TITLE: DETECTING DEFECTS ON A WAFER

CROSS-REFERENCE TO RELATED AmjCATKxss

This application is a continuation-in-part of U.S. Patent Application Serial No. 12/339,476 entitled "S st ms nd Methods for Detecting Defects o a Wafe / ⁵ filed January 26„ 2009, which is incorporated by reference as if fully set forth herein.

BACKGROUND OF THE I V ION

Field of t he invention

The present invention generally relates to detecting defects on a wafer.

2- Description of the Related Art

The following description and examples are not admitted to be prior art by virtue of their inclusion in this section.

Inspection processes are used at various steps during a semiconductor manufacturing process to detect delects on wafers to promote higher yield, in the manufacturing process and thus higher profits. Inspection has always been an important pari o fabricating semiconductor devices such as !€s. However, as the dimensions of semiconductor devices decrease, inspection becomes even .more im or ant to the successful maimfaciure of acceptable semiconductor devices because smaller defects can cause the devices to fail. ^' For instance, as the dimensions of semiconductor devices decrease, detection of defects of decreasing size has become necessary since even relatively small defects may cause unwanted aberrations in. the semiconductor devices.

One obvious way to improve the detection of relatively small defects is to increase the resolution of an optical inspection system. One way to increase the resolution of an optical inspection system s to decrease the wavelength at which dse system can. operate. As the wavelength of inspection systems decreases, incoherent light sources are incapable of producing light with sufficient brightness. Accordingly, for inspection systems that are designed, to operate at smaller wavelengths, a more suitable light source is a laser light source that cars generate relatively bright light at relatively small wavelengths. However, laser light sources generate coherent light. Such light is disadvantageous tor inspection since coherent light can produce speckle in. images of a wafer. Since speckle Is a source of noise m the linages, the signahto-uoise ratio (S N) in images generated by inspection systems will be reduced by speckle. In addition, speckle noise in wafer inspection systems (e.g., laser-based inspection sy stems) is one of the main limitations of defect of interest (DOT) detection ability. As wafer design rules continue to shrink, optical inspection system preferably have shorter wavelengths and larger col lection numerical apertures (NAs , Speckle noise consequentl increases to a. more dominant noise source.

Many illumination systems nave been developed tor inspection applications thai reduce the speckle of light from laser light sources. For example, popular approaches to reduce speckle noise currently involve reducing coherence of the illumination laser source by transmitting light through an. optical diffuser or vibrating optical fiber. These approaches usually require increasing the illumination NA on the wafer and therefore are not effective for an outside-4he lens (OIL) oblique angle illumination architecture. Reduction of laser coherence also limits the usage of Fourier filtering and degrades the S/N. Other approaches such as moving an aperture in the pupil plane have been applied to select a spatial sample of light in the pupil plane and then average the linage over a relati ely large number of samples. This approach will greatly reduce the resolution of the optical system thereby decreasing the defect capture rate.

Some -methods for defect detection ut lize output generated by multiple detectors of an inspection system to detect detects on a wafer and or to classify defects detected, on the wafer. Examples of such systems and. methods are illustrated in international Publicatio No. WO 99/67626 by Ravid et ah, which is incorporated by reference as if full set forth herein. The systems and methods described in t s publication are generally configured to separately detect defects in the electrical signals produced by different detectors. In other words, the electrical signals produced by each of the detectors are processed separately to determine if each detector has detected a defect. At any time that a defect is detected in the electrical signals produced by one of the detectors, die electrical signals produced by at least two of the detectors arc anslyxed. collectively to determine scattered tight attributes of the defect such, as reflected light intensity, reflected light volume, reflected light, linearity, and reflected light asymmetry. The detect is then classified (e.g., as a pattern detect or a particle defect) based on these attributes.

Although the methods and systems disclosed in the above-referenced publication utilize scattered light attributes of delects determined from electrical signals generated by more than one detector, the methods and systems disclosed in this publication do not utilize electrical signals generated b more than one detector in combination to detect the defects. In addition, the methods and systems disclosed in this publication do not use a combination of electrical signals generated by more than one detector for any defect- related function other than classification.

Other currently available inspection systems are configured to inspect a wafer wit more than one detection channel, to detect defects on the wafer by separately processing the data acquired by each of the channels, and to classify the defects by separately processing the data acquired by each of the channels. The detects detected by each of the individual channels may also be further processed separately, for exam le, by generating different wafer maps, each illustrating the defects detected by only one of the indi idual channels. Use defect detection results generated by more than one channel of such a system may then be combined using, for example, Venn, addition of the individual, wafer maps. Such inspection may also be performed using output acquired in a single pass or multiple passes. For example, one previously used method for detect detection includes performing two or more scans of a wafer and determining the union of the lot results as the final inspection result for the wafer, in such previousl used .methods, nuisance filtering and defect binning is based on the Veers ID results, AND/OR.

operation, from multiple scans.

Such previously used inspection methods, therefore; do not leverage the output generated by the inspection system at. the pixel level, but rather combine the results ai the wafer map level as ihe final result. Detects are detected by each pass independently based on their relative signal (magnitude) compared to the wafer level noise seen lor each ass. In addition, nuisance filtering and binning in previously used methods- may he based on the AND/OR detection from multiple scans and thereafter separation, in each individual scan. As such, no cross-pass information other than the A D/OR. operation on detection is considered.

Accordingly, it would be advantageous to develop methods and systems for detecting defects on a. wafer thai combine information from differ nt optical states of an inspection system to increase the S/N of defect in image data for the wafer us tor defect detection while decreasing noise (e.g., speckle noise) in the image data. lM;;.l!l :j V >;ii ¾

The following description of various embodiments is not to be construed in any way as limiting die subject matter of the appended claims.

One embodiment relates to a method for detecting defects on. a wafer. The method includes generating output for a wafer by scanning the wafer using first and second different values for focus of an inspectio system, and the same values for all other optical parameters of the inspection system. The method also includes generating first image data for the wafer using the output generated using the first valise for focus and second image data for the wafer using the output generated using the second value for focus, in addition, the method includes combining the first image data and the second image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer. The method farther includ s detecting defects on the wafer using the additional Image data.

Each of the steps of the method ma be further performed as described herei . In addition, the method may include any other siep(s) of any other methodCs) described herein. Furthermore, the method may be performed by any of the systems described herein.

Another embodiment relates to a non-transitory computer-readable medium containing program instructions stored therein for causing a computer system to perform, a computer-implemented m thod for detecting defects on a wafer, The method includes acquiring output .for a wafer generated by scanning the wafer using first and second different values for focus of an inspection system and the same values for all other optical parameters of the inspection system.. The method also includes generating first image data for the wafer using the output generated using the first value for focus and second image data for the wafer using the output generated using tbe second value for focus. In addition, the method includes combining the first image data and the second image data corresponding to substantially the same locations on t re wafer thereby creating additional image data tor the wafer. The method further includes detecting defects on the wafer using the additional image data.

Each of the steps of the computer-implemented, method described, above may be further performed as described herein. I s. addition,, the computer-implemented method may include any other siep(s) of any other methodfs) described herein. The computer- readable medium may be farther configured as described herein.

An additional embodiment relates to a. system configured to detect defects on a wafer. The system includes an inspection subsystem configured to generate output for a wafer by scanning the wafer using first and second different values for feces of the inspection subsystem and the same values for all other optical parameters of ^' the inspection subsystem. The system also includes a computer subsystem configured to generate first image data for the wafer using the output generated using the first value for focus and second image data for the wafer using the output generated -using the second value for focus, combine the first image data, and the second image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer, and det ct defects on the wafer using the additional image data, The system may be further configured as described herein,

Further advantages of the present invention .may become apparent to those skilled i the art with the benefit of the following detailed description of the preferred embodiments and upon reference to the accompanying drawings in which:

Fig, i is a. schematic diagram, illustrating a side view of one embodiment of different optical states of an inspection system thai are defined by different, values for at Least one optical parameter of the inspection system;

Fig, 2 includes different image data .for the same location on a wafer, each of which was generated using output that wa s generated using one of the di fferent optical states of Fig, .1 ;

Fig. 3 includes different output generated for the same location on a wafer using the different optical states of Fig. 1;

Fig. 4 is image data generated using one example of the output of Fig. 3;

Fig, S is additional image data created by combining the image data of Fig, 4 with other image data generated using the other example of the output of Fig. 3;

Fig, 6 is a. block diagram illustrating one embodiment of a non-transitory computer-readable medium containing program instructions stored therein for causing a computer system io perform a con^uter-implenmtted method for detecting defects o a wafer; and

Figs. 7-8 are schematic diagrams illustrating side views of embodiments of a system configured to detect, defects on a wafer.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the ra ings and may herein be described in detail. The drawings .may not be to scale. It should be understood, however, that the drawings and detailed description thereto are not intended to limit the invention to the particular form disclosed, but on the contrary, the intention is to cover all modifications, equivalents and alternatives failing within the spirit and scope of the present invention as defined by the appended claims. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

Turning ROW to the drawings, it is noted mat the figures are not drawn to scale. In particular, the scale of some of the elements of the figures is greatly exaggerated to emphasize characteristics of the elements. It is also noted that the figures are not drawn to the same scale. Elements shown in more than, one figure that may be similarly configured have been indicated using the same reference numerals.

One embodiment relates to a method for detecting defects on a wafer. The method includes generating output for a wafer by scanning the wafer with an inspection system using first and second optical states of the inspection, system... The output generated by scanning the wafer may include any suitable outp t and may a

depending on the configuration of the inspection system and/or the inspection recipe used to perform the scanning. For example, the output may include signals, data, images, or image data responsive to light scattered from the wafer (e.g., in the ease of dark .field (DF) inspection systems). The inspection, system maybe a commercially available inspection system such as the Puma 1.xx series tools, which are commerc.ial.fy available from I ,A- ^' tencor, Mi!piias, California. The inspection system may be configured for inspection of patterned wafers and/or anpatterned wafers. In addition, the inspection system may be configured for DF inspection, possibly in combination with one or more other inspection modes (e.g., an aperture mode of inspection). Furthermor , die inspection system may be configured as an optical inspection system.. Scanning the wafer with the inspection system may be performed in any suitable manner, For example, the wafer may be moved (by a stage of the inspection system) with respect to optics of the inspection system such that the illumination of the inspection system, traces a serpentine path over the wafer as light, scattered from the wafer is detected.

