KR100971173B1 - Method and system for fast on-line electro-optical detection of wafer defects - Google Patents

Abstract

Short optical pulse illumination from a pulsed laser repeatedly, the field of view of an electro-optical camera system with microscopic optics, and the optical-detector plane optically forming on the focal plane of the optical imaging system and formed of six detector ensembles. A method and system for rapidly electro-optical detection of wafer die defects on-line by imaging of a wafer moving over a focal plane assembly. Each detector ensemble includes an array of four two-dimensional CCD matrix photo-detectors, each of which produces a large matrix of electronic images of 2 million pixels, formed simultaneously from other CCD matrix detectors. Images can be processed in parallel using conventional image processing techniques to compare the imaged field of view with another field of view as a reference to find differences in corresponding pixels. The difference in pixels suggests a wafer die defect. Continuously moving wafers are illuminated with a short duration laser pulse, for example 10 nanoseconds, which is considerably shorter than the image pixel stay time, so there is little pixel corruption while the wafer is in motion. The laser pulses have sufficient energy and brightness to give the necessary illumination to the inspection field of view required to form an image of the inspection wafer die. Parallel processing of the entire focal plane assembly comprising 24 CCD matrix photo-detectors provides a total pixel processing speed of close to 1.5 gigapixels per second. In addition, the entire wafer inspection system operates at substantially 100% efficiency, with a laser pulse rate of 30 pulses per second synchronized with a frame rate of 30 frames per second of each CCD matrix photo-detector, and the wafer being separated between consecutive fields of view. Moves at linear speed so that is covered in 1/30 seconds.

Description

On-line electro-optic detection methods and systems for detecting wafer defects quickly {Method and system for fast on-line electro-optical detection of wafer defects}

1 is a flow chart of a preferred embodiment of an on-line electro-optic detection method for rapidly detecting wafer defects in accordance with the present invention.

2 is a schematic representation of a preferred embodiment of an on-line electro-optic detection system for rapidly detecting wafer defects in accordance with the present invention.

3A is a schematic top view of a CCD matrix photo-detector in accordance with the present invention.

3B is a schematic side view of a CCD matrix photo-detector according to the present invention.

4A is a schematic enlarged side view of a detector ensemble including CCD matrix photo-detectors and prisms in accordance with the present invention.

4B is a schematic enlarged side view of another detector ensemble comprising CCD matrix photo-detectors and prisms in accordance with the present invention.

FIG. 4C is an enlarged schematic view of the surface of a glass prism as part of the detector ensemble shown in FIGS. 4A and 4B, including areas of high reflectivity coating, in accordance with the present invention.

FIG. 4D is an enlarged front optical view of the detector ensemble shown in FIGS. 4A-4C, featuring a plurality of CCD matrix photo-detectors in accordance with the present invention, showing the appearance of an optically continuous surface of the photo-detectors.

5A is a schematic enlarged view of a focal plane assembly comprising beam splitting prisms and detector ensembles in accordance with the present invention.

FIG. 5B shows a continuous surface of photo-detectors comprising several CCD matrix photo-detectors in accordance with the present invention and optically formed in the focal plane formed by detector ensembles of the focal plane assembly.

FIG. 6 is an enlarged view of an image acquisition process constituting wafer dies in accordance with the present invention, where each wafer die is inspected sequentially by imaging a plurality of strip views one at a time; do.

The present invention relates to methods and systems for electro-optically detecting fabrication defects in patterned semiconductor structures, such as semiconductor wafers constituting integrated circuit dies or chips, which defects are inherently random. In particular, the present invention relates to short light pulse illumination from pulsed lasers, a field of view of electro-optic camera systems with microscopic optics, and to a focal plane of the optical imaging system. In a method and system for rapidly on-line electro-optical detection of wafer defects by optically forming the surface of the detectors and imaging the wafer moving over a focal plane assembly (FPA) formed from several detector ensembles. It is about. Here, each detector ensemble comprises several arrays of two-dimensional matrix photo-detectors, each of which produces an electronic image in a matrix of image elements (pixels). In order to find out the difference in the corresponding pixels, images made simultaneously in different matrix photo-detectors are compared in parallel using typical image processing techniques to compare the imaged field of view with another field of reference. The differences are hinted at the presence of wafer die defects.

From here on, the term 'wafer' is considered to be characterized by the individual patterned structures, generally known as 'semiconductor wafer dice', 'wafer dice' or 'wafer chips', so called. Current semiconductor technology involves physically separating a single wafer into identical dies for the manufacture of integrated circuit chips, for example, each die is a discrete integrated circuit chip having a predetermined pattern, such as a memory chip or a microprocessor chip. do. The type of chip produced from a given die is independent of the method or system of the present invention.

From here, the field of view generally refers to the portion or section of the wafer, and the wafer die, and in particular refers to a laser that is illuminated with a pulsed laser and imaged with electro-optic camera system inspection optics with FPA. Thus, the entire single wafer die and the entire single wafer constituting the plurality of wafer dies are examined by sequential imaging of the plurality or series of fields of view. The field of view can be thought of as an inspection system electro-optic image footprint on a wafer or wafer die. Continuously produced fields of view while the wafer is moving in one direction are called 'strips' of fields of view. The pixels are referred to as forming an image of the field of view by the electro-optic inspection system. As a reference dimension, the typical level of rectangular wafer die in a wafer is 1 cm × 1 cm or 104 μm × 104 μm.

