Classifications

• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
  • G01B11/00—Measuring arrangements characterised by the use of optical techniques
  • G01B11/30—Measuring arrangements characterised by the use of optical techniques for measuring roughness or irregularity of surfaces
  • G01B11/303—Measuring arrangements characterised by the use of optical techniques for measuring roughness or irregularity of surfaces using photoelectric detection means
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
  • G01B11/00—Measuring arrangements characterised by the use of optical techniques
  • G01B11/30—Measuring arrangements characterised by the use of optical techniques for measuring roughness or irregularity of surfaces
  • G01B11/306—Measuring arrangements characterised by the use of optical techniques for measuring roughness or irregularity of surfaces for measuring evenness
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
  • G01B2210/00—Aspects not specifically covered by any group under G01B, e.g. of wheel alignment, caliper-like sensors
  • G01B2210/56—Measuring geometric parameters of semiconductor structures, e.g. profile, critical dimensions or trench depth

Definitions

• wafer processes are carefully honed to ensure that these materials, equipment, and resources are not wasted.
• irregularities in the surface of a wafer may result in less than optimal functioning of the electronic devices formed by the wafer or even non-functional electronic devices .
• manufacture of modern nanometer-scale electronic devices requires the accurate measurement and control of
• edge profile would ideally have no effect on planarity up to a few tenths of a millimeter from the apex of the edge.
• substrate front-surface grinding and polishing operations are performed after macroscopic rounding is imparted to the edge of the substrate. As a result, some amount of edge roll-off is inevitably imparted to the substrate, generally beginning several millimeters from the edge.
• Edge roll-off is generally quantified by such metrics as ESFQR and ZDD, which are known in the industry.
• ESFQR Electronic flatness metric, Sector based, Front surface referenced, least sQuares fit reference plane, Range of the data within the sector
• the flatness is measured within a sector of the wafer, i.e. a fan-shaped area formed on the outer periphery of the wafer.
• Figure 1A illustrates a plurality of sectors 101 on a wafer 100. In one embodiment, 72 sectors at 5 degree intervals can be provided on the periphery of wafer 100.
• Figure IB illustrates a cross-sectional view of wafer 100 and an
• Figure 1C illustrates an exemplary sector 101 in which a plurality of measurements 102 have been taken.
• these measurements 102 include thickness information provided by a sensor and its associated electronics.
• the data from measurements 102 can be averaged so as to provide a
• Typical processing of information from the sensor would yield a profile curve indicating a thickness profile (thickness vs. radius) .
• a profile curve indicating a thickness profile (thickness vs. radius) .
• such a profile would have a plurality of small surface unevenness with larger anomalies classified as bumps or voids (i.e. inverted bumps) .
• the profile of thickness has a gradual roll-off or reduction in thickness as the edge of the wafer is reached.
• Second derivative processing converts the curve to a ZDD profile.
• An exemplary ZDD profile 105 (also called a ZDD metric) is shown in Figure ID.
• FIG. 2 illustrates a simplified inspection system 200 in which a wafer 201 is spun about its center on its axis 202.
• a wafer measurement tool 203 includes a sensor 203A configured to measure the distance to wafer 201.
• the output 204 of 203A can be processed by a data interpreter 205 and a post-processing tool 206 (e.g. a microprocessor) to develop the ESFQR and/or ZDD profile.
• Output information regarding one or more metrics, profiles, or other parameters can be provided to a user interface 207 via standard I/O devices.
• the tolerance limits for ESFQR and ZDD are generally specified by IC device makers .
• CMP chemical- mechanical polishing
• a method of providing high accuracy inspection or metrology in a bright-field differential interference contrast (BF-DIC) system is described.
• This method can include creating first and second beams from a first light beam.
• the first and second beams have round cross-sections, and form first partially overlapping scanning spots radially displaced on a substrate.
• Third and fourth beams are created from the first light beam or a second light beam.
• the third beam and the fourth beam have elliptical cross-sections, and form second partially overlapping scanning spots tangentially displaced on the substrate. At least one portion of the substrate can be scanned using the first and second partially overlapping scanning spots as the substrate is rotated.
