Factors limiting the doping efficiency in atomic layer deposited ZnO:Al thin films: a dopant distribution study by transmission electron microscopy and atom probe tomography

Abstract

Transparent conducting oxides (TCOs), such as indium tin oxide (ITO), are commonly used as transparent electrodes in a wide variety of devices, such as in displays and solar cells. ZnO has been reported to be a promising alternative TCO for ITO, because of its lower cost. As the conductivity of intrinsic ZnO films is too low for the applications in mind, doping the ZnO film is essential, the most common dopant being Al. Atomic layer deposition (ALD) is an emerging technique for the deposition of doped ZnO thin films, allowing for accurate thickness control and excellent conformality on high aspect ratio topologies. Due to the self‐limiting half‐reactions and cyclic nature of the ALD process, not only the aforementioned characteristics can be met, but also the amount and distribution of dopants can be controlled by selecting the precursors (i.e. the Zn or Al precursors) for each individual half‐cycle. However, thus far, the maximum conductivity that can be obtained in Al‐doped ZnO (ZnO:Al) thin films prepared by ALD appears to be limited by the low doping efficiency of Al. To better understand the origin of this limitation, the 3‐dimensional distribution of Al atoms in ZnO films has been examined using a combination of Transmission Electron Microscopy (TEM) and Atom Probe Tomography (APT). For this study, three ZnO:Al films with different Al:Zn ratios were grown sequentially in one film stack, and capped and separated by intrinsic ZnO films. A diagram of the stack is shown in Fig. 1a. This geometry allowed a single APT or TEM measurement to collect data on all three doped films, keeping the analytical conditions identical. BFTEM studies (Fig 1b) showed that for high Al concentrations the ZnO grains are interrupted, while they continue across the lower doped layers. Scanning TEM – High Angle Annular Dark Field (HAADF) imaging and 2‐D EDX mappings allows for revealing the aluminum distribution as a function of film depth, showing that the Al‐doped layers follow the surface topography of the ZnO grains during growth Fig. 1c,d. However, TEM is limited in providing 3‐D dopant distributions, on the one hand because of the limited sensitivity of EDX, on the other hand because of the projection of rough interfaces in a 2‐D image. The latter is illustrated in Fig. 2a: individual Al‐doped layers can clearly be discerned for larger interspacings, but are poorly recognizable in layer ‘AZO‐3’. One‐dimensional depth profiles extracted from cylindrical sub‐volumes of the 3D APT data (Fig. 2 b) are presented in Fig. 2c. These 1D profiles show that the peaks in Al concentration are no δ ‐functions, as might be expected from the binary nature of the ALD process. Instead, the peaks have a full width at half maximum (FWHM) of ~2 nm. The 3‐dimensional dopant distribution can be used to explain the dependencies of resistivity and doping efficiency on growth recipes used. When the local Al density is too high, the doping efficiency is limited by two proposed limiting factors: the solid solubility limit of Al atoms in a ZnO matrix and the disorder‐induced carrier localization.