The first and second optical states are defined by different values for at least one optical parameter of the inspection system. For example, an optical "stat " (which may also be commonly referred to as an optical "configuration" or "mode ) of the inspectio system can be defined by values for different optical parameters of the inspection, system that are or can be used in combination to generate output for a wafer. The different optical parameters may include, for example, wavelength of illumination, wavelength of collection/detection, polarization of illumination, polarisation of collection detection, illumination angle (defined by elevation angle or angle of incidence and possibly azimuthal angle), angle of collectioti deteetion, pixel size, and the like. The first and second optical states may be defined by different values for only one of the optical parameters of the inspection system and the same values for other optical, parameters of the inspectio system. However, the first and second optical states may be defined by different values for two or more of the optical parameters of the inspection, system.. n one embodiment, the different values include different angles of illumination at which light is directed to the water during the scanning, ^' line different angles of illumination, may include substantially the same elevation angle and different aziniuthal angles. Fig. 1 illustrates one such embodiment of different optical states of an inspection system that are defined by different values for at least one optical parameter of the inspection system. For ex m le, as shown in Fig. L light .10 ma he directed to wafer 12 at azimuth al angle 14 (e.g., an azirnuihal. angle of about 45 degrees). Light 16 maybe directed to the wafer at azimutha! angle 18 (e.g., an azi ruthal angle of about -45 degrees). Light 1.0 and light 36 may be directed to the wafer at the same or substantially the same elevation angle 20 (e.g.. about 15 degrees).. However, light 10 and light 16 may be directed to the wafer at diljerent elevation angles and/or different azhmuhal angles. Light .10 and light 16 may be generated by different light sources or the same light source.

Light 10 ami Hght .1 may have substantially the same characteristics (e.g., wavelength, polarization, etc.). i this manner, in order to separately detect the light scattered from the wafer due to illumination at the different illumination angles to thereby generate separate output for the different optical states, the wafer may be scanned with light at the different illumination angles in different passes (i.e., multiple passes performed in a single process). For example, in a double pass inspection., first pass output may be generated with illumination coming in at a certain elevation angle and a 45 degree azimutha! angle. Second pass output may be generated with, the same optical conditions used for the first pass except at -45 degree azimuths! angle illumination.

Such multiple pass (or multi-pass) output generation may be performed for any different optica! states of the inspection system for which output cannot be

simultaneously and separately generated (e.g., due t a difference in a setting of a single optical element of the inspection system between, the different optica! states). However, if the different optical states of the inspection system can be used to simultaneously generate separate output for the wafer (e.g., using different channels of lire inspection, system), then generating the output using the first and second optical states of the inspection system may be performed in a single pass scan of the wafer.

In. another eovbodimenh scanning the wafer with the inspection system u ing the first and second optical states is performed with coherent. Hght. The coherent light may include light generated by any suitable coherent light source (e.g., a laser) at any suitable wavelength, in addition, the method may be performed using cmtside h.e~le«s (OTL) optical inspection system in which the illumination source is laser light incident on the wafer at an oblique angle of incidence. n one such embodiment, as shown in Fig. 1, light 1.0 and Sight 16 may he directed to the wafer at an oblique angle of incidence and outside of lens 22 of the inspection system. Lens 22 may be configured to collect light scattered from the water due to illumination of the water during scanning. An inspection system thai includes lens 22 may be further configured as described herein.

.In this manner, one adv an tage of the embodiments described herein, is that the embodiments can reduce speckle noise as described further herein: and compared to other common approaches for reducing speckle noise , the coherence of the aser source may be preserved. Therefore, in the embodiments described, herein, a Fourier filtering technique can be applied effectively to eliminate pattern background in a DF geometry. The Fourier filtering technique may include any (optica! or data, processing) Fourier filtering technique known in the art. Although the output may be advantageously generated in. the embodiments described herein by scanning the wafer using coherent light in an OTL illumination configuration, the output may be generated using any suitable light in any suitable illumination configuration, in an additional embodiment, the first and second optical states are defined by the same values lor optical parameters of the inspection system used, for collecting light if om the wafer during the scanning. For xam l , as described above, the valo.es may include different angles of illumination at which light is directed to the wafer during scanning. In addition, the different optical states .may be different only m one or more Illumination optical parameters of the inspection system. As such, no change in. collection optical path between the different optical states (using which output may be generated in possibly different passes or scans of the wafer) may be made since only the illumination may be changed. Using the same collection optical path for the different optical states may advantageously reduce the alignment error and optical error between different image data, which may be generated as described herein using output acquired by scanning the wafer (e.g., in two or more passes) and which ma be combined as described further herein.

In a further embodiment, the different values include different imaging modes, different polarization states, different wavelengths, different pixel sizes, or some combination thereof. For example, the different values may include different polarization states for illumination. In one such example, the first and second optical states may be defined by the same polarization states for collection. For example, the different values .may include the p-poiarked (P) state for illumination in one optical state and the s~ polarized (S) state for illumination in the other optical state, and the polarization state used for collection in both optical states may be impolarized. (N). However, is another such example, the optical states may also be defined by different polarization states for collection. For example, the first optical state may be defined by the S polarization state for illumination and the P polarization state for collection, and the second optica! state may be defined by the P polarization state for ilhnninatkm and the S polarization state for collection. in another embodiment, the different values include different channels of the inspection system. For example, the first optical state may be defined, by a first channel of the inspection system, and the second optical state may be defined by a second channel of the inspection system. In other words, the o tp t for the wafer may be generated for the first optical state using one channel of the inspection system, and the output, for the wafer may be generated for the second opti cal sta te using a. different channel, of the inspection system. The term "channel" is generally used herein to refer to different, detection subsystems or detectors of the inspection system, which may be different in angles (I.e., collection angles) at which light from the wafer is collected and detected by the detection subsystems or detectors, but which may or may not be different in other respects as well {e.g., wavelengthis) at which light is detected by the channels, polarization of the light detected by the channels, elm). hi one such example, if the inspection system includes three channels, the first and second optical states may be defined by the following channel combinations; channel !. and channel 2; channel 2 and channel 3; and channel 1 and channel 3. In addition, as described further herein, the embodiments may he performed using more than two different optical states. In one such example, if the inspection system includes three channels, first, second, and third optical slates may be defined by channel 1, channel 2, and channel. 3„ respectively. Furthermore, each of the different, optical stales may he defined by a different channel of the inspection system (e.g., N optical states defined, by N channels), in one such embodiment; generating the output using the first and second optical states is performed in parallel For example, the output generated using the first ami second optical states may be generated in the same pass or scan. As such, the output from each channel may be collected in parallel.

The method also includes generating first image data for the wafer usi g the output generated using the first optical stale and second image data tor the wafer using the output generated using the seco d optical slate, in one embodiment, the first and second image data includes difference image data. The difference image data ma he generated in any suitable manner. For example, difference image data for the first optical, state may be generated u ing test image data and two references (e.g., image data from dies on the wafer that are adjacent to the die from which the test image data was generated), in such an example, one reference may be subtracted from the test image data, rid the other reference may be separately subtracted from the test image data. The results of both subtraction operations may be multiplied, ami the absolute value of that product ma be the difference image data. Difference image data lor the second optical state may be generated in a similar manner. As such, the difference image data may be separately generated for each optical sta e using only output generated using that optical state. In other words, generating the difference image data is not a cross-optical state operation. In this manner, generating the first and. second image data may include performing a die-to-die subtraction to eliminate pattern background in the outpu However, the d fference image data may be generated in any other suitable manner using an suitable aigorithrn(s) and/or nrethodis). in addition, the first and second Image data- may not be difference image data. For example, the first and second image data may be raw image data of the wafer after any other pattern background elimination operationis) (e,g., after Fourier filtering).

The method also includes combining the .first image data and tire second image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer. In this manner, the first and second image data, ma be combined on a .!ocation-to-ioealion basis. Unlike other methods thai in vol ve combining information about a water acquired using different opti cal states of an inspection system, combining the first image data and die second image data as described herein creates different image data for the water, which can then be used as described further herein (e.g., for defect detection). For example, combining die first and second image data may include performing "ima e fusion" using the first image data, and the second image data, in other words, new image dat of the wafer may be "fused" from two other image data of the water. As such, image fusion may he performed using multiple optical states, which may Include any of those described herein (e.g., optica! states defined by different polarizations, different channels, etc.). For example, image fusion can be achieved by using image data, from any two (or more) channels of the inspection system. In. one such example, if the inspection system includes three channels, image fusion can be performed using the following channel combinations: channel i and channel 2; channel 2 and channel. 3; channel .1 and. channel 3; and channels .1 , 2, and 3. In addition, as described further herein., the output used to generate the first image data and the second image data may be acquired in different passes. In. this manner, the method may include multi-pass image fusion. However, as also described farther herein, the output used to generate the first and second image data may be acquired in single pass (e.g.. output from each channel ma e collected in parallel). As such, the method may include single-pass image fusion. Although the meth d includes combining t e first image data arid the second image data as described above, the method is not limited to combining only the first image data and the second image data. For example, if output is generated for the wafer using a third optical state of the inspection system, hieh is defined by at least one val ue for at least one optica! parameter of the inspection system that is different than the values for that at least one optical parameter which define the first and second optical states, the method may include generating third image data for the water using the output, generated, using the third optica! state, which may be performed as described herein. The third optical state may be defined by any of the different values for any of the optical parameters described herein. In one such example, each of the different optical states may he defined by a different channel of the inspection system. The method may also include combining the first image data, the second image data, and the third image data corresponding to substantially the same locations on the wafer as described, herein thereby creating the additional image data for the wafer. Combining image data generated using output generated by two or more optical states is advantageous as described further herein.

