The impact of doping on the lattice constants of 4H–silicon carbide (4H–SiC) is an important material aspect influencing several steps of material and device production. Dopant incorporation in 4H–SiC causes misfit between the highly N-doped substrate and differently doped epilayers and hence, wafer bowing and the existence of a critical epilayer thickness. In this paper, the wafer bow is determined by geometrical measurements of the substrate prior to and after the epitaxial growth of single epilayers with different epilayer thicknesses and doping states, i.e. with nitrogen (N) or aluminum (Al) doping and different dopant concentrations. The misfit between substrate and epilayer is deduced from these bow measurements based on a model by Stoney. For highly Al-doped epilayers grown on highly N-doped substrates, the misfit is determined additionally by HRXRD measurements. The doping-dependent misfit, obtained from bow and HRXRD measurements, is compared to theoretical values calculated based on a model by Jacobson. For Al-doped epilayers, the experimental and theoretical values of misfit agree well. For N-doped epilayers, an effective covalent radius of N in 4H–SiC of 66 pm has to be introduced to match the theoretical misfit to the experimental one. The critical epilayer thickness is calculated based on the models by Matthews and Blakeslee and by People and Bean. The comparison of calculated values and experimental findings with respect to dislocation behavior at the substrate–epilayer interface and the dislocation densities of epilayers proves that the model by Matthews and Blakeslee underestimates the critical epilayer thickness for 4H–SiC homoepitaxy.
Read full abstract