Computational efficiency in photolithographic process simulation

Abstract

Photolithographic process simulation is described in which fast computation of resultant intensity for a large number of process variations and/or target depths (var,z t ) is achieved by computation of a set of partial intensity functions independent of (var,z t ) using a mask transmittance function, a plurality of illumination system modes, and a plurality of preselected basis spatial functions independent of (var,z t ). Subsequently, for each of many different (var,z t ) combinations, expansion coefficients are computed for which the preselected basis spatial functions, when weighted by those expansion coefficients, characterize a point response of a projection-processing system determined for that (var, z t ) combination. The resultant intensity for that (var,z t ) combination is then computed as a sum of the partial intensity functions weighted according to corresponding products of those expansion coefficients. Prediction of a mask transmittance function as a function of illumination incidence angle for a regional cluster of source emitters is also described.

1 - 23 . (canceled) 
     
     
         24 . A computer-implemented method for simulating a photolithographic processing system, comprising the computer-implemented steps of:
 computing a plurality of partial simulation results associated with a pattern transfer from a mask to a target by the photolithographic processing system, said target comprising one of a photoresist layer and a semiconductor wafer layer;   receiving a first value of a process variation associated with at least one of said mask, said target, and said photolithographic processing system;   computing a first complete simulation result by combining said partial simulation results according to said first value of said process variation;   receiving a second value of said process variation;   computing a second complete simulation result by combining said partial simulation results according to said second value of said process variation.   
     
     
         25 . The computer-implemented method of  claim 24 , wherein said target comprises said photoresist layer, and wherein each said complete simulation result is representative of an optical intensity in said photoresist layer. 
     
     
         26 . The computer-implemented method of  claim 24 , wherein said target comprises said semiconductor wafer layer, and wherein each said complete simulation result is representative of at least one of a final semiconductor two-dimensional pattern, a final semiconductor three-dimensional pattern, and a final semiconductor etch depth profile. 
     
     
         27 . The computer-implemented method of  claim 24 , wherein said computing said plurality of partial simulation results comprises convolving a mask transmittance function associated with said mask with each of a respective plurality of convolution kernels associated with the photolithographic processing system to generate said plurality of partial simulation results, each said convolution kernel at least partially characterizing said photolithographic processing system while being independent of said process variation. 
     
     
         28 . The computer-implemented method of  claim 27 , wherein said first and second complete simulation results correspond in outcome to computing first and second weighted combinations, respectively, of said plurality of partial simulation results according to first and second pluralities of weighting factors, respectively, computed according to said first and second values, respectively, of the process variation. 
     
     
         29 . The computer-implemented method of  claim 28 , wherein said first and second weighted combinations comprise first and second sums, respectively, of squared magnitudes of said partial simulation results as weighted by said first and second pluralities of weighting factors, respectively. 
     
     
         30 . The computer-implemented method of  claim 27 , the photolithographic processing system comprising an illumination system and a projection-processing system, the method further comprising computing the plurality of convolution kernels associated with the photolithographic processing system according to the computer-implemented steps of:
 receiving first information comprising parameters of the illumination system;   computing from said first information second information characterizing said illumination system by a plurality of modes impingent upon the mask being illuminated thereby;   receiving third information comprising a plurality of basis spatial functions at least partially characterizing said projection-processing system while being independent of said process variation; and   computing the plurality of convolution kernels from said second and third information, each convolution kernel corresponding to a product of one of said basis spatial functions and one of said modes.