From here, the detection of a 'wafer defect' refers to detecting whether the same patterns of wafer dies or the same patterns of visual fields are compared to detect whether there is an unevenness or difference. Current methods and systems for detecting defects on wafers are based on comparative analysis of signals from multiple adjacent wafer dies or fields of view characterized by the same pattern. It is assumed that defects that occur during fabrication of the wafer are inherently random. Thus, defect detection is based on a statistical approach, where the probability of random defects at the same location in adjacent wafer dies is very low. Therefore, defect detection is usually based on the identification of nonuniformity through the use of well-known die-to-die comparison methods. A given inspection system is programmed to inspect a pattern of wafer die or field of view, commonly called an inspection pattern. Then, to detect a pattern non-uniformity or difference that would indicate the presence of a wafer defect, it is compared to a second wafer die on the same wafer or the same pattern of view as the function of the reference pattern.

In order to confirm the presence of a defect and to identify the wafer die or field of view having the defect, a second comparison is made between the previously specified inspection pattern and the same pattern of the third wafer die or field of view. In a second comparison, the first wafer die or view is considered a reference and the third wafer die or view is considered to be inspected.

The manufacture of semiconductor wafers is quite complex and very expensive and the fine integrated circuit patterns of the semiconductor wafers are quite sensitive to defects caused by the process, external material particles and equipment malfunctions. The cost associated with the presence of wafer defects increases many times as we move from development to production. Thus, the semiconductor industry relies heavily on very fast ramp-up of wafer yield at the beginning of production, and on achieving and controlling successively high yields during mass production.

The critical dimensions of the integrated circuits on the wafers are constantly smaller and reaching 0.1 μm. Thus, advanced semiconductor wafers are susceptible to defects of smaller size than are currently detected. Current methods of monitoring wafer yield include optically inspecting wafers for defects during the process and forming a feedback loop, with appropriate variable process adjustments between the fabrication process and the fabricated wafer. In order to detect smaller sized defects, optical inspection systems need to have high resolution by scanning wafers using small pixel sizes. Scanning a given size wafer at a small pixel size increases wafer inspection time per wafer and reduces wafer yield to reduce the statistical sample size of the inspection wafers. Conversely, attempting to increase wafer inspection yield using the pixel size of current optical systems as is reduces the detection efficiency, ie resolution, of wafer defects.                         

In addition to reducing the critical dimensions of the wafers, the semiconductor industry is on the way from 8 inch wafer fabrication to 12 inch wafer fabrication. Larger 12 inch wafers have twice as much surface area as 8 inch wafers. In a given inspection system, the inspection time per sheet of a 12 inch wafer is expected to take twice as long as that of an 8 inch wafer. 12 inch wafers manufacturing is more expensive than 8 inch wafers manufacturing. In particular, 12 inch wafers raw material cost is higher than 8 inch wafers. One consequence of wafer size conversion is that the cost effective productivity of future wafer fabrication will depend heavily on the speed and yield increase of wafer inspection systems.

Automated wafer inspection systems are used for quality control and quality assurance of wafer fabrication processes, equipment and products. These systems are used for monitoring purposes and are not directly involved in the manufacturing process. For any element of the overall manufacturing system, it is important that the system of wafer inspection methods and instruments be cost effective relative to the total cost of semiconductor wafers.

Thus, there is a need to inspect semiconductor wafers to detect wafer die defects of wafers characterized by small critical dimensions and large sizes, with higher yields and cost-effectively than are currently possible.

Automated optical wafer inspection systems were introduced in the 1980s when computer platforms involved in electro-optical, software and image processing enabled a shift from manual to automated wafer inspection. However, the inspection speed, and consequently the wafer yield of these systems, is technically limited, making it impossible to keep up with increasing stringent production conditions, for example, fabricating integrated circuit chips of small critical dimensions with large wafer sizes. Current wafer inspection systems generally use a continuous illumination and two-dimensional scan of the wafer to form a two-dimensional image of the wafer compartment. This is a relatively slow process, and as a result, the amount of on-line inspection data obtained during the manufacturing process is small, resulting in only a relatively small statistical sample of inspection wafers, which takes a relatively long time to find wafer manufacturing problems. Slow systems of on-line defect detection may result in significant wafer scrap, low wafer production yield, and turn-around-time of the pin-pointing manufacturing process and / or equipment resulting in totally long wafer defects. cause turn-around-time).

A notable limitation in current wafer defect detection methods and systems relates to the registration of pixel positions in wafer images. Before a wafer defect is detected by a standard technique that compares the difference in pixel intensities in the image of the target or inspection wafer die with the pixel intensities in the image of the reference wafer die, the wafer in the images of the inspection wafer die and the reference wafer die Pixel locations need to be registered. The wafer speed under the wafer inspection camera system is not constant because there is generally mechanical inaccuracy during the movement of the wafer mounted on the translation stage. As a result, the image pixel positions in the field of view of the detector may be distorted and different from that originally programmed. Thus, the best-fit two-dimensional translational pixel registration correction is performed.

