For effective critical dimension metrology at geometries of 0.5 micrometers and smaller, it is essential to understand the true edge location, as represented in the SEM image profile of the feature. Total uncertainty associated with the linewidth measurement is often greater than the entire metrology error budget allocation. Capability to accurately predict (or model) the observed signal output for any type of feature is required to reduce measurement uncertainties to acceptable levels. The model, once validated, can then be used as the basis for measurement algorithms, error analysis, system calibration. In a manner analogous to the development of the optical photomask standards we investigate the capability to model, using a Monte Carlo based technique which incorporates measurement system geometries and fields, the output of an in-line, backscatter detection based, CD SEM. A comparison of line waveform profile results obtained from measurements taken of several simple resist-on-silicon test structures, with varying sidewall profiles, against simulations of the same structures is made. The line geometry for simulation input is derived using extensive analysis of in- line SEM and laboratory SEM systems for top down and crossectional measurement and an Atomic Force Microscope (AFM) for layer thickness and sidewall angle measurements. The results of these measurements is a trapezoidal line geometry for simulation input. Measurement profiles from simulation and experiment are compared and highlight the importance of; SEM operating condition (resolution/spot size) and linewidth subtleties (edge rounding at the top of the line, presence of a 'foot' or antifoot at the bottom of the line). Capability to use simulations to determine edge positions and the future work required to accomplish this are also discussed.