Spin-lattice relaxation in aluminum-doped semiconducting 4H and 6H polytypes of silicon carbide J. Stephen Hartman a,n , Bob Berno b , Paul Hazendonk c , Philip Hens d , Eric Ye e , Alex D. Bain b a Department of Chemistry, Brock University, St. Catharines, Ontario, Canada L2S 3A1 b Department of Chemistry and Chemical Biology, McMaster University, Hamilton, Ontario, Canada L8S 4M1 c Department of Chemistry, University of Lethbridge, Lethbridge, Alberta, Canada T1K 3M4 d Materials Department 6, University of Erlangen, Martensstr. 7, 91058 Erlangen, Germany e Department of Chemistry, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5 article info Article history: Received 20 January 2012 Received in revised form 29 May 2012 Available online 8 June 2012 Keywords: Silicon carbide Polytypes Semiconductors Doping NMR spin-lattice relaxation abstract NMR spin-lattice relaxation efficiency is similar at all carbon and silicon sites in aluminum-doped 4H- and 6H-polytype silicon carbide samples, indicating that the valence band edge (the top of the valence band), where the holes are located in p-doped materials, has similar charge densities at all atomic sites. This is in marked contrast to nitrogen-doped samples of the same polytypes where huge site-specific differences in relaxation efficiency indicate that the conduction band edge (the bottom of the conduction band), where the mobile electrons are located in n-doped materials, has very different charge densities at the different sites. An attempt was made to observe 27 Al NMR signals directly, but they are too broad, due to paramagnetic line broadening, to provide useful information about aluminum doping. & 2012 Elsevier Inc. All rights reserved. 1. Introduction Doping of silicon carbide with atoms such as nitrogen or aluminum converts an insulator to a wide-band-gap semiconduc- tor with many potential uses [1–4]. The unpaired electrons introduced by the dopant atoms cause the semiconductor beha- vior, which can be studied by various methods including the magnetic resonance methods NMR [5] and EPR [6,7]. As well, there have been a number of detailed studies of the electronic structure and transport in p and n doped silicon carbide [8–10]. Silicon carbide has been known as a semiconductor for a long time and has several advantages, but fabrication difficulties [1] have severely limited its use compared to more widely known semiconductors [11]. Silicon carbide exists as numerous poly- types (forms with different layer stacking sequences), each with its own unique silicon and carbon sites [12,13] and its own semiconductor properties [1,3]. We have found NMR spin-lattice relaxation to be particularly useful in our studies of silicon carbide and the characterization of individual sites [14,15]. NMR studies including spin-lattice relaxation have been car- ried out on many inorganic semiconductor systems [5], but only silicon carbide, with its unique ability to form numerous poly- types, provides multiple sites for the same nucleus in the same sample, allowing comparison of spin lattice relaxation efficiency at these different sites. Spin-lattice relaxation provides a very sensitive probe of conduction electron effects in rigid solids such as silicon carbide because most other mechanisms for nuclear spin-lattice relaxation are ineffective, making relaxation extre- mely inefficient. If unpaired electrons are present, their effects dominate the relaxation process [16–18]. This is illustrated by our 13 C and 29 Si spin lattice relaxation studies [14,15,19], where the extremely long intrinsic spin-lattice relaxation times of silicon carbide allow the detection of even very small relaxation effects arising from dopant and impurity atoms, which would be unde- tectable if the pure material had a viable relaxation mechanism. In semiconductor-grade doped silicon carbide it is the mobile conduction electrons or holes that are effective in causing relaxation [20]. In semiconductor-grade doped silicon carbide the efficiency of spin-lattice relaxation should be directly related to the concen- tration of the mobile electrons or holes. We have shown this behavior with the nitrogen-doped 6H [14] and 4H [15] polytypes. Not only does nitrogen doping make 13 C and 29 Si NMR spin-lattice relaxation much more efficient, but this effect is surprisingly site-specific with huge differences in relaxation efficiency at different sites in the same sample. There is precedent for massive site-specific effects in other properties such as luminescence spectra [21]. The existence of these site-specific differences in NMR relaxation, which indicate very different mobile electron densities at the nonequivalent sites, shows the Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/ssnmr Solid State Nuclear Magnetic Resonance 0926-2040/$ - see front matter & 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ssnmr.2012.05.003 n Corresponding author. Fax: þ1 905 984 4864. E-mail address: shartman@brocku.ca (J.S. Hartman). Solid State Nuclear Magnetic Resonance 45–46 (2012) 45–50