Prior art wafer defect detection methods and systems include continuous wafer illumination and two-dimensional wafer scanning using a laser flying spot scanner such as that disclosed in US Pat. No. 5,699,447 to Alumot et al. Or US Pat. It features a combination of acquiring two-dimensional images from one-dimensional wafer scanning using a linear array of photodetectors as shown in 4,247,203, which requires registration correction of all pixels or all pixel lines. These methods limit system speed, that is, test yield and require substantial electronic hardware. Moreover, they result in residual misregistration, because there is no correct calibration procedure for every pixel in the image. Residual misregistration significantly reduces system fault detection sensitivity. In wafer inspection methods and systems in which all focal plane assembly pixels within a given field of view can be generated simultaneously and considered as a unit, there is no need to register image pixels within the field of view of the focal plane assembly.

Thus, only a single two-dimensional alignment correction between the inspected field of view and the equivalent area of the reference field of view is needed, and the single alignment correction will be accurate over the entire focal plane assembly field of view. This process can result in negligible residual misregistration and can improve defect detection sensitivity. Accordingly, what is needed is a method and system for electro-optical on-line detection of wafer die defects that can minimize residual misregistration of image pixel locations.

Mechanisms for inspecting wafers are disclosed in US Pat. Nos. 4,247,203 and 4,347,001 to Levy et al. The mechanism disclosed in these patents detects defects and errors in the photomask by comparing the patterns of adjacent dies on the photomask while tracking the difference. Using two different imaging channels, an equivalent field of view of each die is imaged simultaneously and the image is electronic digitized by two linear diode array photo-detectors containing 512 pixels each. A two-dimensional image of the field of view selected on each die is created by mechanically moving the subject during inspection in one direction and electronically scanning the array elements in orthogonal directions. During the detector exposure time the photomask cannot be moved a distance of more than one pixel or the image is corrupted. Thus, the time to scan and inspect the photomask is very long. Since the photomask is constantly moving while two-dimensional images are being generated, it is necessary for the photomask to move without jitter and acceleration. This movement limitation requires a very large and accurate air bearing stage to move the photomask, which is expensive.

In addition, a wafer inspection instrument such as Levy et al. Can detect a defect of 2.5 mu m with a 95% probability on the photomask. Currently, critical dimensions of semiconductor integrated circuits are approaching 0.1 [mu] m, which means that inspection pixels should be of similar size. The inspection speed increases in inverse proportion to the square of the pixel size, so Levy et al. Will be slowed more than 100 times. In addition, it becomes impossible to realize a moving stage that can meet the required mechanical accuracy.

Wafer inspection is also based on solid-state cameras using two-dimensional CCD matrix photo-detectors, such as those described by IBM scientists Byron E. Dom et al. 'Machine Vision and Applications' [(1998) 1: 205-221]. It has been realized using a single imaging and detection channel. A wafer inspection system named P300 has been described for inspecting patterned wafers with repetitive patterns of cells in each die, such as semiconductor wafers for memory devices. The system captures an image field of 480x512 pixels. This image processing algorithm assumes a known horizontal cell period R in the image and analyzes each pixel in the image by comparing two pixels spaced one pattern period R in both horizontal directions. Comparing the same cells within a single image is called cell-to-cell comparison. The tested pixel is compared with both periodic neighbors to ensure ambiguity that may exist when compared to only one pixel. The system is capable of capturing two-dimensional images of the subject under test at the same time, but is very slow to inspect the entire wafer.

Because millions of image fields are required to image the entire wafer, and because the system uses continuous illumination, as used with standard microscopes, the wafer must move from field to field under the inspection camera and the image must not be compromised. It should be stopped during the exposure. To reach the other field, the mechanical moving stage carrying the wafer must be accelerated and decelerated to stop at the new position. Each movement takes a relatively long time and therefore takes several hours to inspect the wafer.

Thus, semiconductor wafers can be rapidly used for wafer die defects in wafers characterized by small critical dimensions and large sizes, while providing a high level of image resolution of wafer dies with higher yields than are currently available. It would be useful if a method and system for on-line inspection were needed and had such.

Preferred embodiments of the present invention are optically forming the face of the photo-detector on the focal plane of the optical imaging system, and the short optical pulse illumination from the pulsed laser, the field of view of an electro-optic camera system with microscopic optics, For example, a method and system for rapidly electro-optical detection of wafer die defects on-line by imaging of a wafer moving over a focal plane assembly formed of six detector ensembles, for example. Each detector ensemble includes several, for example, an array of four two-dimensional charge coupled device (CCD) matrix photo-detectors, each two-dimensional CCD matrix photo-detector in a large matrix, for example two million pixels. Images produced simultaneously from different CCD matrix detectors using conventional image processing techniques to compare the imaged field of view with another field of view as a reference to find differences in corresponding pixels. Can be processed in parallel. Difference 0 in the pixels suggests a wafer die defect.

