As the minimum feature size decreases, the gap between real experimental lithography process and simulated one increases. This gap should be reduced as small as possible by inserting the correct process parameters to simulation. Unfortunately, we do not have the exact simulation parameters in most cases and we need to get more accurate parameters. Among many methods to obtain the exact parameters, we used a new automatic cross-sectional critical shape error method to get the develop parameters by comparing the experimental scanning electron microscope image with the simulated image. This new bitmap masking technique is much faster than the conventional serial cross-sectional critical shape error method.

Dill's ABC parameters are key parameters for the simulation of photolithography patterning. The exposure parameters of each resist should be exactly known to simulate the desired pattern. In ordinary extracting methods of Dill's ABC parameters, the changed refractive index and the absorption coefficient of photoresist are needed during exposure process. Generally, these methods are not easy to be applied to in a normal fab because of a difficulty of insitu measuring. An empirical E0(dose-to-clear) swing curve is used to extract ABC exposure parameters previously by our group [1]. Dill's ABC parameters are not independent from each other and different values of them would cause the dose to clear swing curve variation. By using the known relationship ofABC parameters, the experimental swing curves are to be matched with the simulated ones in order to extract the parameters. But sometimes this method is not easy in matching the procedure and performing simulation. This procedure would take niuch time for matching between the experimental data and the simulation by the naked eyes. and also the simulations are performed over and over again for different conditions. In this paper, Dill's ABC parameters were extracted by applying the values, which are quantitatively determined by measuring the mean value, period, slope, and amplitude ofthe swing curve, to the neural network algorithm. As a result, Dill's ABC parameters were able to be rapidly and accurately extracted with some of the quantified values ofthe swing curve. This method ofextracting the exposure parameters can be used in a normal fab so that any engineer can easily obtain the exposure parameters and apply them to the simulation tools.

No abstract

The line-width of semiconductor devices is getting smaller, as the demand of highly density devices is increased. As the minimum feature size is decreased, the difference between real and simulation lithography process is increased and it is difficult to predict the difference. So, there is a need to make this difference as small as possible by inserting exact process parameters fitting for each process in several simulators. It is important that more accurate process parameters are extracted to predict the results of each process by simulation.In order to measure the difference, we first obtained SEM images of several chemically amplified resists, and then, the simulation results were produced by a number of develop parameter set. Subsequently, we used cross-sectional critical shape error (CCSE) method to compare the experimental SEM image with simulation result. By using this CCSE method, we extracted develop parameters of some chemically amplified resist (CAR). Furthermore, we needed to decrease the calculation time greatly to predict a set of more exact develop parameter. We improved the speed of CCSE data treatment by distributing the calculation through several computers and made it possible to manage automatically the new developed CCSE method. This extraction method of develop parameters by using CCSE is easier and gives much better results compares other existing methods.