Darkfield defect inspection with spectral contents

The present invention relates to detection of defects and, more particularly, to defect inspection and identification.

The fabrication of semiconductor devices, or similar types of devices, typically includes a number of processing steps to form desired features and multiple layers on a substrate. In general, the processing steps used to form semiconductor devices on a substrate are well known and includes processes such as deposition, etching, photolithography, and chemical mechanical polishing.

Due to the complexity and need for accuracy in semiconductor fabrication, monitoring and evaluation of the processing is often required. One type of necessary monitoring is defect inspection. Defects are sometimes introduced during processing and typically include contaminants, or other non-uniformities, and scratches. The presences of defects may result in malfunctioning devices and a reduction in the overall yield. Accordingly, defect inspection is used to identify chips with defects as well as to identify processing steps that are causing defects to prevent further defect generation in subsequent processing.

In accordance with an embodiment of the present invention, a metrology device produces broadband illumination on a portion of a substrate at an oblique angle of incidence. In one embodiment, the illumination may be a line of illumination. A darkfield detector collects the darkfield image from the illumination, while another detector collects the spectrally reflected light from the illumination. In one embodiment, the angle of incidence of the illumination may be variable to reduce or eliminate interference with the darkfield image from any diffracting structures on the substrate. Alternatively, the darkfield detector may be variable to receive the darkfield image without interference, or a filter may be used to filter out any interference. A processor uses the data from the darkfield detector to determine if a defect is present on the substrate and uses the data from the detector in the spectral path to identify the material composition of the defect.

In another embodiment, a method of detecting and identifying the material composition of defects on a substrate includes producing broadband light illuminating a portion of the surface of a substrate at an oblique angle of incidence. In one embodiment, the broadband light is formed into a line of illumination on the surface of the substrate. The illumination is scanned across the surface of the substrate, e.g., by moving the substrate or the optics (or both), while the darkfield image is collected with a first detector and the spectrally reflected light is collected with a second detector. The presence of a defect on the surface of the substrate is determined using the collected darkfield image and the material composition of the defect is determined using the collected spectrally reflected light.

In another embodiment, an apparatus includes a stage for holding a substrate and a metrology device for detecting and identifying defects on the substrate on the stage. At least one of the stage and the metrology unit is movable with respect to the other to produce a relative movement between the substrate held on the stage and the metrology unit. The metrology unit includes a variable angle light source that produces a polychromatic beam of light that is incident on the substrate at an oblique angle and at least one optical element that produces an illumination line on the substrate held on the stage, the at least one optical element collimates the beam of light in a first direction and focuses the beam of light in a second direction that is orthogonal to the first direction. A variable angle detector collects the spectrally reflected light from the substrate. The detector includes an array of photodetecting elements that collect the spectrum of the spectrally reflected light along a first dimension during the relative movement between the substrate and the metrology unit. Another detector collects a darkfield image from the illumination line on the substrate. The darkfield detector is an array of photodetecting elements that collects the darkfield image of the length of the illumination line during the relative movement between the substrate and the metrology unit.

FIG. 1 is a schematic view of a metrology device that uses darkfield defect inspection along with spectrometry to detect defects as well as identify the material composition of the defect, in accordance with an embodiment of the present invention.

FIG. 2 illustrates a top view of a substrate with an illumination line that is produced by the metrology device.

FIGS. 3A , 3 B, and 3 C illustrate top views of different embodiments of the detector used in the spectrally reflected light detector.

FIGS. 4A , 4 B, and 4 C illustrate top views of different embodiments of the image detector used in the darkfield detector.

FIG. 1 is a schematic view of a metrology device 100 that uses darkfield defect inspection to detect defects along with spectrometry to identify the material in the detect, in accordance with an embodiment of the present invention. As shown in FIG. 1 , metrology device 100 includes a separate spectrometer path 110 and a darkfield path 120 to detect defects on the surface of a substrate 102 that is held on a stage 104 . The darkfield path 120 is used to detect the presence of defects on the surface of the substrate 102 , while the spectrometer path 110 is used to determine the type or material composition of the defect.

The metrology unit 100 includes a light source 112 , which may be, e.g., a broadband light source that produces light with wavelengths between 190 and 800 nm, and an optical element that produces a line of illumination. The optical element may be, e.g., a lens 114 , such as a cylindrical lens or lens system, that focuses the light from light source 112 in one direction and collimates the light in an orthogonal direction to produce the line of illumination that is incident on the substrate 102 . If desired, the reflective optics may be used as the optical element to produce the line of illumination.

FIG. 2 illustrates a top view of a substrate 102 with an illumination line 130 that is produced by lens 114 . The illumination line 130 may be, e.g., 1 mm or larger in length and between 5 μm to 15 μm wide. The resolution is improved, however, with a narrow line, and thus, the illumination line 130 is preferably less than 5 μm wide, e.g., 1 μm to 5 μm or less. In operation, the substrate 102 or the metrology device 100 (or a combination of the two) is moved to scan the illumination line 130 across the surface of the substrate 130 . By way of example, the stage 104 may be movable in the Cartesian Coordinate (XY) system illustrated in FIG. 2 . Alternatively, the stage 104 may move in Polar Coordinate (R-θ) system, however, it may be necessary to correct for rotational effects, such as remapping the density of an inner diameter of a scan compared to the outer diameter because of the difference in speed at the inner and outer diameters of the scans.