In particular, the method and system of the present invention preferably make it possible to capture high pixel density, wide-field images of the wafer die without on-the-fly wafer standstill. High accuracy of wafer transfer speed is not required, and a relatively simple and inexpensive mechanical stage can be used to move the wafer. Continuously moving wafers are illuminated with laser pulses of short duration, for example 10 nanoseconds, which are considerably shorter than the image pixel dwell time so that there is little pixel corruption while the wafer is in motion. During the time interval of the laser pulses, the wafer die image moves less than 1/10 of the pixels. The laser pulses have sufficient energy and brightness to give the necessary illumination to the inspection field of view required to form an image of the inspection wafer die. In addition, as a result of the method and system characterized by optically coupling the separated CCD matrix photo-detectors with the detector ensemble and focal plane assembly, 24 matrix photo-detectors with an imaging capacity of the entire array, for example 48 megapixels The processing time of the array of chips is equivalent to the processing time of a single CCD matrix photo-detector with a 1/30 second level. This is because the processing of all photo-detectors goes in parallel. As a result, the parallel processing of the entire focal plane assembly comprising 24 CCD matrix photo-detectors provides a total pixel processing data rate close to 1.5 gigapixels per second. In addition, the entire wafer inspection system operates at substantially 100% efficiency, with a laser pulse rate of 30 pulses per second synchronized with a frame rate of 30 frames per second of each CCD matrix photo-detector, and the wafer being separated between consecutive fields of view. Moves at a linear speed so that it is covered in 1/30 seconds.

The methods and systems of the present invention preferably provide significant improvements over the methods and systems currently used for electro-optical inspection and detection of wafer defects in the semiconductor wafer manufacturing industry. Improvements include providing high resolution, wide field wafer die images at very high wafer inspection yields, and requiring less electronic and system hardware. In addition, with the direct effect of using an array of several CCD matrix photo-detectors to obtain a high pixel density image of a wafer die illuminated with a single light pulse, the method and system of the present invention provides a method for pixel in wafer die images. Improves defect detection sensitivity by preventing incorrect registration of positions. Such methods and systems for detecting wafer defects result in feedback control of wafer fabrication processes, which is faster, more efficient and cost effective.

In this way, according to a preferred embodiment of the present invention, a method is provided for electro-optical inspection of defects in patterned semiconductor wafer dies. The method includes (a) moving the patterned wafer along an inspection path; (b) repeatedly providing a pulsed laser illumination source; (c) sequentially illuminating each of the plurality of fields of view in each of the plurality of wafer dies using the pulsed laser illumination source; (d) sequentially obtaining an image of each of the sequentially illuminated plurality of fields of view in each of the plurality of wafer dies, using an electro-optical camera comprising at least two two-dimensional matrix photo-detectors Wherein the at least two two-dimensional matrix photo-detectors simultaneously acquire images of each of the sequentially illuminated multiple fields of view in each of the plurality of wafer dies; (e) a wafer by comparing the sequentially acquired images of each of the sequentially illuminated multiple fields of view in each of the plurality of wafer dies using a die-to-die comparison method Detecting a defect.

According to further features described further in the preferred embodiments, the repeatedly pulsed laser illumination source is a Q switched Nd: YAG laser or excimer laser.

According to other features further described in preferred embodiments, the electro-optical camera is placed in the laser beam light path of the pulsed laser illumination source repeatedly and functions as a second harmonic generating crystal. And further comprising a nonlinear optical crystal, wherein the nonlinear optical crystal has wavelengths of laser beam light generated by the repeatedly pulsed laser. The electro-optical camera may further comprise additional nonlinear optical crystals to divide the original laser wavelength by three or four to produce a third or fourth harmonic.

According to a preferred embodiment of the present invention, a system is provided for electro-optical inspection of defects in patterned semiconductor wafer dies. The system includes (a) a mechanism for providing movement of the patterned wafer along an inspection path; (b) a repeatedly pulsed laser illumination source for illuminating the patterned wafer; (c) obtaining an image of each of the sequentially illuminated plurality of views in each of the plurality of wafer dies, wherein each of the sequentially illuminated views in each of the plurality of wafer dies An electro-optical camera comprising at least two two-dimensional matrix photo-detectors operating as a mechanism for acquiring images simultaneously; And (d) processing the sequentially acquired images of each of the plurality of illuminated views in each of the plurality of wafer dies and comparing the sequentially acquired images using a die-to-die comparison method. And an image processing mechanism for detecting wafer defects.

According to a preferred embodiment of the present invention, an electro-optical camera is provided for inspecting defects of patterned semiconductor wafer dies. The camera comprises a focal plane assembly comprising at least one detector ensemble, the detector ensemble operating at least as a mechanism for simultaneously acquiring images of each of a plurality of illuminated fields of view in each of the plurality of wafer dies. It comprises an array of two two-dimensional matrix photo-detectors.

Realizing the method and system of the present invention includes completing or performing the steps or tasks manually, automatically, or a combination thereof. Furthermore, according to the actual equipment and equipment of a given wafer inspection system, several steps of the present invention may be realized in hardware or software, or a combination thereof, for an operating system of certain firmware. For example, as hardware, the indicated steps of the present invention can be realized in a chip or a circuit. As software, the indicated steps of the present invention can be realized as a plurality of software instructions executed by a computer using any suitable operating system. In any case, the indicated steps of the present invention may be described as being performed with a data processor such as a computing platform for performing a plurality of instructions.

The invention will now be described by way of example only with reference to the accompanying drawings.

The invention preferably includes methods and systems for rapid on-line electro-optical detection of wafer defects.                     

The rapid wafer defects on-line electro-optical detection method and system of the present invention are synchronized to an illumination system characterized by repeatedly illuminating a wafer die with short light pulses from a pulsed laser, and is a high resolution, high pixel density In order to obtain a wide field of view (wide field of view) of the wafer die, a unique combination of new imaging systems is characterized by the optically formed surface of the photo-detectors on a focal plane formed from an array of several two-dimensional matrix photo-detectors. Introduce. The duration of the laser light pulse is considerably shorter than the image pixel stay time. Pixel stay time is the time at which a point on the wafer is imaged by the detector pixel while the wafer is moving. The laser light pulse rate is synchronized with the frame rate of the respective matrix photo-detectors.

