Metrologycharacterization and simulation of Line Width Roughness LWR

Metrology-characterization and simulation of Line Width Roughness (LWR) Evangelos Gogolides, Vassilios Constantoudis, George Patsis Institute of Microelectronics NCSR "Demokritos" Attiki - Greece (in affiliation with lithographic materials team of P. Argitis) More Moore-Excite Joint meeting, Athens May 12, 2005

Outline A. LWR metrology-characterization methodology 1. Off line detection of line edges and line widths from top down CD-SEM images 2. Characterization of LWR : The importance of sigma(L) curve 3. LWR and CD variation - LWR descriptors B. Simulation of lithography material and process effects on LWR

A. LWR metrology-characterization methodology Normalized pixel intensity 1. Off line software for the detection of line edges from top down CD-SEM images Line length or gate width SEM image after noise smoothing Determine the pixel size in nm and the noise smoothing parameters line width or gate length middle inner outer threshold Intensity profile for a pixel row Edge detection algorithm at each pixel row using a threshold value Obtain the line edges of the image (outer, inner or middle) A line edge LER : yi, R(L) of all the edges LWR : δyi of all the lines For uncorrelated and parallel edges : Patsis GP, Constantoudis V, Tserepi A, et al. Quantification of line-edge roughness of photoresists. I. A comparison between off-line and on-line analysis of top-down scanning electron microscopy images J. VAC SCI TECHNOL B 21 (3): 1008 -1018 MAY-JUN 2003

A. LWR metrology-characterization methodology 2. Characterization of LWR: The importance of sigma. LWR(L) curve • Why sigma. LWR(L) curve : sigma not a value BUT a function, sigma(L) curve 500 L 400 y (pixels) 300 L 200 100 L 0 0 50 100 150 200 250 300 350 400 450 500 x (pixels) Parameters for sigma(L) curve determination: • 1) sigma(inf): the sigma for the infinite line length • 2) LS: sigma-correlation length sigma(LS)=0. 9*sigma(inf). For L>Ls the line edges look flat (no correlations) and sigma(L) sigma(inf) • 3) α : roughness exponent giving the relative contribution of high frequency roughness to LWR. It is related to fractal dimension d=2 -α Constantoudis V, Patsis GP, Tserepi A, et al. Quantification of line-edge roughness of photoresists. II. Scaling and fractal analysis and the best roughness descriptors. J. VAC SCI TECHNOL B 21 (3): 1019 -1026 (2003)

A. LWR metrology-characterization methodology The physical meaning of α, Ls • Influence of α, Ls on: a) sigma. LWR(L) curve b) edge morphology 1. Different α, the same sigma, Ls α=0. 2 α=0. 8 3. Different Ls , the same sigma, α LS=300 LS =600

A. LWR metrology-characterization methodology 3. LWR and CD variation : estimating sigma. LWR(inf) using finite line lengths The key relationship : sigma. LWR 2(inf)= sigma. LWR 2(L)+ CDvariation 2(L)+ rmssigma 2(L) variation of sigma. LWR values J sigma. LWR(inf): a line length independent parameter that can be estimated using any line length A new parameter for LWR definition ? Also, ØA new meaning for α, Ls : control the partition of sigma. LWR(inf) in LWR and CD variation as line length L decreases (L<L s) See paper 5752 -141 - V. Constantoudis et al. in Metrology Session of SPIE 2005

B. Simulation of material and process effects on LWR Schematic flowchart of simulator Exposure Mask Developer Top-Down View LER Resist Deprotected Molecular structure Deprotected Side LER (Accurate) Front View LER Top-Down View LER in 2 D

SIMULATION OF ACID DIFFUSION AND MOLECULAR WEIGHT EFFECTS ON LER

SIMULATION OF ACID DIFFUSION AND MOLECULAR WEIGHT EFFECTS ON LER Top-Down View LER Side LER (Accurate) Front View LER Top-Down View LER in 2 D