Optical scattering measurements and simulation data for one-dimensional (1-D) patterned periodic sub-wavelength features

Description

This data set consists of both measured and simulated optical intensities scattered off periodic line arrays, with simulations based upon an average geometric model for these lines. These data were generated in order to determine the average feature sizes based on optical scattering, which is an inverse problem for which solutions to the forward problem are calculated using electromagnetic simulations after a parameterization of the feature geometry. Here, the array of features measured and modeled is periodic in one-dimension (i.e., a line grating) with a nominal line width of 100 nm placed at 300 nm intervals, or pitch = 300 nm; the short-hand label for the features is "L100P300." The entirety of the modeled data is included, over two thousand simulations that are indexed using a top, middle, and bottom linewidth as floating parameters. Two subsets of these data, featuring differing sampling strategies, are also provided. This data set also contains angle-resolved optical measurements with uncertainties for nine arrays which differ in their dimensions due to lithographic variations using a focus/exposure matrix, as identified in a previous publication (https://doi.org/10.1117/12.777131). We have previously reported line widths determined from these measurements based upon non-linear regression to compare theory to experiment. Machine learning approaches are to be fostered for solving such inverse problems. Data are formatted for direct use in "Model-Based Optical Metrology in R: MoR" software which is also available from data.nist.gov. (https://doi.org/10.18434/T4/1426859). Note: Certain commercial materials are identified in this dataset in order to specify the experimental procedure adequately. Such identification is not intended to imply recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to imply that the materials are necessarily the best available for the purpose.

Resources

Name Format Description Link
47 https://data.nist.gov/od/ds/mds2-2290/Phi_L100P300_sim_All.csv.sha256
47 https://data.nist.gov/od/ds/mds2-2290/Imean_L100P300_exp.csv.sha256
47 https://data.nist.gov/od/ds/mds2-2290/labels_L100P300.csv.sha256
47 https://data.nist.gov/od/ds/mds2-2290/y_L100P300_Die1_exp.csv.sha256
47 https://data.nist.gov/od/ds/mds2-2290/invV_L100P300_Die1_exp.csv.sha256
0 84 x 84 matrix that is formatted for use with M.o.R. It is the inverse of the V matrix in the regression. It corresponds to the same die as that for "y_L100P300_Die1.csv". Unitless. https://data.nist.gov/od/ds/mds2-2290/invV_L100P300_Die1_exp.csv
47 https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_467.csv.sha256
47 https://data.nist.gov/od/ds/mds2-2290/Phi_L100P300_sim_467.csv.sha256
0 467 x 84 matrix of simulated intensities from the scattering off a periodic structure parameterized using a double-trapezoid cross-section. Normalized and unit-less. Formatted for use in M.o.R., this is a subset of the larger dataset. https://data.nist.gov/od/ds/mds2-2290/Phi_L100P300_sim_467.csv
0 140 x 3 matrix of parametric values used as inputs the scattering code for simulation. Values in nanometers. This is a subset of the larger dataset formatted for use in M.o.R., sized for use in machine learning. https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_140.csv
47 https://data.nist.gov/od/ds/mds2-2290/Phi_L100P300_sim_140.csv.sha256
0 140 x 84 matrix of simulated intensities from the scattering off a periodic structure paramerized using a double-trapezoid cross-section. Normalized and unitless. This is a subset of the larger dataset formatted for use in M.o.R., sized for use in machine learning. https://data.nist.gov/od/ds/mds2-2290/Phi_L100P300_sim_140.csv
34 Image illustrating the sampling of the full three-parameter space using the full set of 2566 simulations and the smaller subset of 467 simulations. https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_Comparison.png
47 https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_20_through_140.png.sha256
34 Image illustrating the 140 x 3 sampling of the full three-parameter space as it developed by multiples of 20 using a Halton sequence. https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_20_through_140.png
0 2566 x 84 matrix of simulated intensities from the scattering off a periodic structure parameterized using a double-trapezoid cross-section. These intensities are normalized and unit-less. https://data.nist.gov/od/ds/mds2-2290/Phi_L100P300_sim_All.csv
0 https://doi.org/10.18434/mds2-2290
0 2566 x 3 matrix of parametric values used as inputs the scattering code for simulation. Values are in nanometers. https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_All.csv
0 467 x 3 matrix of parametric values used as inputs the scattering code for simulation. Values in nanometers. Formatted for use in M.o.R., this is a subset of the larger dataset. https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_467.csv
47 https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_All.csv.sha256
47 README file describing these data files, both experimental and simulation, and the relationships among them. https://data.nist.gov/od/ds/mds2-2290/scatter1D_readme.txt
47 https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_Comparison.png.sha256
47 https://data.nist.gov/od/ds/mds2-2290/Pi_L100P300_sim_140.csv.sha256
0 3 x 84 matrix of labels assignable to the 84 columns. Rows correspond to linear polarization state, plane of incidence, and angle of incidence (in degrees). https://data.nist.gov/od/ds/mds2-2290/labels_L100P300.csv
0 1 x 84 vector formatted for use with our group's published regression software, M.o.R (http://doi.org/10.18434/T4/1502429). It is a subset of "Imean_L100P300_exp.csv" https://data.nist.gov/od/ds/mds2-2290/y_L100P300_Die1_exp.csv
0 9 x 84 matrix of measurements determined from repeated measurements of the intensities scattered off the nine measured "L100P300" targets, specifically within nine dies on a single wafer. Each row of intensities corresponds to a single die. These intensities are normalized and unit-less. https://data.nist.gov/od/ds/mds2-2290/Imean_L100P300_exp.csv

Tags

  • inverse-problems
  • scattering
  • microscopy
  • experimental
  • scatterfield-microscopy
  • electromagnetic-simulations
  • angle-resolved-scattering
  • bright-field-microscopy
  • gratings
  • simulations
  • patterned-semiconductors
  • machine-learning
  • semiconductors

Topics

Categories