In one embodiment, instead of an illumination line 130 , an area of illumination 131 (illustrated with broken lines in FIG. 2 ) may be used. In this embodiment, instead of using an optical element, such as cylindrical lens 114 , that focuses the light in one direction and collimates the light in an orthogonal direction to form a line, another optical element that illuminates an area of the substrate would be used, e.g., collimating the light or projecting a field aperture on the substrate. The area of illumination 131 may be scanned across the surface of the substrate in a manner similar to that described above, as indicated by arrows 133 in FIG. 2 .

In an XY coordinate system, the substrate 102 or the optics in the metrology device 100 (or a combination of the two) is moved to scan the illumination line 130 back and forth over the surface of the substrate 102 , as illustrated by lines 132 , such that the end of the illumination line 130 is adjacent to or slightly overlaps the area scanned in the previous pass. Alternatively, the illumination line 130 may be raster scanned over the surface of the substrate 102 . The scanning of the substrate 102 continues until the desired area of the substrate 102 has been scanned.

As illustrated in FIG. 1 , the light from light source 112 is incident on the substrate 102 at an oblique incidence. In the spectrometer path 110 , a reflective or refractive optical element, illustrated as lens 116 , which may be convex lens or lens system, focuses the illumination line that is spectrally reflected (0 th order) by substrate 102 into a spot that is received by detector 118 . The detector 118 may be, e.g., a spectrophotometer, with a lens or lens system 142 (or reflective elements) and a diffraction grating 144 that spreads the light from the sample 102 into its spectral components onto a detector array 146 , such as a CCD or CMOS imaging sensor, the operation of which is well known.

If desired, the metrology unit 100 may use a spectral ellipsometer in the spectrometer path 110 , including, e.g., a polarizer 115 and analyzer 117 , illustrated with broken lines in FIG. 1 . Thus, the metrology unit 100 may use both spectral information as well as polarimetry information. The polarizer 115 and analyzer 117 may be located before or after the lens elements 114 and 116 respectively and may be rotated to modulate the polarization intensity. It should be understood that other optical elements, such as a compensator, may be used.

FIG. 3A illustrates one embodiment of the array 146 in detector 118 . As illustrated in FIG. 3A , array 146 may be one dimensional. The reflected light from illumination line 130 is focused to a spot on the detector 118 and the spectral information, illustrated by band 152 , is collected along the vertical axis, labeled λ, of the array 146 . As the illumination line 130 is scanned across the surface of the substrate 102 , the data from the array is periodically loaded or transferred and the array is refreshed. The refresh rate of the array 146 should be adequate to record the information from illumination line 130 without a loss of data, i.e., the array 146 should be refreshed when (or before) the illumination line 130 has traveled a full width's distance.

FIG. 3B illustrates a two dimensional array 146 ′ that is another embodiment of the array in detector 118 . The spectral information 152 from the reflected light from illumination line 130 is collected along the first column of pixels in the array 146 ′ along the vertical axis labeled λ. As the illumination line 130 is scanned across the substrate 102 , the collected spectral information 152 is transferred from one column to the next along the horizontal axis labeled λ and then read out of the array 146 ′. Alternatively, the spectral information 152 maybe scanned across the array 146 ′ as the illumination line 130 is scanned across the substrate 102 and the data from the array 146 ′ is transferred after each pass.

FIG. 3C illustrates another two dimensional array 146 ″ that is another embodiment of the array in detector 118 . In the present embodiment, instead of receiving a spot from lens 116 in FIG. 1 , the detector 118 receives a line, e.g., lens 116 may be a cylindrical lens or lens system. The detector 118 spreads the light from the sample 102 into its spectral components onto a two dimensional detector array 146 ″, which is illustrated shown in FIG. 3C . The length of the illumination line 130 is focused along one dimension of the array 146 ″, e.g., the vertical axis label Y, while the spectral information is received along the horizontal axis, labeled λ. As discussed in reference to FIG. 3A , the data from the array 146 ″ is periodically loaded or transferred and the array is refreshed while the illumination line 130 is scanned across the surface of the substrate 102 . The refresh rate of the array 146 ″ should be adequate to record the information from illumination line 130 without a loss of data, i.e., the array 146 ″ should be refreshed when (or before) the illumination line 130 has traveled a full width's distance.

As illustrated in FIG. 1 , the darkfield path 120 of the metrology device 100 may be positioned approximately normal to the surface of the substrate 102 . A reflective or refractive optical element, illustrated as lens or lens system 122 , focuses the darkfield image of the illumination line on the substrate 102 onto an image detector 124 , such as a CCD or CMOS imaging sensor, the operation of which is well known. As is well known in the art, “darkfield” refers to a metrology technique in which a specimen, e.g., a defect, is illuminated from the side and light scattered from the specimen is detected, resulting in a dark background and a luminescent specimen.

