Session X7: Dopants and Defects in Semiconductors: Compound Semiconductors

Electronic structure and scattering property of 4H-SiC(0001)/SiO$_2$ interface

SiC is attracted much attention as a promising material for the high-power electronics devices. We carried out a first-principles calculations to reveal the relationship between the electronic structure and the interface defects appearing in the thermal oxidation. We found interlayer the states along the SiC conduction band edge (CBE), whose location changes depending on which of two possible lattice sites, $h$ or $k$, is at the interface. Excess O atoms at the interface lead to defect structures which alter the electronic structure. Changes to the valence band edge are the same whether $h$ or $k$ sites are at the interface. On the other hand, defects remove the interlayer state of the CBE between the first and second SiC bilayers if an $h$ site is at the interface, but have no effect when there is a $k$ site. The scattering property of the defects was also examined by electron-transport calculations. Carriers at the CBE of the $h$-type interface are easily scattered by the defects because of the absence of the interlayer states while those at the $k$-type interface is not. Since recent SiC-MOSFETs mainly use the conduction band as a channel, the behavior of these interlayer states at the CBE might play an important role in the performance of these devices.   

Measurements of depth dependent modification of optical constants arising from H$^{+}$ implantation in n-type 4H-SiC using coherent acoustic phonons

Silicon carbide (SiC) is an ideal material for new electronics, such as high power/high temperature devices, and a candidate for advanced optical applications such as room temperature spintronics and quantum computing. Both types of applications may require the control of defects created by ion bombardment. In this work, we examine depth dependent modification of optical constants of 4H-SiC due to hydrogen implantation at 180keV and low doses ranging from 10$^{14}$ to 10$^{16}$ cm$^{-2\, }$probed by coherent acoustic phonon (CAP) spectroscopy. For our studies, we used Si-face 10$\mu $m \quad epilayers of n-type 4H-SiC grown by CVD on 4H-SiC substrate. A comprehensive analysis of the reference and implanted spectra shows a strong dependence of 4H-SiC complex refractive index shape versus depth on the H$^{+}$ fluence. We extract the complex refractive index as a function of depth and ion beam dose. Our results demonstrate that the implantation-modified refractive index is distributed over a greater depth range than Monte Carlo calculation predictions of the implantation induced structural damage. These studies provide insight into the application of hydrogen ion implantation to the fabrication of SiC-based photonic and optoelectronic devices.   

Revealing Hidden Interfacial States in NO Passivated 4H-SiC/SiO$_{2}$ Structures using TEM-EELS and XPS

The interface between \textit{4H}-SiC and SiO$_{2}$ in metal oxide semiconductor (MOS) devices contains a high density of electrically active defects, which adversely affect SiC microelectronic devices. Various treatments and altering the substrate's crystallographic orientation can improve electronic performance. We have previously shown an inverse relationship between nitric oxide (NO) anneal time and the width of the transition layer at this interface ($w_{TL}).^{2} $More recent work analyzing $w_{TL\, }$has revealed much narrower interfaces that do not appear to narrow when subjected to an NO post-oxidation anneal, contradicting expectations. To further explore these interfaces, high resolution transmission electron microscopy and spatially resolved electron energy-loss spectroscopy (EELS) have been used. In addition, X-ray photoemission spectroscopy measurements were taken at the interface. Advanced EELS analysis via machine learning techniques has revealed interfacial bonding states for different post-oxidation annealing processes. The nature of these interfacial states is compared for devices made on substrates with different orientations and for NO post-oxidation annealing. $^{2}$J. Taillon \textit{et al.,} \textit{J. Appl. Phys.} \textbf{113,} 044517 (2013).   

Generalization of the van der Pauw Method: Analyzing Longitudinal Magnetoresistance Asymmetry to Quantify Doping Gradients

A longitudinal magnetoresistance asymmetry (LMA) between a positive and negative magnetic field is known to occur in both the extreme quantum limit and the classical Drude limit in samples with a nonuniform doping density. By analyzing the current stream function in van der Pauw measurement geometry, it is shown that the electron density gradient can be quantitatively deduced from this LMA in the Drude regime [1]. Results agree with gradients interpolated from local densities calibrated across an entire wafer, establishing a generalization of the van der Pauw method to quantify density gradients. Results will be shown of various semoconductor systems where this method is applied, from bulk doped semiconductors, to exfoliated 2D materials. \newline [1] W. Zhou, H.M. Yoo, S. Prabhu-Gaunkar, L. Tiemann, C. Reichl, W. Wegscheider, and M. Grayson, Phys. Rev. Lett. {\bf 115}, 186804 (2015).   