Furthermore; although the method includes creating additional image data for the wafer, the method is not limited to creating only the additional, image data for the wafer. For example, the method may include creating the additional image data for the wafer as described, above and creating different additional image data for the wafer in a similar manner. I one such example, the additional image data may be created by combining the first and second image data as described, above. The method may also include generating output for the wafer using a third optical state of the inspection system, which may be defined as described farther herein. That output, may be used io generate third image data for the wafer as described further herein. Thai third image data may then be combined as described herein with the first image data and/or the second image data to create different additional image data. The different additional image dat may be used in a manner similar to using the additional image data in steps described farther herein. In one embodiment, combining the first image data and the second image data includes performing image correlation on the first image data and the second image data corresponding to substantially the same locations on the wafer, For ex m le, new wafer image data or the fused im ge data ma be generated by correlating image data (e.g., from two passes). In one example, the image correlation may include a 5 pixel by 5 pixel correlation. However, the image correlation may be nerformed in any other suitable manner using any suitable image correlation, algorithm(s) and/or mefhod(s). In addition., the i ge correlation may be performed using any suitable image processing technique that can be used for image correlation, in another embodiment, combining the first image data and the second image data is per formed at the pixel level of the first and second image data. In other words, the first and second image data may be combined on a pixel-hy-pixel basis, in still other words, eoni.bin.ing the first and second image data may be performed separately for individual pixels in the first and second image data. By fusing information at tire pixel Level, one can leverage both magnitude (intensity) and phase (correlation) information among different optical states (which may be generated by different inspection, passes). By combining information at the pixel level, new dimension to exploit becomes available, namely the coincidence among different perspectives (optical states).

The first and second image data (e.g., difference image data) for the different optical states may be generated for the entire wafer or the entire portion of the wafer thai is scanned with the inspection, system, using the first and second optical, states. In addition, combining the first and second image data may be performed using all of the first and second image data. In this manner, image fusion may he performed for the entire wafer or the entire portion of the wafer that is scanned using the first and second optical states.

However, image fusion may not be performed for the entire wafer or the entire scanned portion of the wafer. For example, the method may include applying an intensity cut off to the first and/or second im age data and eliminati ng any of the first and/or second image data thai doss not have intensity values that exceed the intensity cut oil ^' in this manner, the first and/or second image data that is not eliminated may be identified as candidates to be used for additional steps performed in the method, In one such example, if the first image data is generated using output generated in a first pass of the wafer and the second image data is generated using output generated in a second pass of the wafer, the intensity cut off may be applied, to the first image data, to ehmin ic any of the first image data that does not have intensity values thai exceed tire intensity cut off. in this manner, the method may include saving image patch data for only the candidates identified in the first pass. In the second pass, only second image data that corresponds to the same locations on the wafer as the candidates .may be stored and/or combined with the first image data., in this maimer, the image data that is saved in the second pass may vary depending on the candidates that were captured in the first pass, and image fusion may then be performed using the image d a saved in the second pass. However, if the first and second image data, is generated using output that is generated n a single pass of the wafer, the intensity cut off may be applied to both the first and second image data, and an of the first and second image data t at has values that, exceed the intensity cut off may he combined with the corresponding image data regardless of the intensity values of that other image data.

The method may include pertorming some (i.e., one or more) dilation steps to ensure proper alignment between the detect signals in the image data,. For example, for each of the candidates identified as described above, a 3 pixel by 3 pixel dilation may be performed. However,, the dilation step(s) may include any suitable dilation, image processing teehniqo.e(s} known in the art and may be performed using any suitable method(s) and/or algoritlim(s). The dilation siep(s) may be performed on both the first and second image data thereby increasing the accuracy with which delect signals in the first and second image data are aligned to each other.

Regardless of whether or not the method includes dilation siep(s) such s those described above, the method may include aligning the first and second image data prior to the combining step, image data alignment may be performed in any suitable manner. For example, image data aiignmeM may be erformed throu h cross-correlation of X and Y projections (e.g., of the average intensity ac oss the image data in the x and y directions) between image data generated for different optical states (e.g., from image data acquired in two passes ;.

In an additional embodiment, defect detection is not performed prior to the combining step. For example, as described above, an intensity cwt off may be applied to the first and/or second image data prior to the combining step. However _* the intensity cut off is not a defect detection threshold, method, or algorithm. Instead, the intensity cut off acts essential ly as a noise -filter to eliminate the image data that does no have relatively high intensity val ues only for the purpose of decreasing t he processing i n volved in other steps of the method. In addition, defect detection may be performed as described further herein using the first and second image data individually, and delect detection using the first and/or second image data may or may not be performed prior to performing the combining step. However, defect detection cannot e performed using the additional image data until after the combining ste in which the addi iona image data is created has been performed. In this manner, unlike methods and systems that involve combining information generated slier defect detection (e.g.. combining defect detection results from different scans of a wafer), the embodiments described herein combine information prior to defect detection, which is advantageous as described further herein.

Since the embodiments described herein include combining first and second image data to create additional image data, a fairly substantial amount of image data, may be stored during the method. Examples of methods and systems that are particularly suitable for storing relatively large amounts of data such as image data are described in. commonly owned U.S. Patent Application Serial No. 12/234,201 by Bhaskar et ah filed September 19. 2008, which published as U.S. Patent Application Publication No.

2009/0080759 on. March 26, 2009, and which is incorporated by reference as if fully set forth herein. The embodiments described, herein may include storing the output and/or image data generated by the embodiments described herein using the methods and systems described in this patent application. For example, a system may include eight image computers, During the first pass of a multi-pass s ec io performed with a first optical state, each image computer may receive and store 1/8 of the image data for each swath scanned on a wafer. During the second pass of the multi-pass inspection performed with a second optical state, each image computer may rece ve and store I./8 of the image data for each swath scanned on the wafer, In addition, each image computer may receive and store image data generated ai substantially the same locations on the wafer in both passes (i.e., image data generated at substantially the same wafer locations and/or substantially the same in-s at positions). The image data generated during the second pass may be stored in the image buffers of the image computers at fixed offset locations from the locations of the stored first pass image data. The stored image data may then be used in any of the step(s) described herein, ^" t he computer systems and computer subsystems described herein may be further configured as described in the above-referenced patent application. The embodiments described herein may also include any step(s) of any method(s) described in the above-referenced patent appl.i cation.

The method further includes detecting defects on. the wafer using the additional image data. Therefore, defect detection is no longer only determined by each optical state (or each pass) independently, but based on information fused from multiple optical states (e.g., all passes). In this manner, the methods described herein use image fusion results, which are generated by combining information from raw (difference) image data generated by multiple optical states, as the input to defect detection. The defects detected on the wafer using the additional image data may include any defects know in the art and may vary depending on one or more characteristics of the wafer (e.g., the water type or the process performed on the water prior to inspection).

Detecting the defects using the additional image data may include applying one or more defect detection thresholds to the additional image data. For example, the additional image data may be compared to one or more defect detection thresholds. The one or more defect detection thresholds can be used to make a decision regarding whether a pixel in the additional image data is defective or not.. Other methods tor defect detection using one or more defect detection thresholds may first select a set of candidate pixels using a simpler (.less computationally involved) test followed by a more complex computation applied only to the candidates to detect defects.

One or more defect detection thresholds that are used io detect the defects on the wafer may be defect detection thresboidis) of one or more defect detection algorithms, which may be included in an. inspection .recipe. The one or more defect detection algorithms that are applied to the additional, image data may include any suitable defect detection, algorithm(s) and may vary depending on, for example, the type of inspection that is being performed on the wafer. Exanipl.es of suitable defect detection algorithms, which can be applied to the additional image data, include segmented auto-thresholding (SAT) or multiple die auto-thresholding {MDAT}„ which are used by commercially available inspection systems such as those from LA-Tencor. In this manner, the additional image data may be treated as any other image data when ii comes to defect detection, in one embodiment, the additional image data has less noise than the first and second image data. For example, combining the image data for the wafer generated for different optical states as described herein offers new context of the noise profile of the wafer and sensitivity to defects of interest (DDI). In addition, by combining (or fusing) information from multiple optical states at the pixel level, the sensitivity to nuisance events or noise can be reduced. For example _* by performing the image correlation as described, above, wafer noise in. the first image data and the second image d a that is .uon-spatia!!y coincident can be substantially eliminated in the additional image data, in this manner, the embodiments described herein leverage the fact that different optical states (e.g., defined by different imaging modes, polarization states, wavelengths, pixel sizes, etc.) provide different perspectives of the water level, noise and nuisance defects thereby offering the potential to suppress noise in the additional image data, which may be used as described further herein (e.g., for defect detection purposes). hi an. additional embodiment, the additional image data has loss speck is rroise than the first ami second image data. For example, the embodiments described herein may use an image correlation process (to create the additional image data) thereby substantially eliminating un-eortelated speckle noise.. In addition, as described above, the first and second optical slates may be defined, by different illumination angles. As such, the embodiments described herein may be used, for speckle noise suppression by varying illumination angle. In other words, the embodiments described herein may be used for speckle noise suppression by correlating image data generated using output that is generated using various illumination angles in an optical inspection system. For example, as the illumination angle changes, the phase relationship of the scattered light from the surface roughness on the wafer changes, The speckle pattern in the image data changes accordingly. When this change is sufficient, a correlation of different image data will help to eliminate the speckle noise.

An example of how the speckle pattern changes as illumination, angle changes is shown in Fig. 2. inspection image data 24 and 26 was generated lor the same location on a wafer. Inspection image data 24 was acquired at an illumination azlmuthai angle of 45 degrees, inspection image data 26 w s acquired at an il (animation, admuthal angle of -45 degrees. Therefore, the inspection image data was acquired at. different illumination angles (e.g., the same elevation angle but different azimuihal angles). Portions 28 of the image data show that the speckle signatures produced by the two azirautha! angle illuminations are substantially different at the same location on a page break. In particular, die bright speckle spot in. inspection image data 26 is caused by surface roughness and wall, contribute as noise or nuisance, I inspection image data 24» the bright speckle spot is not present at the same location as the speckle pattern, changes with the illumination angle. Therefore, the wafer noise changes as the azi.muth.al angle of illumination changes. In particular, the bright speckle disappears as the a miathal angle is changed from -45 degrees to 45 degrees. in one embodiment, portions of the additional image data, that correspond to defects on. the wafer have greater signal ~to~noi so ratios (S/ s) than portions of the Itrst and second image data thai are combined to create the portions of the additional m ge data. For example, by combining (or fusing) information, from multiple optical states at the pixel level weak signal strengths from DOI may be enhanced.. Enhancing the signal strengths irom DO! may be achieved by not only taking advantage of the relative sig als for detects in each optical state (magnitude), but by also exploiting coincidence or correlations among the different optical states (phase). For example, fusing iniorrnauon ai the pixel level thereby leveraging both magn ude (intensity) and phase (correlation.} in.forrn. tkm among different optical states allows one to extract defects with weak signals and suppress noise and nuisance events by exploiting their respecti ve coincidence and non-coincidence for different optical stales, in this manner, the embodiments described herein leverage the fact that different optical states (e.g., defined by different imagi g modes, polarization states, wavelengths, pixel stees, etc.) provide different perspectives of the wafer level noise and nuisance defect* thereby offering the potential to enhance the contrast of the DOI and their separation I om nuisance defects:. In addition, pixel level image fusio across multiple optical states provides opportunities for enhancemen of separation between DO! and nuisances although both may have relatively high S/Ns.