• a radial slope can be determined using measurements obtained from scanning the at least one portion of the substrate using the first partially overlapping scanning spots as the substrate is rotated.
• a tangential slope can be determined using measurements obtained from scanning the at least one portion of the substrate using the second partially overlapping scanning spots as the substrate is rotated.
• the radial slope and the tangential slope can be used to determine a
• the method can further include compensating for chucking
• Substrate topography can be determined by applying filtering to either the integrated height or the substrate shape.
• the enhanced BF-DIC technique can be used to determine a wafer shape before and after the deposition of a layer on the wafer.
• a shape difference can be computed based on the wafer shape before and after deposition.
• a film- stress map can be generated based on the shape difference.
• exemplary wafer process includes integrated circuit chemical- mechanical polishing (CMP) .
• CMP integrated circuit chemical- mechanical polishing
• the wafer can be patterned or unpatterned.
• first and second beams are created from a first light beam.
• the first and second beams have round cross-sections, and form first partially overlapping scanning spots displaced in a first
• Third and fourth beams are created from either the first light beam or a second light beam.
• the third and fourth beams have elliptical cross-sections, and form second partially overlapping scanning spots displaced in a second direction, wherein the first direction and the second direction are
• At least one portion of the substrate is scanned using the first and second partially overlapping scanning spots as the substrate is moved.
• a first slope in the first direction can be determined using measurements obtained from the scanning using the first partially overlapping scanning spots as the substrate is moved.
• a second slope in the second direction can be determined using measurements obtained from the scanning using the second partially overlapping scanning spots as the substrate is moved.
• Substrate curvature, edge roll-off, integrated height information, and substrate shape can be determined using the first and second slopes .
• the method can further include
• Substrate topography can be determined by applying filtering to either the integrated height or the
• a bright-field differential interference contrast (BF- DIC) system configured to provide high accuracy inspection or metrology is also discussed.
• the BF-DIC system includes at least one sub-system, wherein each sub-system includes a prism,
• the prism is configured to receive a light beam and generate two beams from the light beam.
• the focusing optics are configured to direct and focus the two beams onto a substrate as two partially overlapping scanning spots.
• the photo-detectors are configured to receive light reflected from the substrate from the two partially overlapping scanning spots.
• the data acquisition circuit is configured to process outputs of the photo-detectors.
• the system can further include an apparatus for securing and moving the substrate, and a computer operatively coupled to the data acquisition circuit and the apparatus for securing and moving the substrate.
• the two beams are radially disposed with respect to the substrate and the focusing optics provide the two beams with round cross-sections .
• the two beams are radially disposed with respect to the substrate and the focusing optics provide the two beams with round cross-sections .
• the at least one sub-system includes first and second sub-systems, wherein the prism of the first sub-system has the first orientation, and the prism of the second sub-system has the second orientation.
• the at least one sub-system includes first and second subsystems, wherein the first and second sub-systems provide
• this enhanced BF-DIC technique can be embodied as an integral sub-system in other types of wafer inspection and metrology equipment well known to those conversant with practices in the semiconductor industry.
• the technique may also be
• each of these sub-system embodiments can be implemented such that enhanced BF-DIC data acquisition occurs substantially in parallel with standard inspection, metrology, or process functions.
• Figure 1A illustrates a plurality of sectors on a wafer .
• Figure IB illustrates a cross-sectional view of a wafer and an exemplary distance for calculating ESFQR.
• Figure 1C illustrates an exemplary sector in which a plurality of measurements has been taken.
• Figure ID illustrates an exemplary ZDD profile.
• Figure 2 schematically illustrates an inspection system in which a wafer is spun about its axis .
