论文部分内容阅读
The work reported was undertaken to establish an independent and reliable means for checking the impurity doping level in silicon epitaxial layer production. Both evaporation of aluminium and chemical plating of gold were used to make Schottky junctions to silicon epitaxial layers, however, the latter method was found to be simpler and more satisfactory. With suitable experience, circular junctions of diameters-1mm can normally be obtained with leakage currents of the order 1-10μA (at 10V reverse bias) and diameter variation M 1%. The differential junction capacitance was measured as a function of bias voltage on a high frequency bridge at a fixed frequency of 5MHz and a signal level of 10mV and the impurity distribution was deduced from the capacitance measurements by the well-known formulae given originally by Schottky. The maximum relative error and standard error of the distribution thus obtained were estimated to be 12% and 4% respectively. The method was compared with standard 4-point probe measurements on uniformly doped single crystal specimens, giving results in mutual agreement. The method was further tested on several epitaxial specimens obtained by certain non-uniform doping sequences; the impurity distributions determined showed clear correspondence with the doping sequences adopted.
The work reported was undertaken to establish an independent and reliable means for checking the impurity doping level in silicon epitaxial layer production. However, the latter method was checked for the impacted doping level in silicon epitaxial layer production. found to be simpler and more satisfactory. With appropriate experience, circular junctions of diameters-1mm can normally be obtained with leakage currents of the order 1-10μA (at 10V reverse bias) and diameter variation M1%. The differential junction capacitance was measured as a function of bias voltage on a high frequency bridge at a fixed frequency of 5MHz and a signal level of 10mV and the impurity distribution was deduced from the capacitance measurements by the well-known formulae originally originally by Schottky. The maximum relative error and standard error of the distribution thus obtained were obtained were estimated to be 12% and 4% respectively. The method was compared with standa rd 4-point probe measurements on uniformly-doped single crystal specimens, giving results in mutual agreement. The method was further tested on several epitaxial specimens obtained by certain non-uniform doping sequences; the impurity distribution determined determined clear correspondence with the doping sequences adopted.