In one such example, when the change in the speckle pattern, is sufficient between different optical states used to generate the output for the wafer, a correlation of di fferent image data for the different optical states will help to eliminate the speckle noise and improve the S/N as the signal scattering intensity from the defect ma he relatively constant, For example, if the illumination, angles are symmetric in optical architecture, defect signals may be similar in both, optical slates, which is especially true tor relatively small defects. While reducing speckle noise afte linage correlation, the defect signal is maintained during the process. In this manner, the embodiments described herein can. maintain a healthy defect signal level after linage correlation. For example, compared to other common approaches for reducing speckle noise, the speckle noise is selectively eliminated instead of averaging o ver a relatively large sample of speckle patterns.

Selectively eliminating the speckle noise instead, of averaging over a. relatively large sample of speckle patterns helps to reduce the noise floor and improve the S/N. m this manner, as the examples described herein illustrate, the detect S/N in the additional image data is greatly improved over the S/N of the defect in each individual optical state, especially on wafers where speckle noise is dominant For example, a defect that is not detectable using either optical state (using either difference image data generated using one illumination angle) may become detectable after image correlation. In particular, one advantage of the embodiments described herein is that speckle noise can be greatly reduced in the additional image data compared to the first and second image data while defect S/N¾ in die additional image- data are improved compared to the first and second image data. As such, a defect that is not detectable in either of the first and second image data may become detectable In the corresponding additional image data created by image correlation.

However, the embodiments described herein can be used to increase the S Ns for defects that are detectable in either one or both of the first and second image data individually (e.g., using image data from one illumination angle and/or using image data, from a different illumination angle). For example,, even if a delect produces a moderate S/N in one of the optical states defined by one illumination polarisation state and a feeble S/N in another of the optical states defined by a different illumination . polarization state, the defect S/N in the additional image data, can he increased relative to both optical states because fusing the information from the two optical states can both suppress noise and enhance signal, in addition, if a delect produces margins! S/Ns in two optical states defined by different illumination polarization states, the defect S/N in the additional image data can be increased relative to both optical, states _*

Furthermore, i f a defect produces appreciable S/Ns in. two optical states defined by different illumination polarizatio states and different collection polarization, states but noise (e.g., from the grain of the wafer) dominates the first, and second image data, the noise can be significantly reduced in the additional image data compared to the first and second image data by combining the first and second image data as described, herein. In a similar manner, if a defect has S Ns in two optical states, defined by different illumination polarisation states and different collectio polarization states, tha are on par with the ma mum S/ s of noise (e.g., from a grain signature) in the fimi and second image data, the noise can be significantly reduced .in the additional image data compared to the first and second image data by combining the first and second image data as described herein. in another example, different peak noise events may be present m first and second image daia generated ^' using first and second optica! ates defined by different channels of the inspection system, but a defect may have sufficient correlation in the first and second image daia such thai by combining the first and second image data as described herein, the S/N of the defect can be dramatically higher in the additional image data compared to the fi st and second image data, In this manner, the embodiments described herein may be used to enhance the deteetabiiity of DO! for wafer inspection systems ^• using information from multiple optical slates. An example of .how the S/N of a defect can be improved is show in Figs. 3-5. in particular, output 30 shown in Fig, 3 is a raw image of a bridge defect from, an azimuth&! angle of 45 degrees. The S/N (max DiiT) of the bridge defect was 1.265 in this image. The S/N was determined using the signal in signal window 32 and the noise in noise window 34. which includes noise from page breaks. Output 36 shown in Fig. 3 is a raw image of the bridge defect from an asimuthal angle of -45 degrees. The S/N (max Diff) of the bridge defect was 1 , 51 in th s image. The S/N was determined using the same signal window and noise window as described, above, in this manner, the raw images of the bridge defect from aximuihal angles of 45 degrees and -45 degrees show that neither image has a sufficient S/N such that the delect can be captured usi g either image, For example, the S/Ns for the defect are 1.265 and 1.051, which are both much less than the typical threshold value used for defect detection.

Fig. 4 is an example of image data for the wafer generated using one of the raw- images of Fig. 3. For example, image data 38 shown in Fig, 4 is image data generated by die-to-die subtraction and background suppression performed using one of the images shown in Fig. 3 and a corresponding reference image from an adjacent die on the wafer. As shown in Fig. 4, speckle noi e appears as man nuisances this image data, la this manner, speckle noise shows as many nuisances even after die-to-die subtraction and background suppression, in particular, with the background reduced, nuisance is still apparent in the image data, More specifically, signals 40 in image data 38 correspond to nuisance from speckle on a page break, while signal 42 corresponds to a defect

Therefore, die deteetahiiiiy of the defect will be reduced, by tbe nuisance thai remains after die~io~die subtraction and background suppression.

Fig. 5 is an example of additional image data created by combining the image data of Fig. 4 ith other image data generated for the wafer -using the other image of Fig.

3. in particular, image data 44 shown in Fig. 5 was created by performing image correlation using image data 38 and other image data, image data that was generated from output that was generated using a 45 degree azimuthal angle and image data that was generated from, output that was generated using a -45 degree aximuthsl angle. After correlation performed using the 4S degree a imuthal angle difference image data and the -45 degree az.imu.thai angle difference image data, the S/N of the defect in the image data created by image correlation is 2. in this manner, the S/N of the defect Increases from 1 ,265 and 1.051. to 2. As such, the defect is now detectable with noise greatl reduced. For example, as shown in Fig, 5, peak noise 46 is the only speckle peak in the image data, which corresponds to noise that was present in both of the difference image data that was correlated. Peak noise 46 has a gray level of 1044. Second peak noise 48 has a gray level of 171 , In contrast, defect 50 has a gray level of 2060. Therefore, the delect becomes detectable using the correlated image data. In this manner, the embodiments have been shown to detect a defect that is not detectable using either optical state alone (e.g., using image data from one illumination angle).

As described above, variation of illumination may be used to change tbe speckle pattern in image data of the wafer thereby reducing speckle noise after image correlation performed using that image data. In addition, although some embodiments ate described above as using two illumination angles that are defined by a 45 degree azimuiiial angle and a -45 degree anmuthai angle for the first and second optical states, the different optical states can. be extended to various illummation angles, Including changing azinmthal angles and/or elevation angles. Output for each of the di.Oereni illumination angles may be acquired in different passes of the wafer.. Correlating image data generated using more than two various il domination angles can be used to inn her suppress the noise and improve the S/N. For example, besides changing the azimuth al angles of illumination, changing el.evaii.on. angle can also vary the speckle pattern greatly thereby increasing the un-correlaiion of noise and further improving me S/N. in this manner, performing the method using any additional optical state may help to further eliminate the un-correiated speckle noise and improve the S/N for defect on. the wafer. In a similar manner, the correlation can be extended, for any channel and any optical state.

As described above, speckle noise in water inspection systems (e.g.. laser-based wafer inspectio systems) is one of the main limitations on DOI detection ability. For example, speckle noise increases the noise level in. inspection image data and reduces S/Ns. Therefore, speckle noise from wafer surface roughness may be one of the mai limitations on achievable defect capture rates in some inspection systems, in addition, nuisance detected as a result of wafer noise (e.g., speckle-like noise from wafer roughness) Is one of the major limitations on DOi detect ability. In particular, relatively high nuisance rates and wafer noise on "' ough" wafers suc as grainy metal etch wafers may limit the performance of inspection systems tha t otherwise have rela tively good optical resolution. In addition, as water design rules continue to shrink, optical inspection systems preferably use shorter wavelengths and larger collectio numerical apertures (NAs). Speckle noise consequently becomes a more dominant o se source.

However, as described above, combining the first image data and the second image data suppresses speckle noise in the additional image data created b the combining step. As such, the methods described herein can be used to reduce nuisance rates and to improve defect capture rates in wafer inspection systems by reducing a main limiting noise factor, namely speckle noise (e.g., caused by wafer surface roughness). Therefore, the embodiments described herein can be used to increase the sensitivity of wafer inspection systems. I addition, as described above, the embodimen s described herein, allow preservation of Hi a imati n coherence while reducing speckle noise thereby enabling the usage of Fourier filtering and improving the S N. in one embodiment, the method includes detecting defects on the wafer using the first image data., detecting detects on the wafer using the second image data, and reporting the defects detected on the wafer as a combination of tire defects detected using any of the first m ge data, the second image data, and the additional image data. For example, detects are detected as described above using the additional image data.. In a similar manner, defect detection may he separately performed rising the .first image data and the second image data. Defect detection performed separately using each of the different image data may be performed in substantially the same manner (e.g., using the same threshold vaiueCs}}, in this manner, the method may include detecting three stib- populatioRS of de ects (i.e., defects detected using the first image data, defects detected using the second image data, and detects detected sin the additional image data). The fhree-suhpopul tions may then be combined to generate the defect population for the wafer. For example, the defect sub-populations may be combined using an OR function based on the image data in which the def ect was detected. Any delects drat are detected at substantially the same position in. any two or more of the image data, may be reported only once to avoid double reporting of any one defect. In this manner, any one defect that is detected in. two different image data may be reported, only once. The defects detected on the wafer may otherwise be reported in any suitable manner.

As described above, the method may also include generating different additional image data. Thai different additional image data may also be used for defect detection as described above. Any defects defected using that different additional image data may be combined with defects detected using any other image data (e.g., the additional image data, the first image data, the second image data, etc.) as described, herein. Furthermore, if the wafer is scanned using more than two different optical states of the inspection system; image data generated using the output from tire third, fourth,, etc. optical states may also be used for defect detection, and any defects detected using that Image date, may be combined it defects detected using other, image data (e.g., the additional image data, the first image data, the second image data, etc.) as described herein.

As described above, a defect that is not detectable m either of the first and second optical states may become detectable in the additional image data, created by image correlation, m this manner, the additional image dat may be used to detect defects on the wafer that are unique in that they are not or cannot e detected using either the first or second image data. As such, the defects detected using the additional image data ma be used to supplement the inspection results with defects that were not or could not be detected by either optical stale iuiivkiaaily.