• Figure 3A illustrates a portion of an exemplary wafer and four elliptical beams, with one pair of beams displaced in the tangential direction with respect to a spiral scan path on the wafer surface, and the other pair displaced in the radial direction .
• Figure 3B illustrates a portion of an exemplary wafer and two round beams, where the beam displacement is in the radial direction with respect to a spiral scan path on the wafer
• Figure 4 illustrates an exemplary normalized signal S that varies with the surface slope z' .
• Figure 5A illustrates two curves that indicate surface slope as a function of radius r for two types of silicon wafers .
• Figure 5B illustrates curvature profiles (surface curvature ZDD as a function of radius r) derived from the surface slope profiles in Figure 5A.
• Figure 6 illustrates an exemplary enhanced BF-DIC technique for generating an accurate ZDD value.
• Figure 7A illustrates an exemplary processing of information in which both tangential and radial slope information can be used to generate accurate height information of the wafer surface .
• Figures 7B and 7C illustrate radial and tangential slope wafer BF-DIC images, respectively.
• Figure 7D illustrates an exemplary wafer shape image based on the radial and tangential slope wafer images of Figures 7B and 7C.
• Figure 7 ⁇ illustrates an exemplary image showing reference topography measurements for a portion of a patterned wafer
• Figure 7F illustrates a corresponding image showing BF-DIC topography measurements with filtering.
• Figure 8A illustrates an exemplary technique for generating a film-stress map for a wafer.
• Figure 8B illustrates an exemplary substrate with a thin film thereon imparting tensile stress, resulting in concave- up strain of the substrate.
• Figure 9 illustrates a computer-controlled inspection system employing a plurality of BF-DIC sub-systems to collect radial and tangential slope information in a single scan.
• DIC differential interference contrast
• DIC DIC
• a linearly-polarized laser beam is split into two, proximate beams with mutually orthogonal planes of polarization.
• the laser beam can be split into the two beams using a Wollaston prism, which is built from two wedges of a birefringent material having optical axes parallel to the outer surface of the prism, but perpendicular to each other.
• the two beams are focused by other optical elements such as lenses onto the surface of a substrate, wherein the term "substrate” refers to a workpiece of any material composition, including for example a silicon wafer.
• Figure 3A illustrates a Wollaston prism 309 (in a first orientation, 309-1) that uses an input light beam 311 to generate two beams 305 and 306.
• beams 305 and 306 are elliptical (i.e. in cross section).
• Beams 305 and 306 form two, partially overlapping scanning spots 301 and 302 on a wafer 300.
• scanning spots 301 and 302 have minor axes that are co-linear and parallel to the tangential direction, and major axes that are nearly parallel to the radial direction (i.e.
• scanning spots 301 and 302 are positioned on either side of an actual radial direction 312, their major axes are not exactly parallel to the radial direction) .
• scanning spots 301 and 302 have centers that are displaced by approximately one half of the minor axes' length.
• beams 305 and 306 have a beam displacement in the tangential direction with respect to a spiral scan path on the wafer surface.
• the spiral scan path is provided by spinning wafer 300 (i.e. a rotation ⁇ about a center of rotation 313) while linearly translating the center of rotation along a distance equal to or less than the radius r of wafer 300.
• the two scanning spots' major axes are nearly parallel to the radial direction, which shortens inspection time because the pitch between successive spiral scan tracks may be set to approximately one-half of the scanning spots' major axis length.
• scanning spots 301 and 302 reflect from the surface of wafer 300, recombine upon going back through the Wollaston prism 309 in the reverse direction, and create a generally elliptically-polarized beam when the optical path length for one beam is different from that of the other .
• optical path length For substrate materials that are substantially opaque at the DIC system's laser wavelength, optical path length
• the signal S varies with the component of the surface slope along the direction defined by the two displaced scanning spots' centers .
• the functional dependence of S on the surface slope z' is shown in Figur 4. Specifically,
• BF-DIC techniques can detect relatively localized defects in the substrate surface (such as bumps, dimples, pits, and scratches) , as a result of relatively abrupt changes in surface slopes in the neighborhood of such defects.