Preferred steps for operating the components of the method and system of the present invention will be better understood with reference to the embodiments described below in conjunction with the drawings. It should be noted that the examples of the invention shown herein are provided for illustrative purposes and are not limiting.

Referring now to the drawings, FIG. 1 is a flowchart of a preferred embodiment of an on-line electro-optic detection method for rapidly detecting wafer defects in accordance with the present invention. In Fig. 1, the main steps of the inventive method which are generally applicable are individually numbered and included in the frame. Additional steps indicating details of the main steps indicated in the method are indicated by the letters in parentheses. 2-6 are schematic diagrams illustrating exemplary and preferred embodiments of a system and system components for realizing an on-line electro-optic detection method for rapidly detecting wafer defects in accordance with the present invention. The system components shown in FIGS. 2-6 are referenced during the description of FIG. 1 to correspond to the method of FIG. 1. Details and specific examples of the system components of FIGS. 2-6 are provided throughout the description below. The terms and references expressed in the following description of FIG. 1 are consistent with those shown in FIGS. 2 to 6.

In step 1 of the method, a patterned semiconductor wafer 12 featuring a plurality of wafer dies 14 is placed and aligned over the continuous moving XY translation stage 16. This is shown in the system 10 of FIG. 2, which schematically illustrates a preferred embodiment of an on-line electro-optic detection system for rapidly detecting wafer defects in accordance with the present invention. The XY translation stage 16 moves the wafer 12 into a serpentine shape under the optical imaging system 18. The movement of the XY translation stage 16, and thus the movement of the wafer 12, is synchronized by the central control system 20 via the control / data link 22. At this time, the wafer 12 corresponds to an order of about 10 ⁻² of one pixel during exposure to one field of view 24 and some, for example, illumination system 26, during a CCD matrix photo-detector frame time of 33 milliseconds. In order to move by, the multi-element camera system is operated. There is no image smear or loss of image resolution.

In step 2, a multi-element electro-optic camera system is provided. The system includes (a) illumination system 26, (b) optical imaging system 18, (c) automatic focusing system 28, (d) focal plane assembly 30 and (e) central control system 20 ), Each containing its own system control / data links.                     

In substep (a) of step 2, an illumination system 26 is provided. The illumination system 26 repeatedly includes a pulsed laser 32, a laser beam expander 34, a laser beam light path 36, and a control / data link 38. This type of illumination system features a pulsed laser 32 for repeatedly generating and propagating very bright and high energy light pulses in a very short period of time, thereby enabling very fast imaging of the wide field of view 24. .

This contributes to the overall semiconductor inspection method with high yield. Monochromatic laser illumination is also preferably used to simplify the design needs of the wide field optical imaging system 18. This is because there is no chromatic aberration that requires optical correction or adjustment. The lighting system 26 is in communication with the central control system 20 via a control / data link 38.

In the system 10, the pulse rate of the pulsed laser 32, ie the number of pulses per second, is synchronized with the frame rate of the individual matrix photo-detectors array of the focal plane assembly 30. The laser pulse illumination field of view 24 of the wafer die 14 is compared to the millisecond frame time of the temporarily gated camera system focal plane assembly 30 of the matrix photo-detectors, for a duration of nanoseconds, during which the inspection wafer die ( 14 instantaneous field of view 24. Within one very short laser pulse, several, for example, a relatively large number of pixels, for example about 48 million pixels, of the focal plane assembly array 30 of 24 matrix photo-detectors are illuminated simultaneously. Virtually no relative motion between the pixels. The laser light pulse duration is significantly shorter than the image pixel stay time. Pixel stay time refers to the time at which a point on the wafer is imaged by the detector pixel while the wafer is moving.

Preferably, pulsed laser 32 is a Q-switched Nd: YAG laser, optically pumped to light emitting diodes at a pulse rate of 30 pulses per second with a pulse time interval of about 10 nanoseconds, at a wavelength of 1.06 μm. Generate a pulsed monochromatic light beam. The pulse rate of the pulsed laser illumination system 26, which is 30 pulses per second, is synchronized with the frame rate, which is 30 frames per second of the array of CCD matrix photo-detectors of the focal plane assembly 30. The laser 32 may also be an excimer laser.

Optical resolution is a linear function of the illumination wavelength. The resolution of the optical system increases with decreasing illumination wavelength. Thus, in order to increase the resolution of the optical system 18 and eventually increase the defect detection sensitivity of the inspection system 10, one or more crystals that produce at least a 'second harmonic' with nonlinear optical properties. 40 may be placed in the laser beam light path 36 of the illumination system 26. Secondary, tertiary or quaternary harmonic generation by crystal 40 or crystals 40 provides illumination at wavelengths of 0.53 μm, 0.355 μm or 0.265 μm, respectively. This doubles, triples, or quadruples the resolution of the wafer inspection system 10, respectively.

