AVS2017 Session EM-WeM: Charge Transport in Disordered Materials

K. Prasai, P. Biswas and D. A. Drabold, Electrons and Phonons in Amorphous Semiconductors, Semicon. Sci. Tech. 31 073002 (2016).

Transparent, conductive electrodes have many applications in electronic and photonic devices such as touch screens, solar cells, LEDs, and photodetectors. Metal nanowire networks are promising candidates for these applications as a replacement for indium tin oxide, which has problems such as high cost, scarcity, and brittleness. The conduction in metal nanowire networks is governed by percolation theory.

In most computational work, nanowires in these networks have been modeled as straight “sticks”. However, in real experiments, individual nanowires are not perfectly straight, but exhibit some degree of curviness. In this work, we perform systematic Monte Carlo simulations to study the effect of nanowire curviness on the scaling of percolation resistivity in nanowire networks. We generate the curved nanowires using 3^rd-order Bézier curves. These curves are endowed with a curviness angle property that specifies how far away the two intermediate control points of the Bézier curve may lie, in the tangential sense, from a straight path connecting the two ends of the curve. The curviness angle is varied to obtain networks of differing values of curl ratio, which is defined as the ratio between the curved length of a nanowire and the straight distance between its two ends.

We find that, for random networks, the resistivity of the network increases with increasing nanowire curviness and the resistivity exhibits an inverse power law dependence on the curl ratio. We also find that the value of the extracted inverse power law critical exponent is not universal, but depends on other nanowire and device parameters. As a result, we also study the effect of nanowire density, nanowire length, device length, device width, and nanowire alignment angle on the scaling of network resistivity with nanowire curl ratio.

Curviness results in two competing effects on the percolation resistivity. First, curviness decreases the effective length of the nanowire between its two ends, which increases the resistivity. Second, it increases the effective width of the nanowire, which decreases the resistivity. For networks with aligned nanowires, we find that increasing the curviness decreases the resistivity, indicating that the second effect starts to dominate. By simulating networks with varying values of alignment angles, we study the crossover from the first to the second regime.

Infrared plasmonic nanocrystal thin films are promising materials for harvesting thermal radiation, but their use requires an ability to simultaneously control nanocrystal carrier density and interfacial charge transport. Here, we combine in situ infrared spectroscopy and nanoscale four-point probe measurements to show the intimate interplay between nanocrystal surface chemistry, the localized surface plasmon resonance (LSPR), and thin film resistivity. Indium tin oxide (ITO) nanocrystals synthesized via colloidal methods exhibit LSPRs tunable in the mid-infrared (from 0.2 to 0.6 eV) and serve as model infrared plasmonic materials with which to study these effects. Removal of surface hydroxyl groups, via reaction with metal alkyl species (e.g., trimethylaluminum), and formation of a monolayer of oxygen-metal surface bonds (e.g., Al-O) reduces the number of nanocrystal surface traps. The corresponding increase in carrier density is evidenced by a clear blue shift and increase in intensity of the LSPR. This monolayer surface treatment decreases nanocrystal film resistivity by two orders of magnitude (to 1 x 10^-2 ohm-cm). We further demonstrate that conformal coatings (e.g., of Al₂O₃ or ZnO) permit control of nanocrystal thin film resistivity without modifying LSPR energy, a capability that allows us to independently engineer light absorption and device properties. We also find that these plasmonic nanocrystal films exhibit a photoconductivity in response to mid-infrared light illumination.

Ion transport is closely tied to chemical transformations of crystalline ionic solids such as impurity doping and ion exchange. In order to introduce heteroatoms and other impurities into a crystal it is necessary to displace ions, create vacancies, and fill vacancies many times to achieve a significant degree of transformation. Thus ionic mobility plays a deterministic role both as to how fast a transformation occurs and the achievable extent of doping/exchange in a given system. For example exchange of Cd²⁺ for Cu⁺ in CdSe nanocrystals occurs rapidly and full conversion is ultimately reached. However replacing Cd²⁺ for Hg²⁺ does not reach completion and proceeds more slowly. A full explanation for how these two processes differ must address the mobility difference between Hg²⁺ and Cu⁺ through the Se^2- sublattice.

As a separate example: fast ion conduction relies on there being extensive ion transport in a solid. Such transport could be facilitated by cooperative effects, low (<k_bT) activation energy barriers against ion migration, or a combination of these and other factors. An understanding for how ionic mobilities differ and how they might be manipulated or enhanced will guide the design of materials such as nanoscale semiconductors for applications in optoelectronics, solid state battery electrolytes, and thermoelectrics. Specifically, control over superionic phase transitions can make new materials available for applications in solid state electrolytes by permitting new superionic conductors to be realized at working conditions as opposed to at elevated temperatures.

The approach is to employ density functional theory (DFT) electronic structure calculations for the elucidation of ion conduction trends in Li_xCu_2-xSe. The Li_xCu_2-xSe alloys vary in their fraction of Li⁺ or Cu⁺ cations both of which are mobile and thus apt for a study into tuning ion mobility. Total energies from DFT calculations for a large set of Li_xCu_2-xSe structures permit the identification of factors that promote or inhibit ion mobility. By varying both the cation locations and cell parameters the energetic penalties associated with pushing cations from their relaxed positions into interstitial sites can be obtained. Coupled with a exploration of the Li_xCu_2-xSe compositional range, the DFT energies reveal the best alloys for cation transport that have the lowest energetic penalties and smallest activation barriers. Complementary calculations in which the simulation cell volume is increased or decreased study the effects of tensile and compressive strain respectively; the calculated energies inform how strain and pressure can be used to control ion conduction.