• the beam displacement direction can be rotated by 90°. In one embodiment, this rotation can be done by rotating Wollaston prism 309 by 90° (i.e. providing a second orientation 309-2), thereby producing radially displaced beams 307 and 308. Beams 307 and 308 form two, partially overlapping scanning spots 303 and 304 on wafer 300. In this orientation, scanning spots 303 and 304 have minor axes that are co-linear and parallel to the radial direction, and major axes that are parallel to the tangential direction.
• Wollaston prism by 90° produces 50% overlap radially.
• the amount of beam displacement can be altered by choosing a Wollaston prism with a different wedge angle.
• a preferred beam displacement is about 50%, because greater displacement imparts stronger noise and smaller displacement entails weaker signal. (In fact, two perfectly overlapped beams produce zero signal everywhere) .
• ERO For the measurement of ERO, even finer resolution in the radial direction is advantageous. Moreover, ERO need only be measured in a narrow annular region near the substrate edge, so elongation of the beams in the radial direction may be dispensed with.
• focusing optics 312 of the inspection/metrology tool can be configured to generate small round (in cross-section) beams 316 and 317 rather than the elliptical beams 307 and 308 ( Figure 3A) .
• beams 316 and 317 may be focused with diffraction-limited optics. Beams 316 and 317 form two, partially overlapping scanning spots 314 and 315 on wafer 300. In this orientation, scanning spots 303 and 304 have centers that are co-linear and parallel to the radial direction. Note that embodiments implementing this fine spatial resolution are
• the range of slopes that the BF-DIC techniques can detect without phase wrapping is a function of the laser wavelength and the beam displacement.
• Figure 5A illustrates two curves 501 and 502 that indicate surface slope as a function of radius r for two types of silicon wafers .
• Figure 5A indicates that for the specific wavelength and beam displacement of Figure 4 , the radial surface slopes in the ERO region fall within the linear response region for an
• substrate i.e. 300 mm silicon wafers.
• substrates with different surface slope ranges may be measured with wavelengths and
• Another aspect of the enhanced BF-DIC technique is the lateral adjustment of the Wollaston prism to null the signal when the surface slope is zero. Because localized substrate defects will cover both positive and negative values of slope, setting the zero-slope response to zero is thus most convenient for an inspection application. However, for the measurement of ERO, the radial surface slope range often covers a single polarity for the most part (see, e.g. Figure 5A) . Therefore, for substrates with extremely negative-going slopes, the Wollaston prism can be adjusted such that its linear response range is biased to accommodate negative polarity more than positive.
• Embodiments of the BF-DIC techniques described herein can have numerical aperture (NA) on the order of 0.0086, which corresponds to a maximum reflection angle of ⁇ 0.0086 rad, and a surface slope ⁇ 4300 nm/mm. Surface slopes in the ERO region of the wafer are of the order of a few 100 nm/mm. Therefore, higher NA optics are unnecessary for the BF-DIC techniques described herein.
• NA numerical aperture
• the slope data may be processed with numerical methods to yield the surface height profile (via numerical integration) or the surface curvature (via numerical differentiation) .
• FIG. 5A illustrates typical curvature profiles 511 and 512 (surface curvature ZDD as a function of radius r) for the two exemplary silicon wafers referenced in Fig 5A.
• Figure 6 illustrates an exemplary enhanced BF-DIC technique for generating an accurate ZDD value.
• tangential slope information 610 and radial slope information 611 can be used to determine curvature 612 along the gradient, i.e. the direction of maximum surface slope.
• Curvature 612 can be used to compute a generalized ZDD value 614 that does not necessarily correspond to a radial direction.
• Computing a ZDD value is described, for example, in SEMI M68-1109 - Practice for Determining Wafer Near-Edge Geometry from a Measured Height Data Array Using a Curvature Metric, ZDD.
• one or more calibration schemes 613 can be used to further increase the accuracy of the ZDD value 614 , in particular the influence of chucking forces on the wafer, and dynamic forces during a scan.
• accurate ZDD measurements obtained from the industry standard WaferSightTM tool, which is provided by KLA-Tencor
• partially-overlapping, round beams formed by the enhanced BF-DIC technique can be focused onto the wafer and scanned in a radial direction while the wafer is spinning to measure the radial ERO of the wafer with fine spatial resolution.