The embodiments described herein may also include defect feature level fusion across multiple optical states, which may provide opportunities for enhancement of separation, between DOI and nuisances although both may have relatively high S/Ns. For example, in one embodiment, the method includes deiennining values for features of the defects sin the additional Image data. In this manner, ^{^} eross-opdca! state'' featu es of defects can be determined by performing feature calculations based on fused image data. The defect features that are determined using the additional image data may include any suitable defect features, which may be determined in any suitable manner. In this ^• manner, the additional image data may be treated as any other image data when it conies to defect feature determination.

In another embodiment, the method includes detennining values for features of the defects using some combination of the first image data, the second image data, and the additional image data, i this manner, the "cross-optical, state" features can be determined using melhod(s) aud/or aigorithm(s> similar to those used to determine "cross-channel" features. For example, defect features may be determirsed separately using different image data corresponding to the different optical states. Those detect features may then be combined to determine a different defect feature for the defect.. For example, the values for defect features determined separately using different image data corresponding to different optical states may he combined into a new value for the defect feature. In another example, different defect features may be determined for a defect iron* different image data corresponding to different optical, states. Those different defect features may tb.cn be used in some combination to determine another defect feature for the defect. The maimer in which the different image data is us d to determine the detect features may vary depending on the defects for which the features are being determined, the features that are being determined, and the image data itself (e.g,, characteristics of the image dat thai may affect if or how well a feature can b detanined using the ma data), in this manner, values for features of the defects may be determined using all of the information that is available or some subset of the information that is avai lable.

Nuisance filtering can be performed not only in the dimension of each individual optical state, but in the n-dimensional space generated by the multiple optical states, which provides more possibilities for identifying separations between nuisances and DOI. For example, in one embodiment, detecting the defects includes identifying potential defects on the wafer using the additional image data and identifying the defects fey performing nuisance filtering of the potential defects using pixel level information about the potential defects determined using ihe first Image data, the second image data, the additional image data, or some combination thereof. Therefore-, nuisance filtering can be performed by combining the information at the pixel level across multiple optical states (multiple passes), which creates more potential for performance. The potential defects may be identified, in the additional image data, by defect detection, which may be performed as described herein. Nuisance filtering as described above may also be performed for potential defects identified using any other image data described herein. in another embodiment, detecting ihe defects includes identifying potential defects on the wafer using the additional image data and identifying the defects by performing nuisance filtering of the potential delects using values for features of the potential defects determined using the first image data, the second, image data, the additional image data, or some combination thereof Therefore, nuisance filtering can be performed by combining information a the feature level across multiple optical states (e.g., multiple passes), which creates more potential for performance. Identifying the potential defects using the additional image data may e performed as described above. The values lor the features of the potential defects may include any of the values of any of the features described herein and may be determined as described herein. Nuisance filtering as described above may also be performed for potential defects identified using any other image data described herein.

The embodiments described herein may also include image f usion and binning for wafer inspection systems. Binning can. be performed not only in the dimension of each individual optical state, but also in the n^iimensional space generated by the multiple optical states, which provides more possibilities for finding separations between different types of defects. For example, in one embodiment, the method includes binning the defects using pixel level information about the defects determined using the first image data, the second image data, the additional image data, or some combination thereof Therefore, binning the defects may be performed by combining information i the pixel level across multiple optica! stales (e.g., multiple passes), which creates more potential, for performance, in another embodiment, the method includes binning the delects using values for features of the defects determined using the first image data, the second image data, the additional image data, or some combination thereof Therefore, binning the defects may be performed by combining information at the defect feature level across multiple optical states (e.g., multiple passes), which ereai.es more potential for performance. The values for the features of the defects determined using the first mage data, the second image data, the additional image data, or some cem.hvnati.cm thereof may be determined as described further herein.

As described above, different additional image data may be generated for a wafer by combining image data corresponding to different combinations of optical states of an inspection system. fa other words, different fused Image data, may be generated for a wafer. The different fused image data may be used, in all of the steps described, herein in the same manner as the additional linage data. In addition, if different fused image data is generated for a waier, thai different fused image data may be used io determine an appropriate or optimal inspection recipe for a wafer. For example, the different fused image data may be used independently for defect detection, which may e performed as described ^' herein. The defects that are detected using the different fused image data may then be compared. The defects that are uniquely detected using the different fused image data as compared to the individual optical states may be reviewed by defect review to determine which fused image data detected the most unique DOl. Thai same fused image data then be created for other wafers of the same process and/or layer and used for detecting defects on those wafers. In this- maim r, one or more parameters of an inspection recipe may be determined experimentally using the fused image data (i.e., using fused image data generated from experimentally acquired output for a wafer). In addition, the multiple optical states used in the methods described herein ma be determined in this manner using a detect defection method or algorithm ie.g,, the defect detection algorithm in the inspection recipe). As such, the algorithm may be used to perform mode selection for an inspection recipe.

Such an approach to determining one or more parameters of an. inspection recipe may be used to determine the two or more optical states thai can advantageously be used in the methods described herein for a predetermined detect detection method or algorithm. However, such an approach to determinin g one or more parameters of an inspection recipe may also be used to determine one or more defect detection parameters (e,g., a defect detection method or algorithm) that should be used with the two or more optica] states, in this manner, the fused image data may be used to determine any parameter^} of an inspection recipe.

As described above, the image data thai is fused may be generated iron; output generated using different optical states of a single inspection system. However, the methods described herein are not limited to just fusing image data that is generated from output generated using different optical stales of a single inspection system. For example, in addition or alternatively, the image data that is fused may include image data generated from output acquired in different passes (i.e., scans) but with the same optical stale. In o e such embo ment, generating the output as described above using first and second optical states of the inspection system is performed in. one pass. In this embodiment, the method also includes generating additional output for the wafer by scanning the wafer in a different pass with the inspection system using the first or second optical state of the inspection system. In this manner, output using the same (first or second) optical state may be generated i dllietem passes of ibe wafer performed by die inspection system. Generating the additions! output may be further performed as described herein.

Such a method may also include generating different image data for the wafer using the additional output generated in the different pass. Generating the different image data may be performed as described further herein. In addition, such a method may include combining the different image data with the first image data if the different pass is performed using the first optical s ale or the second image data, if the different pass is performed using the second optical state corresponding to substantially the same locations on the wafer thereby creating farther additional, image data for the water. In this manner, the combining step may he performed using the image data acquired in different passes but with the same optical state. This combining step may he further performed as described herein.

Such a method may further include detecting delects on the wafer using the further additional image data. Detecting the defects on the wafer using ibe further additional image data may be performed as described further herein. Such .methods may also include any other step(s) described herein.

Additional methods may include fusing image data that corresponds to different passes of the wafer performed using the same optical state of an. inspection system. For example, another embodiment relates to a different method for detecting defects on a afer. This method includes generating output for a. wa ter by scanning the wafer with an inspection system in first and second passes using a first optical state of the inspection system. Generating the output in this step may be performed as described former herein. The first optical state may include any of the optical, states described herein.

This method also includes generating first image data for the wafer using the output generated in the first pass and second image data, for the water using the output generated in the second pass. Generating the first and second image data, irs this stop may be performed as described -further herein, The first and second image data may include any of the image data described herein,

This method further includes combining the first image data and the second image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer. Combining the first, image data and the second image data in this step may be performed as described further herein, The additional image data, may include any of the additional image data, described herein, in addition, the method includes detecting defects on the wafer using the additional image data.

Detecting the defects on the wafer in this stop may be performed as described former herein. This method may include any other step(s) described herein.

Fusing image data from the same optical state but from different passes may be particularly useful for cases in which the image data is dominated by random noise. For example, if image data, is generated using output generated in a. first pass using one optical state and fused with image data generated using output generated in a second pass using the same optical state, ail. random noise sources may be substantially eliminated, in the fosed image data, while ensuring coincidence of the signal from DOT In addition, as described above, different image data corresponding to the same optical, state can be fused independently, in other words, first and second image data corresponding to the same optical state does not need to be combined with image data corresponding to an already fused, optical state (corresponding to a different optical, state).

Furthermore, in addition or alternatively, the method may be performed using output generated by different inspection systems. For example, in one embodiment, the method includes generating output for the wafer by scanning the wafer with a different inspection system. Generating the output for the wafer with, the different inspection system may be performed as described herein. The different inspection system may be a DF or BF system. For example, the inspection system may be a OF system, and the different inspection system may be a BF system, In another example, the inspection system may be a DF system, and. the different inspection system may be a different DP system (e.g., a DF system having a different configuration than the inspection system). The different, inspection systems may be configured as described herein or may have any other suitable configuration known in the art

Sneh a. method may also include generating third image data for the wafer using the output generated using the different inspection system. Generating the third image data, may be performed according to any of the embodiments described herein. In addition, such a method may include combining the third image dm with the first or second image data corresponding to substantially the same locations on the wafer thereby creating further additional image data for the wafer. Combining d>e third image data with the .first or second image data may be performed in this embodiment according to any of the embodiments described herein. Such a method may further include detecting defects on the wafer using the further additional image data. Detecting the defects on the water using the further additional image data may be performed as described further herein. Such an embodiment may include any other step(s) described herein. In this manner, the method may include using output collected from different inspection systems and fusing image data generated using the output collected from, the different inspection systems.

In. another embodiment, a method for detecting defects on a wafer includes generating output for a wafer by scanning the wafer with first and second inspection systems. Generating the output for the wafer in this step may be performed as described further herein. The first and second inspection systems may include any of the different inspection systems described herein. Scanning the water with the first and second inspection systems may he performed using the same or substantially the same optica! state. Alternatively, scanning the wafer with the first and second inspection systems may be performed using different optica! stales. In this manner, the diff erent optical states corresponding to the different image data that is fused may be different optical, states of different inspection systems. For example, image data corresponding to optica! state A of inspection system X ears be combined as described further herein with image data corresponding to optical state B of inspection system Y. Optical slates Λ. and. B can be identical or different. The same or substantially the same optical state and ire different optical states may include any of the optical states described, herein. The method also includes generating first image data for the wafer usi the output generated using the first inspection system and second image data for the wafer using the output generated using the second inspection system. Generating the first and second image data may be perforated as described further herein. The first and second image data may include any of the image data described herein.