In substep (b) of step 2, an optical imaging system 18 is provided. This includes a focusing lens 42, a beam splitter 44, an objective lens 46, and a control / data link 49. This system is suitable for high magnification of the wide field of view 24 of the wafer die 14, for example 50 times ultra-high resolution synchronous imaging. The auto focusing system 28 adjusts and sets the objective lens 46 position of the optical imaging system 18 for optimal focusing of all wafer dies 14 on the wafer 12. The optical imaging system 18 is in communication with the central control system 20 via a control / data link 49. During operation of the wafer inspection system 10, the focusing lens 42 images laser light 48, which is reflected, scattered, or refracted by the wafer 12 onto the focal plane assembly 30. Represents. This imaging procedure is further described below with reference to FIG. 5A.

In substep (c) of step 2, an autofocus system 28 is provided that features sensors and control elements (not shown). Through the optical imaging system 18, the wafer 12 is automatically held in focus, and thus the wafer die 14 is also in focus.

In substep (d) of step 2, a focal plane assembly 30 is provided comprising several detector ensembles 50 (FIGS. 4 and 5). Each detector ensemble 50 includes several individual two-dimensional matrix photo-detectors, preferably at least two two-dimensional CCD matrix photo-detectors 52 (FIGS. 3A and 3B), and a focal point; It consists of face assembly electronics 54 and control / data links 56, 58, 90 that enable high capacity and ultrafast high resolution synchronized imaging of wafer die 14. Preferred structural and morphological elements and features of the focal plane assembly 30 are shown in FIGS. 3A, 3B, 4A-4D, 5A and 5B, each of which has a separate CCD matrix photo-detector 52. , A schematic enlarged view of detector ensemble 50 and focal plane assembly 30.

In FIGS. 3A and 3B, which show schematic top and side views of the two-dimensional CCD matrix photo-detector 52, respectively, the light sensitive region 60 is surrounded by the light-insensitive region 62, so that the two CCDs side by side. Physical placement of the matrix photo-detectors is prevented to form a continuous, photo-sensitive focal plane, which is preferably, but not limited to. The focal plane assembly 30 (FIGS. 2 and 5 a) comprises several, for example six detector ensembles 50 (FIGS. 4 a and 4b). Each detector ensemble 50 comprises several, for example four, two-dimensional CCD matrices for all, for example, 24 commercially available high resolution, monochrome, silicon two-dimensional CCD matrix photo-detectors 52. Photo-detectors 52. Each CCD matrix photo-detector 52 is capable of providing a very large number of image sensing picture elements or pixels, for example 1940 × 1035 (e.g. 2 million or 2 mega levels), which can provide 30 frames per second on a high definition basis. Have

Focal plane assembly 30 features six detector ensembles 50, each detector ensemble 50 featuring an array 64 of four individual CCD matrix photo-detectors 52 (FIG. 4D). It is done. The focal plane assembly 30 optically couples the 24 CCD matrix photo-detectors 52 to preferably, but not limited to, a continuous surface 66 of the photo-detectors in the focal plane (FIG. 5B). And fill a relatively wide field of view 24 of the 50x magnification microscope optical imaging system 18. This optical configuration enables illuminating the wafer die 14 with a single laser pulse and simultaneous imaging by an array 66 of 24 CCD matrix photo-detectors with a total of 48 million (48 mega) pixels. Image acquisition of the wafer die 14 for a CCD matrix photo-detector with a frame rate of 30 frames per second and an array of 48 megapixels is at a rate of about 1.5 billion (1.5 gigabytes) per second. The focal plane assembly 30 is in communication with the central control system 20 via control / data links 56, 58 (FIG. 2).

4A and 4B are schematic enlarged side views of the detector ensemble 50, which comprises two sets, for example two CCD matrix photo-detectors 52A, 52B, respectively. Preferably, each detector ensemble 50 consists of two glass prisms 68 and 70. Each prism has a right angle and a 45 degree diagonal plane. Diagonal surface 72 of prism 68 includes areas of coating, up to high reflectivity, preferably up to 100%. At least one CCD matrix photo-detector is optically coupled above each prism 68, 70. An exemplary set of two CCD matrix photo-detectors 52A coupled over prism 68 is the same as two CCD matrix photo-detectors 52B coupled over prism 70. In FIG. 4B, two CCD matrix photo-detectors 52A are shown coupled in series over prism 68. Then, two CCD matrix photo-detectors 52B are coupled in series on prism 70. The exact position of the combined CCD matrix photo-detectors 52A, 52B is one continuous straight line when all the photo-sensitive regions 60 of each of the CCD matrix photo-detectors 52A, 52B are viewed from line of sight A. FIG. It is chosen to be a strip.

4C is an enlarged view of the diagonal face 72 of the glass prism 68, where the diagonal face 72 includes an area of high reflectivity coating. 4C shows the B-B cross section of FIG. 4A. The area 74 over the diagonal surface 72 is coated with a high reflectivity coating and is opposite the surface 72 to the light-sensitive area 60 of the CCD matrix photo-detectors 52A coupled over the prism 68. ) Is arranged. Light incident on the light incident detector ensemble 50 along the line of sight A, opposite the reflecting region 74, is reflected by the reflecting region 74 and deviates 90 degrees and strikes the CCD matrix photo-detectors 52A. Light incident on the light incident detector ensemble 50 along the line of sight A, but not opposite the reflective region 74, strikes the CCD matrix photo-detectors 52B unbiased past the prisms 68, 70.

