Optics: measuring and testing – Photometers – Integrating spheres
Reexamination Certificate
2002-05-10
2004-03-02
Pham, Hoa Q. (Department: 2877)
Optics: measuring and testing
Photometers
Integrating spheres
C356S600000, C250S228000
Reexamination Certificate
active
06700657
ABSTRACT:
THE FIELD OF THE INVENTION
The present invention relates to non-contact inspection and characterization of a surface. More particularly, the present invention relates to capturing and segmenting portions of scattered light to characterize surface roughness.
THE BACKGROUND ART
The ability to accurately measure physical properties of a surface is important in a variety of applications. Such physical properties include roughness, texture, waviness, and information relating to the profile of the surface. The measure of such physical properties is generally referred to as “characterizing” a surface.
In the field of computer hardware, computer hard disks are manufactured with a known roughness, generally referred to as “texture.” The texture may be implemented as circumferential scratches. Hard disks are typically textured by a process that yields the circumferential scratches through use of an oscillating rotating pad. Other than the circumferential scratches, the hard disk surface is relatively smooth. As a quality control measure, hard disk manufacturers desire a measurement device which permits them to quickly and easily measure the surface roughness as precisely as possible. Current technology trends are moving toward surface texture levels requiring surface measurement down to about the
10
Angstrom level. It would be preferable if surface roughness could be measured to within 1 Angstrom or less.
Other applications where precise roughness measurements are desirable include the computer chip wafer industry. In manufacturing chip wafers, it is desirable that the front surface of the wafer be as smooth as possible and that the back side of the wafer be finished to a known roughness.
Also, the optical industry, particularly mirror manufacturers, desires high-precision measurement devices to gauge the quality of the surfaces of their optics. Such optics are typically employed in imaging systems such as those utilized in telescopes and satellites.
Some surface characterization instruments operate by contacting the surface. A profilometer is an example of such a device. A profilometer operates by dragging a stylus across a surface. The stylus is physically connected to a recorder which traces the profile of the surface. Mathematical analysis of the profile may be conducted to determine physical properties of the surface.
For many applications, such contact-based instruments and methods are unacceptable because of the risk of contamination or other damage to the surface. Additionally, they are extremely slow and do not provide sufficient resolution to be effective for use in many applications. Thus, there exists a need for non-contact surface characterization devices and methods.
Surface inspection devices based on optics have generally proved to be effective at non-contact surface characterization. Such optical devices typically operate by directing a beam of light at the surface and measuring the amount and direction of non-specular light scattered off the surface. Through the analysis of such data, much information regarding the character of the surface can be ascertained. This information includes roughness, texture, waviness, and information relating to the profile of the surface.
One such non-contact, optical-based device is the scatterometer. To measure roughness, for example, the scatterometer measures the scatter intensity of the scattered light at every scatter angle in a selected plane. This information can then be used to generate the “power spectral density” function for that plane. The power spectral density function illustrates the distribution of the power scattered by each spatial frequency. The roughness of the surface can then be approximated by integrating the power spectral density function.
The scatter of a surface is distributed throughout an upper hemisphere above the surface receiving a beam of light. One disadvantage to the use of such scatterometers is that because the scatterometer measures only one plane of the scatter hemisphere, only a small portion of the total information about the surface is obtained. If the surface has an isotropic roughness, such methods are generally accurate. For isotropic surfaces, the total roughness is determined by performing three-dimensional integration on the power spectral density function. However, if the surface is non-isotropic, wherein anisotropic structures are present on the surface such that the surfaces have a “lay” to them or randomly rough surfaces, a scatterometer may produce grossly inaccurate results.
One method for characterizing non-isotropic surfaces is to measure the scatter intensity at every point in the scatter hemisphere. The sample data can then be manually integrated to determine the roughness. Such a method can be performed with an “out-of-plane” scatterometer. This method is extremely time consuming and is therefore not practical for most applications which require rapid inspection and analysis.
In an attempt to make scatter measurement more efficient and versatile, it has been noted that plotting the power spectral density versus the spatial frequency on a log-log plot will generally result in a straight-line curve. Thus, by obtaining two representative points on this line, the curve can be approximated. By integrating this function over selected spatial frequency limits, surface roughness can be determined.
One difficulty with this process is that the power spectral density data is two dimensional; thus, the process only works well for isotropic surfaces. Additionally, because of the limitations on the physical size of the detector, the representative points used to generate the curve are close together. Hence, any noise in the data could substantially decrease the accuracy of the fit of the curve.
Measuring additional data points to improve the fit of the curve becomes difficult because of the complexity of the necessary instrumentation. Additionally, the math to include additional data points becomes unduly complicated. Also, the inclusion of more data points still does not account for non-isotropic variations in the surface. Thus, attempting to add additional data points to improve the curve fit is not viable for many applications.
A conventional method for characterizing non-isotropic surfaces is the “total integrated scatter” method. According to this method, an optical integrating device, such as a hollow sphere, generally referred to as an “integrating sphere,” is placed over the surface of the sample. The integrating sphere has an input aperture through which a beam of light may be directed into the device. A sampling aperture on the other end of the sphere permits the light to be directed onto the surface and allows light scattered off the surface to enter the sphere. An output aperture is also configured into the sphere for permitting the reflected specular beam to exit the sphere. Thus, the light scattered off the surface remains within the sphere and its intensity can be measured with a detector. This method measures most of the scattered light regardless of variations in the surface. However, some scattered light does escape from the output aperture and is therefore not measured.
An additional component which improves the collection of scattered light is the use of a second optical integrated device, such as a focusing mirror. The focusing mirror is disposed and configured to capture scattered light reflected off the surface which passes through the output aperture of the integrating sphere. The scattered light reflected by the mirror is focused to a detector which measures the intensity of this portion of scattered light. In this manner, substantially all of the scattered light is collected for measuring the total integrated scatter. Such an apparatus and method is disclosed in U.S. Pat. No. 5,625,451 to Schiff et al. which is hereby incorporated by reference.
The total integrated scatter method is useful in determining the total integrated scatter reflected from a surface and allows for a detailed analysis of the roughness of the surface. However, it would be advantageous to be a
Pham Hoa Q.
Schmitt Measurement Systems, Inc.
Stoel Rives LLP
Thompson John R.
LandOfFree
System and method for detecting surface roughness does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for detecting surface roughness, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for detecting surface roughness will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3252495