The method further includes combining the first image dm and the second, image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer. This combining step may be performed as described further herein. in addition, the method includes detecting defects on. the wafer using the additional image data, ^' Detecting the defects on the wafer in this step .may be performed as described further herein.

As described above, the method may include generating output using different inspection systems and generating different image data using the output generated using the different: inspection, systems. However, the method may not necessarily include generating the output using all of the different inspection systems. For example, the output generated using one or more of the different inspection systems may be acquired from one or .more storage media in which the output has been stored (e.g., by the one or more different inspection systems). The acquired output generated using the different inspection systems may then be used as described, further herein. In this manner, the methods described herein can perform image fusion regardless of the origin of the output used to generate the image data that is fused., In addition, the output generated by the different inspection syst ms that is used in the embodiments described herein may necessarily be generated using different optical states (as would be the case if the different inspection systems include a DF inspection system and a BF inspection system or DF inspection systems having completely different (e,g., n n-overlapping) configurations). However, the output generated by the different inspection systems that is used in die embodiments described herein may be generated using the same optical state or substantially the same optical state (as may be the case if the different inspection systems include two inspection systems having the same configuration or relatively similar configurations).

The embodiments described herein may also include storing results of one or more steps of one or more methods described herein in a storage medium. The results may include any of the results described herein. The results may be stored in an manner known in the an. The storage medium may include any suitable storage medium known in the art. After the results have been stored, the results can be accessed in the storage medium and used by any of the method or system embodiments described herein, any other method, or any other system. Furthermore, the results may be stored

"permanently," "semi-permanentl ," temporarily, or for sonic period of time. For example, the storage medium may be random access memory CRAM), and the results may not necessarily persist indefinitely in the storage medium.

Another embodiment, relates to a non- transitory computer-readable medium containing program instructions stored therein for causing a computer system to per.lb.rnr a compiler-implemented method for detecting defects on a wafer. One such embodiment is shown in Fig. 6. For example, as shown, in Fig. 6, non-transitory computer-readable medium 52 contains program instructions 54 for causing computer system 56 to perform a computer-implemented method for detecting defects on a wafer.

The computer-implemented method includes acquiring output tor a wafer generated by scanning the wafer with an inspection system using first and second optical states of the inspection system, The first and second optica! states are defined by different values for at [east one optical parameter of the inspection system. The first and second optical states may include any of the optical states described herein. The different values for the at least one optical parameter of the inspection s stem may include any of the different values described herein. The at least one optical parameter of the inspection system may include any of the optical parameters described herein. The inspection, system may include any of the inspection systems described herein.

Acquiring the output genera ted for the wafer may be perforated using the inspection system. For example, acquiring the output, may include using the inspection system to scan light over the aier and to generate output responsive to light scattered from the wafer detected by the Inspection system daring scanning, In this manner, acquiring the output may include scanning the wafer. However, acquiring the output does not necessarily inciude scanning the wafer,. For example, acquiring the output may include acquiring the output from a storage medium in which the output has been stored (e.g., by the inspection system). Acquiring the output from the storage m diu may he perforated In any suitable manne , and the storage medium from which the output is acquired may inciude any of the storage media described herein. The output may include any of the output described herein.

The computer-implemented method also includes generating first image data for the wafer using the ouipui generated using the Inst optical state and second image dais for the wafer using the ouipui generated using the second optical state. Generating the first image data and the second image data may be performed as described further herein. The first and second image data may include any such image data described herein.

The computer-implemented method further includes combining the first image data and the second image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer. Combining the first image data, and the second image data may be performed as described further herein. The additional image data may include any of the additional image data described herein. In addition, the method includes detecting defects on the wafer using the additional image data. Detecting the defects on the wafer may be performed as descsibed further herein. The delects detected on the wafer may mclude any of the defects described liereim The computer-implemented method for which the program instructions are executable may include any other step(s) of any other memod(s) described, herein.

Program instructions 54 implemeutirsg methods suc as those described, heret .may be stored on non-transitory compirier-readable medium 52. The nou ransitory computer-readable med um may be a storage medium such as a read-only memory, a RAM, a magnetic or optica) disk, or a magnetic tape or any other suitabl computer- readable medium known in the art.

The program instructions may ^" be implemented in any of various ways, including procedure-based techniques, component-based techniques,, and/or object-oriented

techniques, among others. For example, die program instructions may be implemented using atlab. Visual Basic, ActiveX controls, C, OH- objects, C# _> JavaBeans, Microsoft

Foundation Classes f "MFC"), or other technologies or methodologies, as desired.

Computer system 56 may lake various forms, including a. personal computer system, .mainframe computer system, workstation, system computer, image computer, programmable image computer, parallel processor, or any other device known m the art. In general, the term, "computer system" may be broadly defined to encompass any device having one or more processors, which executes instructions from a memory medium.

An additional era ltr sM relates to a system configure to detect detects on a wafer. One embodiment of such, a system is shown in Fig. 7. As shown in Fig. 7, system. 58 includes inspection subsystem 60 and computer subsystem 80. The inspection

subsystem ts configured to generate output for a. wafer by scanning the wafer using first and second optical slates of the inspection subsystem. For example, as shown in Fig. ?, the inspection subsystem includes light source 62. Light source 62 may include any suitable light source known in the art such as a laser, Light source 62 is configured to direct light to polarizing component 64, which may mclude any suitable polarizing component known in. the art. In addition, the inspection subsystem may include more than one polarizing component (not shown), each of which may be positioned

independently in the path of the light front the light source. Each of the pokriziog components may be configured to alter the polarisation of the light from the light source in a different manner. The inspection subs stem may be configured to move the polarizing components Into and out of the path of the light from the light source in any suit-able manner depending on which polarization setting is selected for illumination of the wafer during a scan. The polarization setting used for the illumination of the wafer during a scan may include F, S, or circularly polarized (€},

Light exiting polarizing component 64 is directed to wafer 66 at an oblique angle of incidence, which may include any suitable oblique angle of incidence. The inspection subsystem may also include one or more optical components (not shown) that are configured to direct light from light source 62 to notarizing component 64 or from polarizing component 64 to water 66. The optical components may include any suitable optical components known in the art such s, but not limi ed to, a reflective optical component. In addition, the light source, the polarizing component, and/or the one or more optical components may be configured to direct the tight to the wafer at one or more angles of incidence {e.g., an oblique angle of incidence and/or a substantially normal, angle of incidence). The inspection subsystem, may be configured to perform the scanning by scanning the light over the wa er in any suitable manner. light scattered from water 66 may be collected and detected by multiple channels of the inspection subsystem during scanning, For example, light scattered from wafer 66 at angles relatively close to normal may be collected by lens 68, Lens 68 may include a refractive optical element as shown i Fig, 7. In addition, sens 68 may include one or more refractive optical elements and/or one or more reflective optical elements. Light collected by lens 68 may be directed to polarizing component 70, which may include any suitable polarizing component known in the art. In addition, the inspection subsystem may include more than one polarizing component (not shown), each of which may be positioned independently in the path of the light collected by the lens. Each of the polarizing components may be configured to altar the polarization of the light collected by th lens in a different manner. ^' The inspection subsystem may be configured to move the polarizing components into and out of the path of the light collected by the lens in any suitable manner depending on which polarisation setting is selected tor detection of the light collected by lens 6H during scanning, The polarization setting used for the detection of the Sight collected by km 68 during scanning may include any of the polarization settings described herein (e.g., P _s S, and N) _t

Light exiting polarizing component 70 is directed to detector 72, Detector 72 may include any suitable detector known in the art such as a charge coupled device (CCD) or another type of imaging detector. Detector 72 is configured to generate output that is responsive to the scattered light collected by lens 6S and transmitted by polarizing component 70 if positioned in the path of the collected scattered, light. Therefore, leas 68, polarizing component 70 if positi ned in the path of the light collected by lens 68, and detector 72 form one channel of the inspection subsystem. This channel of the inspection subsystem may include any other suitable optical components (not shown} known in. the art such as a Fourier filtering component.

Light scattered from wafer 66 at different angles may be collected by lens 74, Lens 74 may be configured as described above. Light collected by lens 74 may be directed to polarizing component 76, which may include any suitable polarizing component known in the art. In addition, the inspection subsystem may include more than one polarizing component (not shown), each of which ma he positioned independently in the path of the light collected by the lens. Each of the polarizing components may be configured to alter the polarization of the light collected by the lens in a differen manner. The inspection subsystem may be configured to move the polarizing components into and out of the path of the light collected by the lens in any suitable manner depending on which polarization setting is selected for detection of the light collected by Sens 74 during scanning. The polarization setting used for detection of the light collected by lens 74 during scanning may include P„ S, or N, Light exiting polarizing component 76 is directed to detector 78, which may be configured as described above. Detector 78 is also configured to generate output that is responsive to the collected scattered light that passes through polarizing component 76 if positioned in the pain of the scattered light. Therefore, lens 74, poiarking component 76 if positioned in the path of the light collected by lens 74, and detector 78 may form another channel of the inspection subsystem. This channel may also include any other optical components (not shown) described, above. n some embodiments, lens 74 .may be configured to collect light scattered from the wafer at polar angles from about 20 degrees to abou 70 degrees. In addition, lens 74 may be configured as a reflective optical, component (not shown) that is configured to collect, light scattered from the wafer at azimuthal angles of abou t 360 degrees.

The inspection subsystem shown in Fig, ? may also include one or more other channels (not shown). For example, die inspection subsystem may include an additions! channel, which may include any of the optica.! components described herein such as a le s, one or more polarising components, and a detector, configured as a side channel The lens, the one or more polarizing components, and the detector may be farther configured as described herein. In one such e am le, the side channel may be configured to collect and detect light that is scattered out of the plane of incidence (e.g., the side channel may include a lens, which is centered in a plane thai is substantially

•perpendicular to the plane of incidence, and a detector configured to detect light collected by the lens).

The first and second optical states are defined by different values for at least one- optical parameter of the inspection subsystem. The first and second optical states may be defined by any of the different values for any of the optical parameters of the inspection subsystem described herein. In addition, the values of any of the optical parameters may be altered in any suitable manner if necessary between passes. For example, if the different values arc different values of illumination polarization states, between passes polarizing component 64 may be removed, and/or replaced as described herein with a different polarising component, in another example, if the different values are different illumination angles, the position of the light source and/or any other optical components (e.g., polarizing component 64) used to direct the light to the wafer may be altered between passes in any suitable manner.