• this technique can be used for additional areas of the wafer as time permits. Indeed, these radial slope measurements in combination with the standard tangential slope measurements generated with partially- overlapping, elliptical beams can be used to construct an
• A 1 - sin[ 5 ⁇ p(t)] cos[ ⁇ ( ] - cos[ 5 ⁇ p(t)]sin[ ⁇ ( ]
• amplitudes on the differential signal is negligible (limited to a cosine effect) .
• vibration can enter the inspection system directly as 2 sin[ ⁇ ( ⁇ )] .
• typical frequency ranges encountered for these vibrations is low compared to the desired differential signal .
• these frequency ranges result in spurious signals and can be filtered out electronically.
• the inspection system is more susceptible to wobble compared to when the scanning spots are tangentially displaced.
• radial wafer wobble is typically more severe than tangential wobble given its axisymmetry, and radial wafer wobble has no impact on the lateral positions on the
• Wollaston prism for the tangential displaced beams .
• the signal bandwidth in the radial direction is determined by the spin rate of the wafer, which is far lower than the video rate in the tangential direction. Therefore, preferred BF-DIC inspection system embodiments that measure both tangential and radial slope must use standard techniques to reduce the amplitude and frequency of wobble and other types of vibration.
• Figure 7A illustrates an exemplary processing of information in which both tangential and radial slope information can be used to generate accurate height information of the wafer surface.
• First BF-DIC measurements 701 can be obtained from using the two tangentially-displaced elliptical beams that generate two partially-overlapping, tangentially-displaced scanning spots (e.g. scanning spots 301 and 302, Figure 3A) .
• Second BF-DIC measurements 703 can be obtained from using the two radially-displaced round beams that generate two
• partially-overlapping, radially-displaced scanning spots e.g. scanning spots 314 and 315, Figure 3B. Measurements from these partially-overlapping, radially-displaced scanning spots can generate radial slope results 704.
• the radially-displaced beams can be obtained by rotating the Wollaston prism by 90 degrees relative to the position of the prism when generating the tangentially- displaced beams.
• Tangential slope results 702 and radial slope results 704 can be used to generate integrated height results 705.
• the tangential and radial slope information can be integrated to generate an accurate height for any location on the wafer.
• Such height information obtained of the wafer may include shape distortion induced by virtue of the force of gravity and other clamping forces acting on the wafer, while being held horizontally on a mounting or a clamping (chucking) system (such as a 3-point kinematic mount or an edge-handling chuck) .
• These results 705 can subsequently be used to generate the wafer shape 707.
• the results 705 can be adjusted based on chucking distortion information 706 to obtain wafer shape information with improved accuracy 707.
• chucking distortion information 706 can be subtracted from wafer shape 707 to generate a corrected wafer shape 708.
• chucking distortion information 706, provided as tangential and radial information of a known shape (e.g. bowl- shaped or flat) imparted to a wafer secured by a chuck, can be subtracted from enhanced BF-DIC wafer images .
• Figures 7B and 7C illustrate exemplary radial and tangential slope wafer images, respectively. In Figures 7B and 7C, dark (light) shading
• Figure 7D illustrates an exemplary wafer shape image based on the radial and tangential slope wafer images of Figures 7B and 7C.
• dark (light) shading indicates negative (positive) height. Note that such images are typically generated in various gradations of color, and therefore may show finer gradations of slope and height .
• chucking distortion information 706 may be derived from a finite element (FE) model (which is known by those skilled in the art of semiconductor manufacturing) depending on chuck configuration.
• FE finite element
• the chuck itself may directly contribute to surface distortions of the wafer, particularly in the areas spanning the grooves when the vacuum is applied.
• Other types of chucks may contribute different amounts of distortion to various areas of the wafer when secured by the chuck.
• Some state-of-the-art chucks may contribute only negligible distortion to the wafer and therefore chucking distortion information 706 can be ignored.
• BF-DIC provides sufficiently high lateral resolution to derive the topography of the wafer substrate.
• Wafer topography may be obtained from either the integrated height data or the corrected wafer shape data by applying a high-pass filter to remove low-frequency features (high-pass filter may include filters such the Laplace filter etc . ) .