Because the sample 102 may include diffractive elements that are not defects, the incident angle of the light from light source 112 may be varied, as indicated by arrow 113 in FIG. 1 , to ensure that light diffracted from non-defect diffractive elements does not coincide with the darkfield path 120 and therefore interfere with the darkfield image. As the angle of incidence is varied, the position of the detector 118 may be correspondingly varied, as indicated by arrow 119 , so that detector 118 receives the spectrally reflected (0 th order). If desired, the angle of incidence may be held stationary and the position of the darkfield path 120 , e.g., the image detector 124 and lens 122 , may be varied to receive the darkfield image, as illustrated by broken arrow 121 . Alternatively, a filter 125 may be used at the pupil plane of the imaging path of the detector 124 to block the first order light from the periodic structure on the substrate 102 .

FIG. 4A illustrates a top view of one embodiment of the image detector 124 used in the darkfield path 120 of the metrology device 100 . The image detector 124 may be a one dimensional array that receives the length of the darkfield image from illumination line 130 along the axis labeled Y. The darkfield image 154 of the illumination line 130 is illustrated superimposed over the detector 124 with a defect 156 illustrated as a white spot in the darkfield image 152 . Similar to array 146 , described in reference to FIG. 3A , the data from the array 124 is periodically loaded or transferred and the array is refreshed as the illumination line 130 is scanned. The refresh rate of the array 124 should be adequate to record the information from illumination line 130 without a loss of data, i.e., the array 124 should be refreshed when (or before) the illumination line 130 has traveled a full width's distance.

In another embodiment, illustrated in FIG. 4B , a two dimensional image array 124 ′ is used. The darkfield image 154 from the illumination line may be imaged on the first column of pixels in the array 124 ′, e.g., the length of the darkfield image 154 is imaged along the vertical axis labeled Y. As the illumination line 130 is scanned across the substrate 102 , the collected darkfield image 154 is transferred from one column to the next along the horizontal axis labeled X and then read out of the array 124 ′. Alternatively, darkfield image 154 maybe scanned across the array 124 ′ as the illumination line 130 is scanned across the substrate 102 and the data from the array 124 ′ is transferred after each pass.

In another embodiment, illustrated in FIG. 4C , a two dimensional image array 124 ″ is used and the darkfield image is projected over an area of the array 124 ″. It should be understood that while the area of the array 124 ″ that is covered by the darkfield is shown to be square in FIG. 4C , the darkfield image may have other geometric shapes. In this embodiment, instead of a using a line of illumination on the surface of the substrate, an area of illumination 131 is used, such as that shown in FIG. 2 . The data from the array 124 ″ is periodically loaded or transferred and the array is refreshed while the area of illumination is scanned across the surface of the substrate 102 . The refresh rate of the array 124 ″ should be adequate to record the information from area of illumination without a loss of data, i.e., the array 124 ″ should be refreshed when (or before) the area of illumination has traveled a full width's distance. Alternatively, the image of the area of illumination is transferred from one column to the next synchronized with the scanning of the area of illumination across the substrate.

As shown in FIG. 1 , the detectors 124 and 146 are coupled to a processor 160 , which receives and processes the data from the detectors. Using the data from detector 124 in the darkfield path 120 , the processor 160 can determine the size, geometry, and intensity of a defect on the substrate. Processing darkfield data to detect the presence of a defect is well known in the art. Additionally, when a defect is detected, the processor 160 can use the spectral data from detector 146 to identify the type and material composition of defect. For example, when the darkfield image indicates the presence of a defect, but the spectral information from the detector 146 indicates that the optical parameters, such as refractive index or absorption coefficient, does not change, the defect may be a scratch or other structural defect. On the other hand, if the spectral information indicates that one or more optical parameters change, the defect may be contamination, such as a foreign particle or fingerprint. The processor 160 may determine the composition of the defect using the spectral data using well known spectrometer and/or ellipsometer techniques. The composition of the foreign particle can then be used to identify the source of the defect.

The processor 160 may include a data structures and software code for automatically implementing the one or more acts described in this detailed description, which can be implemented by one of ordinary skill in the art in light of the present disclosure and stored on a computer readable storage medium. The storage medium may be any device or medium that can store code and/or data for use by a computer system such as computer 160 in FIG. 1 . This includes, but is not limited to, magnetic and optical storage devices such as disk drives, magnetic tape, compact discs, and DVDs (digital versatile discs or digital video discs), and computer instruction signals embodied in a transmission medium (with or without a carrier wave upon which the signals are modulated.

Although the present invention is illustrated in connection with specific embodiments for instructional purposes, the present invention is not limited thereto. Various adaptations and modifications may be made without departing from the scope of the invention. Therefore, the spirit and scope of the appended claims should not be limited to the foregoing description.