Optics: measuring and testing – Inspection of flaws or impurities – Surface condition
Reexamination Certificate
2001-03-12
2004-07-06
Lee, Michael G. (Department: 2876)
Optics: measuring and testing
Inspection of flaws or impurities
Surface condition
C356S237300, C356S327000
Reexamination Certificate
active
06760100
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to methods and apparatus for optically detecting defects in or on a smooth surface of a substrate such as a silicon wafer, and for classifying the defects in terms of defect type and size.
BACKGROUND OF THE INVENTION
Optical inspection methods are frequently used for inspecting the quality of a smooth surface of a substrate such as a silicon wafer, computer disk, or the like. In most such inspection systems, the surface is impinged with a beam of laser light and the light scattered and reflected from the surface is collected and converted into electrical signals that are analyzed so as to infer the presence and size of certain defects on the surface. At least in the case of optical inspection of silicon wafers that are used as the starting material for making integrated circuit chips, the types of defects of major concern include particles of foreign materials on the surface, pits formed in the surface, scratches in the surface, and others.
Particles on the wafer surface can interfere with the lithography process by which lines of electrically conductive material are formed on the surface. As a general rule, any particle whose diameter is larger than half the width of the electrical lines to be laid onto the surface constitutes an unacceptable defect. If there are too many such particles, the wafer must be rejected. Currently, integrated circuits are being made with line widths as small as 0.25 &mgr;m (250 nm), so that particles larger than 125 nm in diameter occurring on the wafer surface would be cause for rejecting the wafer, while particles smaller than 125 nm would be tolerable. The semiconductor industry is quickly moving towards production of circuits composed of 0.18 &mgr;m and then 0.15 &mgr;m lines, which means that much smaller particles will soon cause concern.
Wafer inspection systems must be calibrated in order to function properly to accurately determine the diameter of a particle. The calibration is typically done by intentionally placing a plurality of particles of various known diameters on the wafer surface and inspecting the wafer with the inspection apparatus, so that the scattered light intensities produced from the various sizes of particles can be correlated to the particle sizes. These calibration particles are usually spheres made of polystyrene latex (PSL).
One difficulty that has been encountered in wafer inspection processes is that identically sized particles of different material types can produce substantially different scattered light intensities. Stated differently, two particles made of different materials and having substantially different diameters may produce virtually the same measured scattered light intensity. For example, it has been found that silicon particles of a given diameter will produce a much larger scatter intensity than the same diameter PSL particle. In fact, among the various types of materials that can commonly appear on a wafer surface in the form of particles, PSL particles tend to be one of the lowest sources of light scattering. Thus, after the inspection apparatus has been calibrated with PSL particles, the apparatus will tend to overestimate the diameters of silicon particles and those of many other materials. Accordingly, wafers are rejected as having particles larger than half the line width, even though in reality the particles may be smaller than half the line width. Therefore, the accuracy with which particles can be sized by light scattering can be greatly increased if something is known about the particle material.
Another advantage to the semiconductor industry in being able to identify particle material is that this information provides a strong clue as to the source of the contamination. Because particle contamination has to be reduced to a level where useful product can be produced, finding and eliminating contamination sources quickly is economically important.
For light of a given wavelength, every material has an index of refraction, which indicates how much the speed of light is reduced within the material, and an absorption coefficient, which is generally indicative of how opaque the material is to the light. The combination of the index of refraction and absorption coefficient, which are known as the material constants, is unique for each different material. The combinations of these material constants can be roughly separated into four groups: (1) dielectrics such as PSL, SiO
2
, and Al
2
O
3
(low index of refraction and zero absorption coefficient); (2) semiconductors such as silicon (large index of refraction and small absorption coefficient); (3) gray metals such as tungsten (large index of refraction and large absorption coefficient); and (4) good conductors such as silver (small index of refraction and large absorption coefficient).
The combination of particle material, along with particle shape and the material constants for the substrate surface (which are known), completely and uniquely define the pattern of light scattered by the particle for any given light source. Moreover, for particles whose average diameter is less than about one-fifth of a wavelength of the illuminating light, the particle shape does not play a significant role in determining the scatter pattern. Thus, for visible light and particles smaller than about 100 nm, knowledge of the average particle diameter and the various material constants are enough to calculate the scatter pattern for a given scattering geometry. This fact has allowed the development of scattering models that predict scatter patterns for a given set of conditions. These models have been experimentally confirmed and the results published.
What would be desirable is a system and method for solving the more difficult inverse problem. That is, it would be desirable to be able to determine the particle material and average particle diameter from a knowledge of the scatter pattern. Heretofore, methods have been developed for determining average particle diameter by analyzing the scatter pattern, for example as described in commonly owned U.S. Pat. No. 5,712,701, which is hereby incorporated herein by reference. However, as noted above, the accuracy of such methods depends on the calibration of the system, and currently the calibration must be performed using PSL spheres, which have substantially different material constants from some of the other materials that can appear as particles on a wafer.
Methods for identifying particle material have been proposed. For instance, U.S. Pat. No. 5,037,202 to Batchelder et al. discloses methods and apparatus in which two parallel light beams that are initially mutually coherent but of different polarizations are focused onto a focal plane (such as the surface of a wafer) such that they are displaced apart from each other at the focal plane. After the beams are reflected from the surface, a further optical system intercepts the beams and combines them so that a particle-induced phase shift in one of the beams is manifested by a change in the elliptical polarization of the combined beams. A first detector is responsive to the combined beam's intensity along a first polarization axis to produce a first output, and a second detector is responsive to the combined beam's intensity along a second polarization axis to produce a second output. The first and second outputs are added to provide an extinction signal and are subtracted to provide a phase shift signal. The phase shift and extinction are correlated with index of refraction of the particle material, and hence the identity of the material purportedly can be determined based on the phase shift and extinction values. The size of the particle purportedly can be inferred from its position on a curve of extinction versus phase shift. Thus, in Batchelder's system and method, information about the particle is inferred by analyzing the specularly reflected beams. A disadvantage of this approach is that the reflected light is relatively insensitive to changes in particle properties, such that small particles
Ivakhnenko Vladimir I.
Stover John C.
ADE Corporation
Kim Ahshik
Lee Michael G.
LandOfFree
Method and apparatus for classifying defects occurring at or... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for classifying defects occurring at or..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for classifying defects occurring at or... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3215009