Method and apparatus of measuring line structures with an optical microscope by data clustering and classification

Claim 1

The method of measuring line-widths of a structure on the surface of an object with an optical microscope comprising the steps of obtaining an intensity image of optical signals at a plurality of positions, each position defined by a horizontal position x along an x-axis and focus position z along a z-axis; obtaining a cloud plot of said intensity images as a function of horizontal position x and focus position z; linescanning said cloud plot in the z direction at positions along the horizontal line to provide linescan vectors; clustering and classifying the linescan vectors into groups representing raised and non-raised portions on a surface of said structure on the surface of the object, including defining centroid of each group, and creating a similarity curve from the linescan vectors and centroids which is a function of position along the horizontal line wherein a first attribute of said curve indicates a raised surface and a second attribute of said curve indicates a non-raised surface; and measuring the distance between points in the similarity curve which are along a preselected line parallel to the abscissa of the similarity curve to give a measure of the width of the raised or non-raised surface.

Claim 2

The method of measuring line-widths of a structure on the surface of an object with an optical microscope comprising the steps of obtaining an intensity image of optical signals at a plurality of positions, each position defined by a horizontal position x along an x-axis and focus position z along a z-axis; obtaining a cloud plot of said intensity images as a function of horizontal position x and focus position z; linescanning said cloud plot in the z direction at positions along the horizontal line to provide linescan vectors; clustering and classifying the linescan vectors into groups representing raised and non-raised portions on a surface of said structure on the surface of the object, including defining centroid of each group, and creating a similarity curve from the linescan vectors and centroids which is a function of position along the horizontal line wherein a first attribute of said curve indicates a raised surface and a second attribute of said curve indicates a non-raised surface; and measuring the distance between points in the similarity curve which are along a preselected line parallel to the abscissa of the similarity curve to give a measure of the width of the raised or non-raised surface.

Claim 13

The method of measuring line-widths of a surface of a semiconductor device with an optical microscope comprising the steps of obtaining optical images of the surface of said semiconductor device, said optical images having intensity signals representing the surface of the semiconductor device as the semiconductor device is moved in a z direction, and forming from said optical images a cloud plot, which represents intensity of said optical images as a function of a position x along a horizontal line and a focus position z along a z-axis, obtaining linescans from said cloud plot at various focus positions z for a given position x from an area surrounding a focal spot and weighing said linescans to obtain a linescan column vector, obtaining linescan column vectors at different positions x along the horizontal line, clustering and classifying the linescan column vectors into groups representing lines and trenches by defining a centroid of each group and creating a similarity curve from the linescan column vectors and the centroids, said similarity curve being a function of position along the horizontal line wherein a first attribute of said curve indicates a line and a second attribute of said curve indicates a trench, measuring the distance between adjacent points in the similarity curve which are along a selected line parallel to the horizontal line to give a measure of the width of the line or trench.

Claim 14

The method of measuring line-widths of a surface of a semiconductor device with an optical microscope comprising the steps of obtaining optical images of the surface of said semiconductor device, said optical images having intensity signals representing the surface of the semiconductor device as the semiconductor device is moved in a z direction, and forming from said optical images a cloud plot, which represents intensity of said optical images as a function of a position x along a horizontal line and a focus position z along a z-axis, obtaining linescans from said cloud plot at various focus positions z for a given position x from an area surrounding a focal spot and weighing said linescans to obtain a linescan column vector, obtaining linescan column vectors at different positions x along the horizontal line, clustering and classifying the linescan column vectors into groups representing lines and trenches by defining a centroid of each group and creating a similarity curve from the linescan column vectors and the centroids, said similarity curve being a function of position along the horizontal line wherein a first attribute of said curve indicates a line and a second attribute of said curve indicates a trench, measuring the distance between adjacent points in the similarity curve which are along a selected line parallel to the horizontal line to give a measure of the width of the line or trench.

Claim 16

The method as in claims 12 or 13 in which the linescan vectors are transformed prior to clustering and classification.

Claim 21

A method of measuring line-widths of a structure on the surface of an object with an optical microscope comprising the steps of: obtaining an intensity image of optical signals at a plurality of positions, each position defined by a horizontal position x along a horizontal line and a focus position z; linescanning a cloud plot of the intensity images in the z direction at each position x to provide linescan vectors, clustering and classifying said linescan vectors into groups, said groups representing raised and non-raised portions on said structure on the surface of the object, including defining a centroid of each group, and creating a similarity curve from the linescan vectors and centroids which is a function of position along the horizontal line wherein a minimum of said curve indicates a raised surface and a maximum of said curve indicates a non-raised surface, reiterating the clustering and classifying step by using an improved estimate of the centroid of each group, and measuring a width of said raised and non-raised portions from said similarity curve.

Claim 22

A method of measuring line-widths of a structure on the surface of an object with an optical microscope comprising the steps of: obtaining an intensity image of optical signals at a plurality of positions, each position defined by a horizontal position x along a horizontal line and a focus position z; linescanning a cloud plot of the intensity images in the z direction at each position x to provide linescan vectors, clustering and classifying said linescan vectors into groups, said groups representing raised and non-raised portions on said structure on the surface of the object, including defining a centroid of each group, and creating a similarity curve from the linescan vectors and centroids which is a function of position along the horizontal line wherein a minimum of said curve indicates a raised surface and a maximum of said curve indicates a non-raised surface, reiterating the clustering and classifying step by using an improved estimate of the centroid of each group, and measuring a width of said raised and non-raised portions from said similarity curve.

Claim 24

A system for measuring line-widths of a structure on the surface of an object comprising: a translation stage that moves in any of three mutually perpendicular directions; image acquisition means for acquiring an optical image at a plurality of positions, each position representing at least a portion of the object, wherein each optical image is converted to a binary signal and multiple bits of the binary signal represent a picture element of the optical image; storage means for storing bits of the binary signal; a cloud plot generator for generating a cloud plot from the optical images; linescan vector generator to linescan the cloud plot and store optical image intensities in vector form at each linescanned position; clustering and classifying means for data clustering and classifying the vectors into groups representing raised and non-raised structures on the surface of the object and generating a similarity curve; and output display means for displaying the distance between two preselected points on the similarity curve to represent the width of the raised or non-raised structures.

Claim 25

A system for measuring line-widths of a structure on the surface of an object comprising: a translation stage that moves in any of three mutually perpendicular directions; image acquisition means for acquiring an optical image at a plurality of positions, each position representing at least a portion of the object, wherein each optical image is converted to a binary signal and multiple bits of the binary signal represent a picture element of the optical image; storage means for storing bits of the binary signal; a cloud plot generator for generating a cloud plot from the optical images; linescan vector generator to linescan the cloud plot and store optical image intensities in vector form at each linescanned position; clustering and classifying means for data clustering and classifying the vectors into groups representing raised and non-raised structures on the surface of the object and generating a similarity curve; and output display means for displaying the distance between two preselected points on the similarity curve to represent the width of the raised or non-raised structures.