• Figure 7 ⁇ illustrates an exemplary image 711 showing reference topography measurements for a portion of a patterned wafer (e.g. obtained from the industry standard WafersightTM tool)
• Figure 7F illustrates a corresponding image 712 showing BF-DIC topography measurements with filtering.
• dark (light) shading indicates negative (positive) height
• Figure 7F dark (light) shading indicates positive (negative) height.
• Images 711 and 712 advantageously show substantially the same topography
• Wafer topography is useful in getting insight into the higher-frequency height variation of the front-side of the wafer which may lead to defocus related issues on a lithography scanner during IC manufacturing on such a wafer substrate.
• BF-DIC may be calibrated to measure the topography of a patterned wafer.
• a predetermined set of wafers can be measured using both a reference metrology tool, such as the WaferSightTM tool, and a tool configured to provide the enhanced BF-DIC measurements. Comparing these measurements (like comparing Figures 7 ⁇ and 7F) can provide calibration feedback for the tool configured to provide the enhanced BF-DIC measurements, thereby improving its accuracy.
• the shape information generated by the enhanced BF-DIC technique can be used to determine the residual stress of the wafer.
• the deposition of a film layer on a substrate can induce stresses, which cause the substrate to curve (bow) . Stresses that remain in a material without application of an external load are called residual stresses.
• residual stresses can be induced by the deposition of a film on the substrate.
• Figure 8A illustrates an exemplary technique for generating a film-stress map for a wafer.
• a pre-deposition shape 801 of a wafer and a post-deposition shape 802 of the wafer can be used to generate a shape difference 803.
• both pre-deposition shape 801 and post-deposition shape 802 can be generated using the enhanced BF-DIC technique
• Stoney's equation 804 can be used to generate an accurate film- stress map 805.
• ⁇ is the elastic modulus of the substrate
• v s is Poisson's ratio of the substrate
• Roc is the effective radius of curvature
• t s is the substrate thickness
• t f is the film thickness
• FIG. 8B illustrates an exemplary strained substrate 810 with a thin film 811 thereon having tensile stress.
• the effective radius of curvature can be defined as: where Ri is the initial radius of curvature and R 2 is the radius of curvature after film deposition.
• Ri and R 2 refer to a global or an average value of radius of curvature per wafer.
• local stress variation shown variation in stress across the wafer
• this modeling can take into account local boundary conditions.
• stress induced by film depositions may be accurately measured using BF-DIC by determining the difference between the corrected shape
• laser diodes of specific wavelengths can be used as light sources.
• channels each of which may operate with a predetermined spot size and separation appropriate for surface slope measurement, defect detection, and/or review imaging.
• Figure 9 illustrates an exemplary
• enhanced BF-DIC system 930 including a plurality of separate DIC sub-systems.
• BF-DIC system 930 has two subsystems 931 and 932 , each sub-system having substantially similar optical components.
• sub-system 931 includes a light source 917 that produces a light beam 915.
• a Wollaston prism 909 receives light beam 915 after passing through a beamsplitter 911.
• Focusing optics 907 can be configured to focus the two light beams generated by Wollaston prism 909 onto a substrate 900 as first scanning spots.
• Substrate 900 can be secured by a chuck 901, which is movable using a spindle motor 902 (providing rotation ⁇ ) and a linear motor 903 (providing x-y movement, e.g. moving in a radial direction r) .
• Both spindle motor 902 and linear motor 903 can be controlled by a central control and data acquisition computer 906 via a motor control cable 904 (shown connected to linear motor 903 for illustration simplicity) .
• Central control and data acquisition computer 906 receives the processed data from data acquisition circuit 919 via a data acquisition cable 905.
• Sub-system 932 includes similar optical components to those of sub-system 931. Specifically, sub-system 932 includes a light source 918 that produces a light beam 916. Note that in one embodiment, light source 918 can be the same as light source 917 , wherein the output light beam can be directed to sub-systems 931 and 932 using standard optical components . In this
• the light source can be characterized as being external to one or any sub-system.
• a Wollaston prism 910 receives light beam 916 after passing through a beamsplitter 912.