We report on the fabrication and electrical characterization of bottom gate thin-film transistors (TFTs) based on sol-gel derived ZnO channel layer. The effect of annealing of ZnO active channel layers on electrical characteristics of the ZnO TFTs was systematically investigated. Photoluminescence (PL) spectra indicate that the crystal quality of the ZnO improves with increasing annealing temperature. Both the device turn-on voltage (V_on) and threshold voltage (V_T) shift to a positive voltage with increasing annealing temperature. As annealing temperature is increased, both the subthreshold slope (SS) and the interfacial defect density (D_it) decrease. The field effect mobility (μ_FET) increases with annealing temperature, peaking at 800 ℃ and decreases upon further temperature increase. Besides that, the temperature dependent field effect measurement (from 296 K to 330 K) was used to gain an insight on the annealing temperature effect on sol-gel derived ZnO TFTs electrical characteristics. The conduction processes in these TFTs are thermally activated and the drain current, especially in subthreshold regime, obeys Arrhenius equation. Thermal activation energy and a corresponding prefactor parameter were extracted from temperature dependent field effect measurement. The Meyer-Neldel (MN) rule widely observed in the intrinsic material property study is also obeyed by the relation between thermal activation energy and the prefactor parameter. By combing the MN rule and applying the self-consistent procedure, the localized sub-gap density of states ( DOSs) of different temperature annealed sol-gel derived ZnO TFTs were successfully extracted. The DOSs for subthreshold regime decrease from 10¹⁹ eV^-1cm^-3 level to 10¹⁷ eV^-1cm^-3 level with increasing annealing temperature from 600 ℃ to 800 ℃ and no substantial change was observed with further temperature increase to 900 ℃. The results show that DOSs decrease with annealing temperature increase in general, but annealing at a very high temperature may not be beneficial to reducing DOSs. It was found that the TFTs with ZnO layers annealed at 800 °C for one hour shows the best electrical performance with the highest field effect mobility and lowest localized DOSs.

Organic-inorganic hybrid nanostructures have been proposed to improve the performance of organic photovoltaic devices. In this work, we investigated creation and transport behaviors of photo-generated carriers in P3HT-based hybrid nanostructures, such as P3HT-coated Si-nanopillar (NP) arrays and P3HT layers with embedded upconversion nanorods (UCNs). Optical resonance strongly concentrated incident light in the Si NPs, significantly increasing the number of photo-carriers near the NPs. The UCNs converted near-infrared (NIR) light to visible light . We studied spatial distribution of the surface photovoltage (SPV) in the P3HT/Si-NP and P3HT/UCN samples using Kevin probe force microscopy. The magnitude and sign of SPV are determined by the number of the photo-carriers and band profiles near the sample surface [1-3]. Large SPV values appeared in the P3HT layers near the Si NPs under illumination of visible light and near the UCNs under illumination of NIR light in the P3HT/Si-NP and P3HT/UCN samples, respectively [3,4]. This suggested that the concentrated light in the NPs and photon upconversion in the UCNs increased the local density of photo-carriers. Our work shows that the scanning probe microscopy technique can visualize the creation and transport behaviors of photo-generated carriers in nanostructures.

1. E. Kim et al., Curr. Appl. Phys.16, 141 (2016).

2. M. Gwon et al., Sci. Rep.5, 16727 (2015).

3. E. Kim et al., Sci. Rep.6, 29472 (2016).

4. Y. Jang, E. Kim et al., J. Phys. Chem. Lett. 8, 364 (2017).

Leakage currents in dielectric thin films utilized in present day integrated circuitry are important reliability concerns. Among the most important dielectric materials are amorphous hydrogenated silicon nitrides (a-SiN:H). These relatively high dielectric constant materials are utilized in many applications such as a passivating layer, an etch stop layer, a diffusion barrier to water, and a gate dielectric. Electron paramagnetic resonance (EPR) studies of a-SiN:H films have identified the K center (a silicon dangling bond back-bonded to three nitrogen atoms) as the single dominating paramagnetic defect in stoichiometric films. Previous studies of spin dependent trap assisted tunneling (SDTAT) detected via electrically detected magnetic resonance (EDMR) provide us with energy levels for these K center defects. [1] However, the effects of varying N/Si stoichiometry on defect levels and defect chemistry have not been studied with EDMR. We have initiated such an EDMR study of SDTAT in a-SiN:H dielectric samples of several stoichiometries.

In our SDTAT/EDMR measurements, a slowly varying magnetic field and an oscillating rf or microwave frequency magnetic field are applied to the thin film samples. As in conventional EPR, energy is absorbed by paramagnetic sites when the resonance condition is met. In the simplest cases, this condition may be expressed by hv=guB, where h is Planck’s constant, g is an orientation dependent parameter often close to 2, u is the Bohr magneton, and B is the magnetic field. In EDMR, the EPR is detected through a change in current, in our case due to SDTAT.

The devices in our study include 3 nm a-SiN:H stoichiometric samples, and 25 nm a-SiN:H samples with three N/Si ratios of 1, 1.35, and 1.5. The overall device structures under observation consist of Ti/a-SiN:H/p-Si capacitors. A comparison of EDMR measurements taken at high field and frequency (X-band frequency ~ 9.75GHz, 3500G) and low field and frequency (frequency ~ 85-350MHz, 30-125G) provide us with information about defect structure. These comparisons allow us to extract information about the g matrix as well as hyperfine interactions. (The g and hyperfine details provide information about defect structure.) A comparison of EDMR measurements under various biasing conditions allow us to approximately determine the energy levels of the defects involved. This energy level and defect structure information should lead to a better understanding of transport in these technologically important materials.

[1] M.J. Mutch, P.M. Lenahan, and S.W. King, Appl. Phys. Lett. 109(6), 062403 (2016).