Dark Field Surface Inspection Illumination Technique

1. Field of the Invention

This invention pertains to methods of evaluating optical scratches and other damage to coated and uncoated optical samples. In particular, the present invention pertains to an apparatus and method permitting detailed scanning throughout the depth of an optical sample to identify the exact location of imperfections within the sample.

2. Description of the Prior Art

Previous techniques for evaluating the quality of optical samples have included:

a. Straight through or bright field technique which does not discern subtle contrasts,

b. Normarski or back reflection technique which does not look at the subsurface but only the surface scratches, and

c. the dark field or grazing angle reflection technique. None of these techniques provide information on the size of small imperfections beneath 500 angstroms or the position of such imperfections.

Previous internal microscopy has been limited to thin plates with beveled edges. This technique creates numerous reflections within the plate. Since each reflection must lose some light, this method is limited to loss spectroscopy.

A well collimated beam of light such as linerally polarized light from a laser is transmitted into a prism. When the light reaches the exit side of the prism, it encounters a layer of index matching fluid which it transmits through and an optical sample which it continues to transmit through. The optical sample, index matching fluid and prism are all of similar index of refraction to permit continued transmittance of the light. The purpose of the arrangement is to permit light to be present within the optical sample at or greater than the critical angle whenever it reaches the surface. Upon reaching the surface of the optical sample the light undergoes total internal reflection off of the surface because it is at the critical angle and is transmitted back through the sample, index matching fluid and prism. The light exits from the prism.

For an optical sample without any imperfections, viewing the surface of the sample above the area of internal reflection will show a dark field. However, any irregularities within the sample or on the surface of the sample will scatter the light. Some of this light will be scattered towards the surface of the sample at a non-critical angle. This light will emerge and be seen as a bright feature. The polarization of the light is dependent on the nature of the scatterers. Observation of such features throughout the optical sample permits a detailed analysis of the imperfections of the sample. A viewing device ideal for such observations is a microscope. A polarization analyzer can be placed between the microscope and the sample to determine the state of polarization of the scattered light. The sample can be moved over the prism to permit all areas of the sample to be observed.

The linearly polarized light within the optical sample will establish a standing wave pattern. As described herein, the standing wave pattern properties occur for monochromatic polarized light. This pattern will have peaks and nulls within the depth of the sample. If a very small imperfection occurs at a point where a null occurs, it will not be observed. For a small imperfection to be observed, it will have to be illuminated with the peak of a standing wave pattern. For larger imperfections, the site which is positioned at a null will remain visible because it is larger than the null. The change in visibility is used to determine the size of the imperfection. The peaks and nulls within the sample can be alternated by changing the polarization of the light present. Such polarization changes will also give an indication of the depth of any imperfection within the sample. A further refinement of identification of imperfections can be obtained by coating the surface of the optical sample with a drop of oil. This oil will again have an index of refraction similar to the sample. The oil will fill all scratches on the surface of the sample and effectively eliminate them. Thus a second viewing of the sample will only show imperfections which are imbedded in the sample and omit surface scratches.

An alternate technique is to tune the monochromatic light to different monochromatic wavelengths and thus change the standing wave pattern. As the peaks and nulls are shifted by the changing wavelength, features will appear and disappear as the peaks and nulls change depth within the sample. An alternate way of using the same effect is to change the angle of orientation of the light within a sample to angles greater than the critical angle. This changes the distribution of the peaks and nulls of the standing wave by broadening them over a larger scale. This last technique coupled with the use of an oil drop on the surface will permit peaks and nulls to be brought up through the surface of the optical sample itself.

FIG. 1 is a diagram of a preferred embodiment for the present invention.

FIG. 2 is a diagram of the change in peaks and nulls obtained by changing polarity.

FIGS. 3 and 4 show alternate methods of coupling light into an optical sample.

FIG. 5 is a graph of surface light intensity vs. angle of incidence for external and internal illumination of an optical sample.

FIG. 1 is an embodiment of the present invention which permits surface irregularities of an optical sample 10 as well as imbedded imperfections in optical sample 10 to be viewed by a dark field technique. The general term defects will refer to all possible observed features including imbedded imperfections, surface scratches and contaminents on the surface. Linearlly polarized monochromatic light from a tunable laser 12 is transmitted along optical path 14. For unpolarized light, path 14 can first pass through a polarizer 13. For a polarized source the beam passes through a Babinet-Soleil compensator 15 which rotates the polarization of the light to a predetermined orientation. The beam of light travelling along optical path 14 then reflects off a mirror 22 and enters a prism 16 which has an index of refraction similar to or identical to optical sample 10. Mirror 22 is a steering device that controls the angle .theta. shown in FIG. 1. The use of prism 16 permits the light to enter the prism at a non-critical angle. Once the light is transmitted within prism 16, it will encounter the second surface of the prism. To avoid reflection, a layer of index matching fluid 18 is used to coat one surface of prism 16. The thickness of fluid 18 in FIG. 1 is greatly exaggerated for purposes of example. Light travelling along path 14 now continues into index matching fluid 18. On top of index matching fluid 18, optical sample 10 is placed. Optical sample 10 now has light enter it along path 14. The light within optical sample 10 now encounters the upper surface of sample 10 at an angle which is equal to or greater than the critical angle for total internal reflection. If optical sample 10 has no defects in the illuminated area, a person observing the surface area of optical sample 10 from an observation point 20, which can be a microscope, will see nothing. As shown in FIG. 1, observation point 20 views light exiting normal to the surface of optical sample 10. Index matching fluid 18 also serves as a lubricant to permit movement of sample 10 over prism 16 so all of sample 10 can be observed except for a narrow edge about sample 10. An alternate method of viewing all of sample 10 is to move prism 16. Plane surfaces on the bottom of sample 10 and the top of prism 16 are exemplary. Matching contoured surfaces can be used to study lenses and such.