• Focusing optics 908 can be configured to focus the two light beams generated by Wollaston prism 910 onto substrate 900 as second scanning spots. Light reflecting from substrate 900 from the second scanning spots is redirected through Wollaston prism 910, which recombines the light.
• Mixing optics 934 direct the recombined light from Wollaston prism 910 (via beamsplitter 912) onto photo-detectors 914.
• a data acquisition circuit 920 A data acquisition circuit 920
• Central control and data acquisition computer 906 receives the processed data from data acquisition circuit 920 via data
• the optical components of subsystems 931 and 932 can produce light beams with dimensions, shapes, orientations, and displacements suited to their intended uses.
• Wollaston prism 909 can have a first orientation that generates two beams that are tangentially-displaced, whereas Wollaston prism 910 can have a second orientation that generates two beams that are radially- displaced.
• focusing optics 907 could be configured to generate elliptical beams
• focusing optics 908 could be configured to generate round beams.
• each sub-system could interrogate a different region of the wafer, e.g. sub-system 931 could focus at a center of
• substrate 900 and sub-system 932 could focus at the wafer edge. In one embodiment, the positions of such focusing can be changed during the r- ⁇ scan.
• the wavelengths for the light sources of a set of DIC sub-systems are different, although in other embodiments the wavelengths could be the same. In either case, isolation of the sub-systems to prevent optical or electronic cross-talk is an important consideration for the implementation of a particular embodiment, using common methods and techniques well-known to those skilled in the art.
• a plurality of miniaturized BF- DIC systems may operate at distinct wavelengths .
• the wavelength can be extended to the near-IR, where the optical skin depths of important substrate materials, e.g. such as silicon, increase to 10-100' s of microns to nearly complete transparency depending on dopant type and concentration.
• substrate materials e.g. such as silicon
• the results of the enhanced BF-DIC techniques described herein can characterize and/or monitor a wafer process.
• the results can be used as feedback or fed forward to another wafer process .
• the results of the scanning can be used to characterize and/or monitor an integrated circuit manufacturing processes such as chemical-mechanical polishing (CMP) , Rapid Thermal Processing (RTP) , Chemical Vapor Deposition (CVD) , etc.
• CMP chemical-mechanical polishing
• RTP Rapid Thermal Processing
• CVD Chemical Vapor Deposition
• the polarized light can be split using a Wollaston prism
• Nomarski prism which also consists of two wedges of a
• a first wedge is configured as the above-described Wollaston prism and a second wedge has its optical axis obliquely positioned, thereby providing an
• the Nomarski prism can be located outside the aperture plane of the objective lens, thereby providing further flexibility of component positioning.
• enhanced BF-DIC techniques can be used for any substrate.
• an x-y scan can be performed on the substrate.
• first and second beams are created from a first light beam.
• the first and second beams have round cross-sections, and form first partially overlapping scanning spots displaced in a first direction.
• Third and fourth beams are created from either the first light beam or a second light beam.
• the third and fourth beams have elliptical cross-sections, and form second partially overlapping scanning spots displaced in a second direction, wherein the first direction and the second direction are orthogonal.
• At least one portion of the substrate is scanned using the first and second partially overlapping scanning spots as the substrate is moved.

Landscapes

• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Testing Or Measuring Of Semiconductors Or The Like (AREA)
• Length Measuring Devices By Optical Means (AREA)
• Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
• Constituent Portions Of Griding Lathes, Driving, Sensing And Control (AREA)
• Finish Polishing, Edge Sharpening, And Grinding By Specific Grinding Devices (AREA)
• Mechanical Treatment Of Semiconductor (AREA)

Priority Applications (1)

Applications Claiming Priority (2)

Publications (1)

Family

ID=51525937

Family Applications (1)

Country Status (4)

Cited By (3)

Families Citing this family (8)

Citations (5)

Family Cites Families (3)

• 2013
  • 2013-03-12 US US13/797,901 patent/US9052190B2/en active Active
• 2014
  • 2014-03-11 JP JP2016501356A patent/JP6171080B2/ja active Active
  • 2014-03-11 WO PCT/US2014/023824 patent/WO2014164935A1/en not_active Ceased
  • 2014-03-12 TW TW103108800A patent/TWI623725B/zh active

Patent Citations (5)

Also Published As

Similar Documents

Legal Events