4D is an enlarged front optical view of the detector ensemble 50 shown in FIGS. 4A-4C, showing the shape of the optically continuous surface of the photo-detectors constituting the plurality of CCD matrix photo-detectors 52. FIG. 4D shows the line of sight A of FIG. 4B, illustrating what is formed by the optical means of the continuous surface 64 characterized by four photo-sensitive photo-detector regions 60. In surface 64 photo-sensitive region 60 opposite reflective region 74 is connected to CCD matrix photo-detectors 52A coupled over prism 68. A photo-sensitive region 60 other than the reflective region 74 is connected to CCD matrix photo-detectors 52B coupled over the prism 70. The photo-detectors 52A, 52B are on different surfaces or planes, and the photo-sensitive regions 60 are not continuous, but the detector ensemble 50 forms the surface 64 by optical means.

FIG. 5A is a schematic enlarged view of focal plane assembly 30 including beam splitting prisms 76, 78 and detector ensemble 50 in accordance with the present invention. In FIG. 5A, the focal plane assembly 30 includes six detector ensembles 50, two of which are designated 50A, two of 50B and two of 50C. Light 48 representing the reflected, scattered, and refracted laser illumination light exiting the wafer 12 is directed and focused into the focal plane assembly 30 by the focusing lens 42. Light 48 passes through beam splitting glass cube 76 that reflects about 33% of light 48 at 90 degrees to form imaging channel 80 and about 67% of light 48 passes. . Passing light 82 from beam splitting cube 76 reflects about 50% of light 82 at 90 degrees to form imaging channel 84 and passes about 50% of light 82 to imaging Pass through a second beam splitting cube 78 forming a channel 86.

This structure of a combination of beam splitting cubes 76, 78 forms three imaging channels 80, 86, 84, each having an equal light energy and having the light of the initial incident light beam 48. Has about 33% of the energy. An optical cube 88 is inserted into the imaging channel 80 to equalize the amount of glass in the optical path of all three channels. This allows similar image quality to be formed in all three channels. At the focus of the focusing lens 42, for each of the three imaging channels 80, 86, 84, two sets of detector ensembles 50 are placed. One set of two detector ensembles 50A is placed in imaging channel 80, one set of two detector ensembles 50B is placed in imaging channel 86, and one set of two detector ensembles 50C. Is placed in the imaging channel 84.

FIG. 5B is a schematic front view of the focal plane assembly 30 from the optical line of sight A, showing six detector ensembles 50 and twenty-four two-dimensional CCD matrix photo-detectors 52 positioned at different geometries. Using this, a continuous surface 66 of the photo-detector is shown optically formed in the focal plane.

Referring back to FIG. 2, in substep (e) of step 2, the control / data links 38, 49, 54, 56, 58 and the central control system 20 are electronically connected between other systems and system elements. It features wiring and enables proper automation and synchronization of the various steps of the method of detecting wafer defects. For example, the automated movement of the wafer 12 through the movement of the XY translation stage 16 is electronically set at a linear speed so that the wafer 12 emits at the pulsed laser 32 of the illumination system 26. It moves by one field of view 24 distance between two pulse times. The temporarily gated opening and closing, or frame rate, of the focal plane assembly 30 including all the CCD matrix photo-detectors 52 are synchronized with the pulse rate of the pulsed laser 32 of the illumination system 26. have.

In step 3, the camera system of step 2 is adjusted, focused, and set via the central control system 20 signals to a position on the inspected field of view 24 within the wafer die 14. The pulse rate of the pulsed laser 32 of the illumination system 26 is synchronized with the frame rate of the CCD matrix photo-detectors 52 included in the detector ensembles 50A, 50B, 50C of the focal plane assembly 30. It is. This step moves the wafer 12 and the inspection wafer die 14 at a speed that covers the inspection field of view 24 during the time interval of one frame of the CCD matrix photo-detectors 52 of the focal plane assembly 30. It is done to make it possible.

In step 4, via the central control system 20 signal, a wafer die with a short duration, e.g., 10 nanoseconds, which is several orders of magnitude smaller than the pulse rate synchronized with the frame time of the camera system CCD matrix photo-detectors 52 By concentrating the laser pulse on (14), instantaneous illumination of the inspected field of view 24 of the inspection wafer die 14 of step 3 is obtained. In a 10 nanosecond laser pulse, about 48 million pixels of focal plane assembly 30 with 24 CCD matrix photo-detectors 52 are illuminated simultaneously, so there is no relative motion between the pixels. During short laser pulses, there is virtually no wafer movement during wafer exposure time. This is because the laser pulse duration is much shorter than the pixel stay time, the time at which a point on the wafer is imaged to the detector pixel while the wafer is moving. Thus, there is virtually no image corruption that degrades image resolution. This is unique only when wafer inspection methods and systems feature continuous illumination of the wafer.

In step 5, a focal plane assembly 30, optically connected to detector ensembles 50A, 50B, 50C, characterized by 24 two-dimensional CCD matrix photo-detectors 52, via a central control system 20 signal. ), The illuminated field of view 24 illuminated in step 4 is imaged by the optical imaging system 18.