First principles modeling of grain boundaries in CdTe

The role of extended defects is of significant interest for semiconductors, especially photovoltaics since energy conversion efficiencies are often affected by such defects. In particular, grain boundaries in CdTe photovoltaics are enigmatic since the achievable efficiencies of CdTe photovoltaics are higher in polycrystalline devices as compared to single crystalline devices. Yet, despite recent advances, the efficiency of poly-CdTe devices are still substantially below the theoretical maximum. We carry out an atomistic-level study using Scanning Transmission Electron Microscopy (STEM), together with first principles density functional theory (DFT) modeling, in order to understand the properties of specific bicrystals, i.e. artificial grain boundaries, constructed using wafer bonding. We discuss examples of bicrystals, including some involving large scale DFT calculations, and trends in defect and electronic properties.   

Origin of High Electronic Quality in Solar Cell Absorber CH$_{\mathrm{3}}$NH$_{\mathrm{3}}$PbI$_{\mathrm{3}}$

Thin-film solar cells based on CH$_{\mathrm{3}}$NH$_{\mathrm{3}}$PbI$_{\mathrm{3}}$ halide perovskites have recently shown remarkable performance. First-principle calculations and molecular dynamic simulations show that the structure of pristine CH$_{\mathrm{3}}$NH$_{\mathrm{3}}$PbI$_{\mathrm{3}}$ is much more disordered than the inorganic archetypal thin-film semiconductor CdTe. However, the structural disorders from thermal fluctuation, point defects and grain boundaries introduce rare deep defect states within the bandgaps; therefore, the material has high electronic quality. We have further shown that this unusually high electronic quality is attributed to the unique electronic structures of halide perovskite: the strong coupling between cation lone-pair Pb $s$ orbitals and anion $p$ orbitals and the large atomic size of constitute cation atoms. We further found that although CH$_{\mathrm{3}}$NH$_{\mathrm{3}}$PbI$_{\mathrm{3}}$ GBs do not introduce a deep gap state, the defect level close to the VBM can still act as a shallow hole trap state. Cl and O can spontaneously segregate into GBs and passivate those defect levels and deactivate the trap state.   

\textbf{Mn Doping Effects on the Electronic Band Structure of PbS Quantum Dot Thin Films: A Scanning Tunneling Microscopy Analysis}

A thorough understanding of the phenomena associated with doping of transition metals in semiconductors is important for the development of semiconducting electronic technologies such as semiconducting quantum dot sensitized solar cells (QDSSC). Manganese doping is of particular interest in a PbS QD as it is potentially capable of increasing overall QDSSC performance [1]. Here we present scanning tunneling microscopy and spectroscopy studies about the effects of Manganese doping on the energy band structures of PbS semiconducting QD thin films, grown using pulsed laser deposition. As a result of Manganese doping in the PbS QD thin films, a widening of the electronic band gap was observed, which is responsible for the observed increase in resistivity. Furthermore, a loss of long range periodicity observed by XRD, upon incorporation of Manganese, indicates that the Manganese dopants also induce a large amount of grain boundaries. [1] Qilin Dai \textit{et al.},APL 104,183901(2014).   

Electronic structure and defect properties of selenophosphate Pb$_{\mathrm{2}}$P$_{\mathrm{2}}$Se$_{\mathrm{6}}$ for $\gamma $-ray detection$^{\mathrm{1}}$

Heavy metal chalco-phosphate Pb$_{\mathrm{2}}$P$_{\mathrm{2}}$Se$_{\mathrm{6}}$ has shown a significant promise as an X-ray and $\gamma $-ray detector material. To assess the fundamental physical properties important for its performance as detector, theoretical calculations were performed for the electronic structure, band gaps, electron and hole effective masses, and static dielectric constants. The calculations were based on first-principles density functional theory (DFT) and employ the highly precise full potential linearized augmented plane wave method and the projector augmented wave method and include nonlocal exchange-correlation functionals to overcome the band gap underestimation in DFT calculations. The calculations show that Pb$_{\mathrm{2}}$P$_{\mathrm{2}}$Se$_{\mathrm{6}}$ is an indirect band gap material with the calculated band gap of 2.0 eV, has small effective masses, which could result in a good carrier mobility-lifetime product $\mu \tau $, and a very high static dielectric constant, which could lead to high mobility of carriers by screening of charged scattering centers. We further investigated a large set of native defects in Pb$_{\mathrm{2}}$P$_{\mathrm{2}}$Se$_{\mathrm{6}}$ to determine the optimal growth conditions for application as $\gamma $-ray detectors. The results suggest that the prevalent intrinsic defects are selenium vacancies, followed by lead vacancies, then phosphorus vacancies and antisite defects. The effect of various chemical environments on defect properties was examined and the optimal conditions for material synthesis were suggested. $^{\mathrm{1}}$Supported by DHS (Grant No. 2014-DN-077-ARI086-01).   

