The continuous increase of computational resources and the availability of sophisticated optimizers and device simulators makes new doping characterization methods feasible which depend heavily on inverse modeling.
Most of the newly developed methods do not measure the doping concentration directly. After measuring electrical properties extensive inverse modeling is necessary to extract information on the doping. Assumptions on the investigated structure have to be made and data known from the manufacturing process is used. These methods are often limited to a small number of device structures.
In a method described in [16] an inverse modeling procedure is described in which the results of capacitance measurements, SIMS measurements and device simulation are combined to model the doping distribution by a tensor spline product.
In the SCM method [8] a tip is scanned across the surface of a cross-sectioned and polished device. The tip-sample capacitance is connected to a tuned circuit and a DC bias voltage is applied between tip and sample to set the depletion width. An additional low frequency AC voltage is applied to modulate the depletion width and the capacitance. The measurements are repeated using different bias conditions. The doping information is extracted from the measurement results using device simulation.