The viewing of no light gives rise to the title dark field. However, if there are surface scratches or imbedded imperfections, light travelling along path 14 will undergo scattering. Some of the scattered light will be directed upwards out of optical surface 10 and encounter the upper surface at less than the critical angle. This scattered light will be emitted from optical surface 10 and scratches and imperfections will show up as a pattern of bright spots against a dark background. A polarization analyser 21 can be placed between sample 10 and observation point 20 to determine the state of polarization of the scattered light.

Surface scratches can be separately identified from imbedded imperfections by raising the upper surface of optical sample 10 with a coat of material of matching index of refraction. The coating will provide a smooth surface on optical sample 10. The coating fills in all surface scratches and raises the surface where total internal reflection occurs. Any surface scratches previously viewed will now be eliminated and the only imperfections seen will be those imbedded in the sample. Care should be taken that the coating is a uniform thickness and that it does not contain dust or other impurities. Impurities would appear as new bright spots. The coating can be a drop of oil or any other suitable material deposited on the sample. On a known optical sample, this method can be used to view impurities such as microbes on the surface of the sample or suspended within the coating.

FIG. 2 shows the standing wave patterns of linearlly polarized light near the surface of optical sample 10. Line 30 represents the upper surface of optical sample 10. Line 31 represents a level a few microns beneath line 30. The standing wave pattern exits in the region beneath the surface of optical sample 10 where the incoming beam and reflected beam overlap. White light obviously cannot be used because the coherence length for interference is so short. In contrast, monochromatic light has large coherence lengths which provide the necessary depth to search for small imbedded imperfections. While this region is small compared to the depth of the sample, it contains thousands of standing waves. Standing wave pattern 32 and standing wave pattern 34 are compared to the reference level of light itensity, lines 32a and 34a. Wave patterns 32 and 34 represent "s" and "p" polarization respectively. Peaks and nulls of the standing wave patterns are alternated at a given wave length. One polarization, curve 32, has a maximum intensity at the surface of the sample while the alternate polarization 34 has a null at the surface of the sample. If the incident light has an intensity of I.sub.o and it enters the sample for curve 32, the surface intensity will be 4I.sub.o and for curve 34 the peaks will be approximately 2I.sub.o. Thus, it can be seen that only one component of polarization illuminates small imperfections when viewed from directly above. As a result, a small imperfection, such as 36, will not be seen for polarization 32 but will be brightly illuminated for polarization 34. The result of this change in information by changing the polarization is to provide a way to measure the depth of imperfection 36 within the optical sample. For the standing wave pattern shown, for any given polarization, there will be several possible depths for which imperfection 36 would appear. For polarization 32, there are several possible depths at which an imperfection could be illuminated. To identify which of the locations is the correct one, the laser light coming along path 14 can be tuned to cause the standing waves to vary their nulls and peaks. As the standing wavelength increases, the possible location of imperfection 36 will consist of a different set of values. Location of imperfections can be identified this way. In a real sample, several different wavelengths will have to be used to isolate the impurity. Also as previously discussed, surface irregularities or scratches can be eliminated as a possible area of consideration by using a drop of oil on the surface.

Another way of identifying the depth of the imperfections within the optical sample is to change the incident angle to one greater than the critical angle. When this is done, the standing wave pattern expands. By using the same polarization, but expanding the peaks and nulls, imperfections can effectively be tuned in and out of a peak and null for a given polarization. This can be done in FIG. 1 by mirror 22 which can be used to deflect the beam to angles greater than the critical angle.

Lastly, combining two of these methods, placing an oil drop on the upper most surface and then changing the incident angle permits the upper most null to go all the way to the surface of the oil drop. Without the oil drop, the upper most null shown for polarization 34 can never get closer to the surface and only expand away from it. With the oil drop, the upper most null can actually be shifted above the optical sample itself and into the layer provided by the oil drop.

FIG. 3 shows an alternate way of coupling light into optical sample 10. Two prisms 40 are used to couple light in and out as described previously. The difference is that now the bottom of optical sample 10 is not obstructed. Light travelling along optical path 14 now undergoes several total internal reflections. The advantage of the FIG. 3 arrangement is that see-through laser evaluation of a sample can be made at various stages of a test lifetime without repeated disassembly for testing. For each internal reflection point, light will be scattered both up and down. Thus, two possible observation points 20 are possible.

FIG. 4 shows a dry technique for coupling light into optical sample 10. A soft plastic rounded rod 44 is slightly deformed by contact with optical sample 10. This permits controlled access and egress points for light along optical path 14.

FIG. 5 is a graph which shows the difference in the surface intensity of light, I.sub.o, vs angle of incidence for internal and external reflection. The critical angle is .theta..sub.c. In this graph an index of refraction, n=1.47, for fused quartz has been used. The polarization shown is S-polarization which corresponds to curve 32 in FIG. 2. The amount of light available to illuminate defects is far greater for internal reflection at or near the critical angle, .theta..sub.c. In addition, the external light does not create a useful standing wave pattern as internal light does.

The above invention can be used with many optical areas and particularly has been used with visible light for glass optics. However, it is clear to those skilled in the art that this method will be applicable to a wide range of optical areas.