Challenges in p-type Doping of CdTe

We have made progress in defect identification of arsenic and phosphorous doped CdTe to understand the self-compensation mechanism which will help improve minority bulk carrier lifetime and net acceptor density. Combining previous measurements of un-doped CdTe, we performed a systematic comparison of defects between different types of crystals and confirmed the defects impacting the doping efficiency. CdTe bulk crystals have been grown via vertical Bridgman based melt growth technique with varying arsenic and phosphorous dopant schemes to attain p-type material. Furnace temperature profiles were varied to influence dopant solubility. Large carrier densities have been reproducibly obtained from these boules indicating successful incorporation of dopants into the lattice. However, these values are orders of magnitude lower than theoretical solubility values. Infrared Microscopy has revealed a plethora of geometrically abnormal second phase defects and X-ray Fluorescence has been used to identify the elemental composition of these defects. We believe that dopants become incorporated into these second phase defects as Cd compounds which act to inhibit dopant solubility in the lattice.   

Diversity of surface conduction in pyrite FeS$_{\mathrm{2}}$ single crystals

Pyrite FeS$_{\mathrm{2}}$ has long been recognized as an attractive material for solar cells because of its high absorptivity, potential low cost, high abundance, and low toxicity. Despite having appropriate band gap (0.95 eV) and minority carrier diffusion length (100-1000 nm), low open circuit voltages ($V_{oc}\le $ 0.1 V) have plagued FeS$_{\mathrm{2}}$-based cells. Surface conduction has been proposed as a contributing factor for the low $V_{oc}$, particularly a $p$-type surface inversion layer on $n$-type crystals [1]. Here we report a detailed electronic transport study of a large number of well-characterized CVT-grown $n$-FeS$_{\mathrm{2}}$ single crystals. Abundant evidence of surface conduction is found from the $T$ dependence of resistivity, resistance anisotropy, low $T$ behavior at the 2D quantum resistance, thickness dependence, and the influence of contact metal work function. However, striking diversity in this surface conduction is found, even in nominally identical crystals at similar doping. The results cannot be understood by surface inversion alone, pointing to as yet uncontrolled surface factors. [1] Limpinsel \textit{et al}. Energy Environ. Sci. (2014). Work supported by NSF.   

Grain boundaries and surfaces in polycrystalline photovoltaics

Despite the fact that polycrystalline photovoltaics materials such as CdTe and CIGS are an established commercial technology, the precise role of grain boundaries in their performance remains poorly understood. The high defect density at grain boundaries is generally detrimental to carrier lifetime, however the electric fields surrounding charged grain boundaries may separate electrons and holes, effectively passivating the grain boundary. One difficulty in ascertaining the properties of grain boundaries is that high spatial resolution experimental techniques needed to probe individual grain boundaries are generally surface sensitive. For this reason, extracting quantitative grain boundary and other material properties from this data requires a quantitatively accurate model of the exposed surface. Motivated by these considerations, we present a theoretical analysis of the response of a polycrystalline semiconductor to a localized excitation near a grain boundary, and near the surface. We use our analytical results to interpret electron beam induced current (EBIC) data on polycrystalline CdTe solar cells.   

Individual iso-electronic N and Bi centers in GaAs studied by Scanning Tunneling Microscopy.

Nitrogen and bismuth iso-electronic doping centers in GaAs have received considerable interest in the last few years due to their peculiar behaviour in dilute nitrides and bismides. In these materials effects such as a strong band bowing and the formation of resonant states in respectively the conduction and valence band have been reported. In this contribution we will report our exploration of individual nitrogen and bismuth atoms in the outermost layers of a freshly cleaved (110) GaAs surface by STM. Depending on the tunnel conditions we are able to either visualise the lattice distortion or image the charge distribution of the resonant state. We clearly observe that nitrogen pulls its neighbouring atoms inwards whereas bismuth pushes its neighbouring atoms away. A straightforward geometrical model based on the covalent radii of the dopants and substrate atoms is used to interpret the observed crystal deformation seen in our STM images of nitrogen and bismuth under the appropriate tunnel conditions. At small positive voltages we could observe the charge distribution of the resonant state induced by iso-electronic nitrogen atoms in GaAs. Tight Binding Modelling (TBM) was used to explain the observed strongly anisotropic charge distribution.