZnO.…”

Fundamental Definitions for Axially-Strained Piezo-Semiconductive Nanostructures

“…The Dirichlet condition (V = 0) is applied at the center of the nanowire. In order to consider the effect of charge carriers, the equations for semiconductors have been added to the simulations and, similar to [ 22 , 23 , 34 ], for simplicity, the effects of external charges [ 37 , 38 , 39 ] have not been considered. The nanowire is assumed to be uniformly n-type doped, with different doping conditions (intrinsic, 10 16 cm −3 , 10 17 cm −3 ).…”