Output generated by the detectors during scanning may be provided to computer subsystem SO. For exam le, the computer subsystem may be coupled to each f the detectors (e.g., by one or more transmission media shown by the dashed lines in Fig. 7, which may include any suitable transmission media kno n the art) such that the computer subsystem may receive the output generated by the detectors. The computer subsystem may be coupled to each of the detectors in any suitable manner. The output generated by the detectors during scanning of the wafer may include any of the output described herein.

The computer subsystem is configured to nerate first image data for the water using the output generated using the first optical state and second image data for the wafer using the output generated using the second optical state. The computer subsystem may be configured to generate the first image data, and the second image data according to any of the embodiments described herein. The first image data and the second image data may include any such image data described herein.

The computer subsystem is also configured to combine the first image data and the second image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer. The computer subsystem, may be configured to combine the first and second image data according to any of the embodiments described herein. The additional image data, may include any of the additional image data described herein.

The computer subsystem is further configured to detect defects on the wafer using the additional image data. The computer subsystem may be configured to detect ie defects according to any of the embodiments described herein. The detects may include an of the defects described herein. The computer subsystem may be configured, to perform any other step(s) of any ^• method einbodiment(s) described herein. The computer subsystem may be further configured as described herein. The inspection subsystem may also be further configured as described herein. Furthermore, the system may be further configured as described herein.

Another embodiment relates to a method ibr detecting defects on a wafer. In general, the embodiments described further herein relate to detecting delects on a wafer by "focus fusion " For example, as described further herein, wafer noise In wafer inspection systems is one of the main limitations of defect detection ability. Wafer noise can be caused by the imperfection of processed pattern features and surface profiles. in a laser-based coherent system, speckle noise, introduced by surface roughness and line edge roughness, is the dominating wafer noise. As wafer design rules continue to shrink, shorter wavelengths and larger collection NAs are preferred for optical, inspection ste s. Speckle noise consequently increases to a more dominant, noise source. The embodiments described further herein improve defect capture rate and reduce false count in wafer inspection systems by suppressing speckle ooise.

The method includes generating output for a wafer by scanning the wafer using first and second differen values for focus of an inspection system and the same values for all other optical parameters of the inspection system. For example, in other embodiments described above, which may be generally referred to "image fusion" methods and systems, two or more images are acquired of the same wafer locatio by different optical modes, which can be different in illumination angle, collection ^' A, wavelength, or polarisation. By carefully choosing the different optical modes, the defect signals in the images can all be healthy while the noise can be suppressed due to its uncorreiated nature thereby improving defect capture rate and false count.

Image fission embodiments described above have been proven to improve delect detection ability. However, some difficulties can exist during implementation. For example, two or more o tic s modes have to be- chosen among which noise Is

uneorrelated. Modern inspection systems can be substan.tial.iy complex due to the large number of degrees of freedom in the systems such as illumination angle, collection NA, wavelength, illumination and collection polarization, etc. As a result, the num er of optical modes for a certain inspection can. be enormous, The number of possible combinations of any two optical modes is even more enormous. More specifically, the .number of possible combin ations of any two optica! modes is *( - ! ).-·¾ where H is the number of optical modes for a single pass. If no prior knowledge is availa le, marry, if not ail, of the possible mode combinations have to be tested in. recipe setup before the optimal combination is found. This testing will put a huge burden on application, engineers and may make recipe setup for a new wafer difficult. in addition, while noise is not correlated in the chosen modes, the improvement in defect detection is mainly a function of the signal strengths of the de-feels- in each mode. The stronger the defect signal is in each mode, the better delect capture Is after image fusion. However, with a constraint of mode selection being that images acquired by the modes need to have- uneorrelated noise, it is safe to say that the chosen modes are usually not the modes of best S/N indi idually. In fact, if ike defect signal in one .mode is particularly weak, there may not be any improvement from image fusion. As described further herein, focus fusion can suppress noise while using the one mode with the best S Therefore, the embodiments described further herein c n provide improvements to other image fusion embodiments.

Furthermore, depending on the differenceis) between the modes used tor image fusion, there will be- loss in either throughput or light budget. For example, if two modes differ in .illumination angles, two images cannot be generated simultaneously. A. double pass inspection is needed and therefore throughput is halved. If two modes are different in collection A, the image has to be split at the pupil plane so that each channel will have individual control over its pupil. However, the image has to be combined in the latter part of the optical train to pass through the same zoom lens.. This splitting and combining process will gener lly introduce Loss in optical power. I ierstbre, the light budget is reduced, in one embodiment, generating the output is performed with coherent lig . For example, former approaches used to reduce speckle noise usually involve reducing coherence of die illumination light from laser sources by transmitting the light through an optical diffuser or vibrating optical fiber. However _* reduction of illumination light coherence may reduce usage of Fourier niters arid arm pa erns cannot be suppressed, as a result. These limitations greatly degrade the defect detecting ability in an OTL oblique angle illumination DF architecture, Generating the output using coherent light may be further performed as described herein, and the coherent light may include any of the coherent light described .further herein.

In. some embodiments, the same values for all other optical parameters of the inspection system include values for iliumiuadon and collection parameters of the inspection system. Generally, the iihiniioadon and collection parameters collectively define the optical state of the inspection system that is used for performing an. inspection pass or an inspection. In this manner, te first and second output is preferably generated using the same optical state but with different focus values. In one embodiment, the same values for all other optical parameters of the inspection system, define an optical state of the inspection system that produces the best S N for the wafer compared, to other optical states of the inspection system. For example, the embodiments may include selecting the one optical mode with the best S/N as in the case of one channel, inspection. In addition, generating the output may include acquiring two images of the same Held of view on the wafer with slightly different focus.

In one embodiment, the first and second output is separately generated, during the generating step. Separately generating the first and. second, output may be performed, as described further herein. For example, the first output may be generated using one detector during a first pass of the inspection while the second output may be generated using the same detector during a second pass of the inspection, and the first ami second passes .may be performed using different values of focus. However, the first and second output may be generated separately ami simultaneously as described further herein, for example, in another embodiment, the first and second output is generated separately and simultaneously using different detectors of the inspection, system.. The different detectors may be configured as described further herein,

In. another embodiment, the first a d second output is generated separately and simultaneousl using first and second detectors of the inspection system, and the first and second detectors have offset positions along the optical axis of me inspection system. For example, acquiring two images of the same field of vie w on. the wafer with slightly different focus can be achieved by offsetting two channel sensors along the optical axis or by inserting a focus offset glass plate in. one of channels. In some embodiments, the first and second output is genera ed separately and. simultaneously using first and second detectors of the inspection system, the first detector is positioned on a. first side of the focal plane of the inspection system, and the second detector is positioned on a second, side of the focal plane opposite to the first side. For example, the image plane of the- first, channel may be at the near side of the focal, plane, and the image plane of the second channel may he at the far side of the focal plane. In one embodiment, the first and second different values for focus are different by a value equal to the depth of focus of the inspection system. For example, the defocus shift between two image planes such as those described above may be about the depth of focus. The two images can also be spin right before the sensors. In this manner, the two images can be acquired by splitting the Image path to two channels with different sensor locations or any other method.

In. one embodiment, the first and second different values for focus are out~o.f- f cus values of the inspection system, in another emb iment, the first and second different values for focus do not include an in- ocus value of an inspection system. For example, many in spection processes include altering the focus value(s) of the inspection system as a. wafer is scanned to account for variations in the wafer surface to thereby keep the wafer surface in focus during as notch of the scan as possible. ^' Therefore, the inspection system may perform the scanning at a number of different focus values with respect to an absolute reference. However, with respect to the wafer surface itself, he inspection system will perform the entire scan at roughly the in-ibeus value with generally negligible variations from the m~ioeus value. In contrast, she embodiments described herein perform enti e scans at intentionally oe -ofobcus values for the inspection system and with values thai, are cmi-of-foeus with respect lo the wafer surface being inspected. Therefore, the embodiments described, herein are entirely different than methods of scanning that include dynamically altering the locus of the inspection system to account for variations in the wafer.

The method also includes generating first image data, for the wafer using the output generated using the first value for focus and second image data for the wafer using the output generated using the second value or focus. Generating the first image data and the second image data may be performed as described further herein, In addition, the first and second image dam may include any of the image data described further herein, in one embodiment, noise in the first image data is ^■ ^correla ed with noi e in the second image data. For example, a speckle noise correlation function, between two adjacent image planes at different defocus positions is invariant to the image plane position. The function is only dependent on the distance between these two image planes along the optical axis. In this manner, the speckle pattern from wafer noise {e.g., surface roughness. Sine edge roughness, etc.) varies when an image is defbeused (which can he illustrated using finite-difference time-domain (FDTD) simulation which is a known modeling technique), in particular, speckle noise varies as the image plane moves .from defocus to focus and then to defocus in the other direction. As such, the speckle pattern, in different image planes at both ends of the depth of focus is uaeorre!ated, lo. this manner, if two image planes are placed far enough along the optical axis (e.g., approaching the depth of focus), speckle noise between the first and second image data is uncorrected. In addition, when two channels are taken at a defocus offset equal to the depth of locus, the uncotre!ated nature between noise of the two channels is guaranteed. As such, the uncorrected nature of the speckle noise between the first and second dsfocused image data ma be guaranteed, in this manner, there is no need to search for two optical .modes, among the enormous number of ossible optical mode combinations for any one inspection system, that provide uncorrected noise. This will tremendously reduce the effort of recipe setup because the process is simplified to fed the single best mode and to use focus fusion to suppress the noise. in another embodiment, a signal strength of a detect in the first image data is th same as a signal strength of the detect in die second image data, In other words _* the defect signal intensity may be the same between ire two defoc sed image data, For example, detect signal strengths in the two channels described herein may be identical In this manner, there is no concern that defect signals of one channel of different modes selected for image fusion are particularly weak (e.g., S/N less than one) and. that therefore image fusion will not provide improvement over single channel inspection. For example, for each channel, if the image plane is offset from the optimal focus plane by about half of the depth of focus, the detect signal may be reduced in theory by only 30% from the maximum signal at the ideal focus plane of the best optical mode. Furthermore, in many cases, a defect signal that is reduced by about 30% from the best mode is still stronger than the signal of any other mode, and the focus fusion, embodiments described herein, remove ail the uncertainty in the defect signal strengths in the modes used for linage fusion. In addition, with current autofocus mechanisms, signal reduction from the optimal focus location is accepted as a result of auto focus error and system slope (stage, chuck, and wafer), in this manner, for practical applications, as long as the image plane is within the depth of focus, which i is in this case, the detect signal should he considered optimal on average,

