Optics: measuring and testing – By light interference – For dimensional measurement
Reexamination Certificate
2001-06-06
2003-09-23
Kim, Robert H. (Department: 2882)
Optics: measuring and testing
By light interference
For dimensional measurement
Reexamination Certificate
active
06624893
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention is related in general to the field of scanning interferometry and, in particular, to a novel approach for correcting motion errors in the scanning arm of an interferometer.
2. Description of the Related Art
Conventional scanning interferometry utilizes a light source, such as white light or a laser beam, to produce interference fringes at a light detector as the optical path difference (OPD) between a test surface and a reference surface is varied during a vertical scan. All measurement techniques based on phase-shifting interferometry (PSI), vertical-scanning interferometry (VSI, also referred to as white-light interferometry), and phase-shifting interferometry on-the-fly (PSIOTF) rely on an analysis of the interference between two beams of light. One beam (the object beam) is reflected from the sample; the other beam (the reference beam) is reflected from the reference mirror; and the two beams are recombined to create an interference pattern (the interferogram) at each scanning step. A detector (usually a CCD camera) registers the interferogram in a number of frames while the optical path difference (OPD) between the interfering beams is changing in a predefined fashion, which is realized either by moving the object, or by scanning the interferometric objective or the reference mirror at a preferably constant speed. The resultant shape of the sample object is calculated based on the intensity patterns in the interferograms and can be used to describe the height profile of the sample.
PSI is preferably used for measurements of smooth surfaces with small changes in profile (see K. Creath, “Temporal Phase Measurement Methods,” Interferogram Analysis, Institute of Physics Publishing Ltd., Bristol, 1993, pp. 94-140). VSI is generally used to measure smooth and/or rough surfaces with large interpixel height ranges (K. G. Larkin, “Efficient Nonlinear Algorithm for Envelope Detection in White Light Interferometry,” J. Opt. Soc. Am., A/Vol. 13, 832-843 (1996). The combination of VSI and PSI in used in the PSIOTF technique to improve measurements of smooth surfaces in the medium height range (A. Harasaki et al., “Improved Vertical Scanning Interferometry,” Appl. Opt. 39, 2107-2115, 2000).
In VSI the interference fringes are localized only in a small region around the focus because of the low coherence source or the high numerical aperture of the microscope objective employed. While scanning through focus, fringes for different parts of the sample surface are produced and analyzed resulting in an unambiguous measurement of the object shape. As well understood by those skilled in the art, the height information from a low-coherence interferogram can be retrieved in many ways, such as, for example, by peak detection of the coherence envelope or fringe, by calculation of the centroid of the intensity signal, or by determination of the slope of the phase of the average wavelength. All these methods of analysis estimate the best vertical-scan focus position for a given pixel by sensing the coherence peak and utilize this information to determine the relative height difference between pixels (that is, a height profile of the test surface).
In PSI techniques the interference fringes are examined only around a single focus position using typically a quasi-monochromatic light source, such as a laser. Algorithms employed to analyze PSI fringes produce phase measurements that may be ambiguous (because of so-called 2&pgr; ambiguities) and therefore require further processing to remove the ambiguity by unwrapping the phase data. Because of this limitation, only objects with relatively small inter-pixel height differences can be measured using conventional PSI analysis. On the other hand, the advantage of PSI versus VSI techniques is that PSI allows for more precise measurements.
The PSIOTF technique affords the high precision of PSI combined with the lack of ambiguity of VSI measurements. Accordingly, this combined technique is often utilized for determining the shape of smooth surfaces with large height differences (that is, surfaces with steep profiles). The procedure first involves finding a coarse focus position for each pixel using a VSI technique; then a PSI algorithm is applied to the intensity data collected at frames around this focus position to achieve a high precision measurement where the 2&pgr; ambiguity has already been resolved by VSI. The PSIOTF technique can be used with a low-coherence light source or with a combination of both a low- and a high-coherence source. In addition, the technique can be applied to each pixel individually or to groups of pixels in separate areas, as in the case of a stepped sample with separate smooth and flat regions.
All of these techniques require a well-calibrated and, preferably, a constant scanning motion of the sample object along the optical axis because the height calculations are based on the scanning distance traveled between interferogram frames. Scanner miscalibration, nonlinearities and vibrations affect the surface-profile measurements produced by PSI, VSI and PSIOTF algorithms by attributing an incorrect size to each scanner step. Thus, the rate of change in the OPD, normally referred to in the art as the “phase step,” needs to be well calibrated in order to achieve an accurate measurement. The most commonly used algorithms for the three interferometric techniques mentioned here require a predefined, nominal phase step which, once the scanner is calibrated, is assumed to remain constant along the whole scan for each measurement. However, this is not always the case and the errors in scanning speed influence the final result and need to be accounted for accurate measurements.
A common example of a source of scanning error is linear phase-step miscalibration, where the phase step is constant but different than the calibrated phase step, which may produce height magnification errors (in VSI), residual ripples (in PSI), or a saw-tooth profile (in PSIOTF). Such effects may be visible in the calculated profile. For example,
FIGS. 1A and 1B
show a saw-tooth effect in the phase map (x and y profiles, respectively) of a chromium-coated step height standard using a PSIOTF technique. Phase step miscalibration was present, causing saw-tooth effects following the fringes. The PSI phase calculated around each frame reflects real changes in the optical path, but the frame number position is only assumed based on the device's calibrated phase step; therefore, a correction is needed for a precise profile measurement. Improved calibration and a feed-back loop control in the scanner system have been the conventional approaches used to reduce this source of error. FIGS.
2
A,
2
B and
3
A,
3
B illustrate similar scanner-error effects produced by VSI and PSI techniques, respectively.
Another example of a scanning-error source is high frequency vibrations, which introduce random errors in the intensity values of the interferograms. They may also produce a decrease of fringe contrast, residual ripples, or a saw-tooth profile in all interferometric measurement techniques. The most common way of reducing of this type of error is through better vibration isolation, such as with air tables and a more rigid structure for the instrument.
Low-frequency vibrations, miscalibration, and nonlinearity in the scanner motion result in erroneous height-difference readings, which may produce ripples, or a saw-tooth profile, or height magnification errors. In practice, these errors cannot be easily corrected. In addition, they are usually not repeatable from measurement to measurement; therefore, they are the most difficult errors to correct. Thus, eliminating or reducing the influence of these errors would be a great advance in the art.
Various studies have characterized the errors associated with scanning perturbations and improved algorithms have been developed to reduce sensitivity to these error sources, but all state-of-the-art corrective techniques require intense calculations that greatly affe
Olszak Artur
Schmit Joanna
Artman Thomas R
Durando Antonio R.
Durando Birdwell & Janke PLC
Kim Robert H.
Veeco Instruments Inc.
LandOfFree
Correction of scanning errors in interferometric profiling does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Correction of scanning errors in interferometric profiling, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Correction of scanning errors in interferometric profiling will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3078108