The value of the image fusion embodiments described herein has been proven by theoretical analysis ami tests done on current tools and benches. However,

implementation of the image fusion embodiments can require more complexity of optical design and inspection recipe setup. Therefore, tool cost can increase and. user operation ca become relatively time consuming and difficult. Also, depending on the combination of optical modes chosen, the throughput or light budget, may be reduced. m contrast, the focus fusion embodiments described herein, will simplify the optical architecture and reduce optics costs compared to the image fusion embodiments. For example, in image fusion in which the pupil may be split, the image may be split at the pupil plane and thou combined before a zoom lens. As such, the optical geometry ear- be complicated and therefore tool cost increases. Also, a 50% loss of optical power may be inevitable under other currently used designs, hi the ease of focus fusion, however, image splitting can be performed right before the detectors. Therefore, there is minimal, optical geometry change to the system, until, the last element, in tkk manner, the locus fusion embodiments require minimal change from a single channel inspection system. As such, the focus fusion embodiments can achieve the same or better defect capture enhancement as in image fusion with the simplicity of a single channel inspection system. In addition, the focus fusion embodiments preserve the same throughput and light budget as a single channel inspection system. Furthermore, there will be no optical power loss introduced by image split and recombination.. Moreover, since the optical state does not need to be changed between the focus fusion channels except tor a slight de-focus in the sensor or last lens element in the collection optics, there is minimal effort in system engineering to align the images from the two channels.

The method also includes combining the first image data and the second image data corresponding to substantially the same locations on the wafer thereby creating additional image data, fo the wafer. In one embodiment the combining step includes performing image correlation on the first image data and the second image data corresponding to substantially the same locations on. the wafer. For example, combining the first and second image data may include applying a correlation algorithm

( ^' multiplication, for example) to combine two images into one image. The combining step ma be further performed as described herein. For example, in another embodiment, the combining step is performed at the pixel level of the first and second image data. In one embodiment, as described further herein, portions of the additional image data. that, correspond to the defects: have greater S/Ns than portions of the first and. second image data that are combined to create the portions of the additional image data, in another embodiment, the additional, image data has less noise than the first and second image data. In an additional embodiment, the additional Image data has less speckle noise than the first and second image daia. For example, as described further herein, since noise and speckle noise will be ttncorrelated in the first and second mage data, combining the first and second image daia will reduce the o se and speckle noise in the image data resulting from the combining step.

The method further includes detecting detects on the wafer using the additional image data. Detecting defects on the wafer using the additional image data may be performed as described further herein (e.g., applying a suitable defect detection algorithm to the additional image data and or performing a defect detection method using the additional image data}, hi some embodiments, defect detection is not performed prior to the combining step. For example, a correlation algorithm (e.g., multiplication) may be applied as described above to merge the first and second image data into additional image data., and then a defect detection, algorithm may be applied once on the additional image data or the "final image." In one embodiment,, the method includes detecting defects on the w fe using the first image data, detecting defects on die wafer using the second image data, and reporting the defects detected on the wafer as a combination of the defects detected using any of the fi st image data, the second image data, and the additional image data. These steps may be performed as described farther herein. For example, defect candidates can be identified using both the first and second image data by setting the same threshold or using any other defect detection algorithm. In addition, the coordinates of the defect candidates may be compared between two images, and a defect may be identified as a true defect only when the coordinates of the defect candidates match between two images and the defect densities at candidate sites are close within a certain range.

Each of the .focus fusion embodiments described herein may include any other siep(s) of any other embodiments described herein. Each of the focus iosio

embodiments described herein may be performed by an of the systems described herein. An additional embodiment relates to a non-transitory computer-readable med um containing program instructions stored therein for causing a computer system to per form a computer-implemented method for detecting defects on a wafer. The computer- implemented method includes acquiring output for a wafer generated by scanning the wafer using first and second different values for focus of an inspection system and the s me vales for all other optical parameters of the inspection s stem, which may be performed as described further herein. The computer-implemented .method also includes generating first image data for the wafer using the output generated using the first value for focus and second image data for the wafer using the output generated using the second value for focus, which may be performed as described further herein, lb addition, the computer-implemented method includes combining the first image data and the second image data corresponding to substantially the same locations on the wafer thereby creating additional image data for the water, which may be performed as described, riher herein. The compiUer-impiemenied method further includes detecting defects on the water using the additional image data, which, ma be performed as described further herein. The computer-implemented method may include any other step(s) described, herein. In addition, the non-transitory computer-readable medium may he configured as described further herein.

An additional embodiment relates to a system configured to detect defects on a water. One embodiment of such a system is shown in Fig, 8. As shown in Fig. 8, system 82 includes inspection subsystem 84 and computer subsystem 80. The inspection subsystem is configured to generate output for a wafer by scanning the wafer using first and second different values for focus of the inspection subsystem and. the same values for all. other optical parameters of the inspection subsystem. For example, as shown in Fig. 8, the inspection subsystem includes light source 86, which may include any of the light sources described herein. Light source 86 is configured to direct light to polarizing component 88. which may include any suitable polarizing component known in the art. In addition, the inspection subsystem, may include more than one polarizing component (not shown), each of which may be positioned independently in the path of the light, from the light source. Each of the polarizing eonrponenls may he configured as described herein. The inspection subsystem may be configured io ove the polarizing components as described further herein. The polarisation, setting used for the illumination of the wafer dining a scan may include P. S, or C.

Li ght exiting polarizing component $8 is directed to wafer 66 at an oblique angle of incidence as described farther herein. The inspection subsystem ma also include one or more optical components (not shown) as described further herein that are configured to direct light from light source 86 t polari ing component SS or from pokrking component 88 to wafer 66. In addition, the light source, the po larking component, and/or the one or more optical components may be configured as described further herein. Th inspection subsystem may he configured to perform the scanning as described further herein.

The inspection subsystem acquires two defbe sed images (channels) of the same field of vie on the wafer. For example, light scattered from wafer 66 ma be collected and detected by multiple detectors of the .inspection subsystem during scanning. In one such example, light scattered .from wafer 66 at angles relatively close to normal may he collected by lens 90, which may be configured as described further herein. Light collected by lens 90 may be directed to polarising component 92, which may include any suitable polarizing component known in the art In addition, the inspection subsystem may include more than one polarising component (not shown), each of which may be positioned independently in the path of the light collected by the lens, Each of the polarizing components may be configured as described further herein. The inspection subsystem may he configured to move the polarizing components as described further herein. The polarization setting used for the detection of the light collected by lens 90 during scanning may include any of the polarization settings described herein (e.g., P, S, and N).

Light exiting polarizing component 92 i directed to beam splitter 94, which may include any suitable beam splitter (such as a 50/50 beam splitter), The beam splitter in the image path may not necessaril be at the pupil lane and can create two channels and enable s ngle pass focus fusion Inspection. Light reflected by the beam splitter may be directed to first detector 96, and sight transmitted by the beam splitter may be directed to second detector 98. Detectors 96 and 98 may be configured as described further herein. Therefore, lens 90, polarizing component 92 if positioned in the path of the light collected by lens 90, and detector 96 form one channel of the inspection subsystem,, and lens 90, polarizing component 92 if positioned in the path of the light collected by lens 90, and detector 98 form another channel of the i nspecti on subsystem. The channels of the inspection subsystem may include any other suitable o tica! components (not shown) such as those described further herein. The inspection, subsystem shown in Fig. 8 may also include one or more other channels (not shown) such as a side channel described further herein, in this manner, the inspection subsystem can acquire two de focused images (channels) of the same field of view on the wafer without changing the optical state between these two channels except for a slight defoeus in the sensors or last lens element in the collection optics.

Output generated by the detectors during scanning may be provided to computer subsystem 80, For example, she computer subsystem, may be coupled to each of the detectors as described further herein. The output generated by the detectors during scanning of the wafer may include any of the output described herein . Computer subsystem 80 is configured to generate first image data, for the wafer using the output generated using the first value tor focus and. second, image data for the wafer using the output generated using the second value for focus. The computer subsystem may generate the first and second image data as described further herein. The computer subsystem, is also configured to combine the first image data and the second image dat corresponding to substantially the same locations on the wafer thereby creating additional image data for the wafer. The computer subsystem may be configured to combine the first and second image data as described further herein, m addition, the computer subsystem is configured to detect defects on the wafer using the additional image data. The computer subsystem may he configured to detect the detects as described former herein. The computer subsystem may be configured to e form any other step(s) of any method e bodimenti ) described herein. The computer subsystem may be further configured as described herein. The inspection subsystem may also be further configured as described herein. Furthermore, the system may be further configured as described herein.

It is noted thai Figs, 7-8 are provided herein to generally illustrate con.figu.ra lions of an inspectio subsystem that may be included in the system embodiments described herein. Obviously, the inspection subsystem configurations described herein may be altered to optimize the performance of the inspection subsystems as is normally performed when designing a commercial inspection system. In addition, the systems described herein may be implemented using an existing inspection system (e.g., by adding functionality described herein to an existing inspection system} such as the Puma 9000 and 9ixx series of tools that are commercially available from KLA- eucor. For some such systems, the methods described herein may be provided as optional functionality of the system (e.g., in. addition to other functionalit of the system).

Alternatively, the system described herein may be designed "from scratch" to provide a completely new system.

Further modifications and alternative embodiments of various aspects of the invention may be apparent to those skilled, in the art in view of this description. For example, systems and methods for detecting defects on a wafer are provided.,

Accordingly, this description is to be construed as illustrative only and. is for tire purpose of leaching those skilled in the art the general manner of carrying out the i vention, it is to be understood that the forms of the invention shown and described herein are to be taken as the presently preferred embodiments. Elements and materials may be substituted .for those illustrated and described herein, parts and processes may be reversed, and certain features of the invention may be utilized independently, all as would he apparent to one skilled in the art alter having the benefit of this description of the invention. Changes may be made in the elements described herein without departi g from the spirit and scope of the invention as described in the following claims.