In step 6, the focal plane assembly 30 is optically formed via the central control system 20 signal, which optically forms a continuous surface of at least two CCD matrix photo-detectors 52, preferably but not limited to. And synchronizing the opening of the temporarily gated CCD matrix photo-detectors 52, the digital image of step 5 consisting of approximately 58 million pixels of the inspection field 24 of the wafer die 14 (not shown). ) Is obtained. During the frame time interval of each activated CCD matrix photo-detector 52, the wafer 12 and eventually the wafer die 14 move equally through one field of view through the XY translation stage 16. This corresponds to a pixel stay time that is relatively larger than the laser pulse time interval, about 10 ⁻² of one pixel while exposed to the array of CCD matrix photo-detectors 52 of the focal plane assembly 30 (FIG. 5B). Only a fraction of the wafers are moved in order to prevent image corruption or loss of image resolution. In sub-step (a), the acquired digital image is acquired by a series of image processing channels 90 configured in parallel as an image grabber 92 to thereby form part of the image processing system 100 (FIG. 2). It is stored in the constituent image memory buffer 94.

In step 7, steps 3 through 6 are repeated sequentially to obtain an image of the next fields of view within the same inspection wafer die 14, thereby removing the strip of fields of view until it includes the first equivalent field of view of the nearest neighbor wafer die in the strip. Form, and the neighboring wafer die serves as a reference. This automated sequential image processing is clearly shown in FIG. 6, which is an enlarged view of the image acquisition process constituting the wafer dies, where each wafer die is configured by imaging multiple strip fields of view one at a time. , They are checked sequentially. In FIG. 6, image acquisition of the first field of view 24A at the first inspection wafer die 14A is followed by image acquisition of the second field of view 24B at the same first inspection wafer die 14A. In synchronization with the sinuous movement of the wafer 12, image acquisition of successive fields of view, in turn, is performed until the image for the first field of view 24J in the second inspection wafer die 14B is obtained. Proceeds past all of (14A). This process results in the formation of successive strips 110 of imaged wafer dies 14 until the entire wafer 12 is imaged in its entirety.

In step 8, an image processing system is used to process digital image data of each field of view that is equally located within the nearest neighboring wafer die serving as a reference and each field of inspection wafer die. Referring to FIG. 2, the image processing system 100 includes an image processing channel 90, an image buffer 94, a defect detection unit 96, a defect file configured in parallel to acquire an image with the image acquirer 92. 98, and control / data link 102. Image data obtained with focal plane assembly 30 featuring 24 two-dimensional CCD matrix photo-detectors 52 are processed in parallel, and each of the two two-dimensional CCD matrix photo-detectors 52 In parallel with the other CCD matrix photo-detectors 52 of the focal plane assembly 30, it individually communicates with the image acquirer 92 via 24 separate image processing channels 90. Instead of processing image data using a single series of channels 48 megapixels at a CCD frame rate acquisition rate of 30 per second, this results in a single channel with a very high throughput of 1.5 gigapixels per second, Each of the 24 separate image processing channels 90 having about 2 megapixels of image data and acquired at a rate of 30 per second can be used to process at a medium speed of 60 megapixels per second. In this structure, a total image processing speed of 1.5 gigapixels per second is obtained using significantly slower discrete channels, which makes it easier to realize in a wafer defect detection system 10 with commercially available hardware. This configuration of processing the acquired image data in parallel significantly contributes to the high yield of the wafer inspection method of the present invention. The image processing system 100 is in communication with the central control system 20 via a control / data link 102.

Step 8 includes: sub-step (a) performing image alignment of the inspected field of view and the reference field of view, sub-step (b) identifying the presence of potential wafer defects, and sub-step (c) storing the comparison data in a defect file. And sub-step (d) deleting unnecessary image data in the first field of view of the first inspection wafer die.

In substep (a) of step 8, prior to confirming the presence of potential wafer defects in the inspection wafer die, an image alignment is performed between the image of each inspected visual field and the corresponding visual field serving as a reference. Due to minor mechanical inaccuracies during the movement of the XY translation stage 16, the speed of the wafer 12 below the camera optical imaging system 18 is not constant. As a result, image pixel positions within multiple fields of view of CCD matrix detectors may differ from those originally programmed upon intersystem synchronization. Thus, two-dimensional translational image alignment correction of the examined field of view and the reference field of view is performed. More complex rotation registration corrections may also be performed but are ignored for the standard realization of the method and system of the present invention. This process of aligning the images of the fields of view before comparing the images and detecting defects is shown in FIG. 6 for exemplary strips 110 of the same fields of view. The pixel positions in the image of the first field of view 24A of the first inspection wafer die 14A and the pixel positions in the first field of view 24J of the nearest neighboring wafer die 14B that are equally located are from the image buffer 94. It is extracted and subjected to image alignment correction. In this process, the first field of view 24J of the nearest neighboring wafer die 14B serves as a reference to the equivalent field of view 24A of the first inspection wafer die 14A.                     

Prior art methods and systems for detecting wafer defects, such as those mentioned above, are characterized by a combination of continuous wafer illumination and the scanning in one or two directions to obtain a two-dimensional image. Requires registration correction of pixels or all pixel lines. This limits system speed, that is, inspection yield, and increases the need for electronic hardware and the total system cost. Moreover, not all pixels in the image can be corrected correctly, resulting in residual misregistration. Residual misregistration significantly reduces system fault detection sensitivity. In contrast, for a preferred embodiment of the method and system of the present invention, all focal plane assembly CCD matrix detection pixels in a given field of view of the focal plane assembly are considered as one unit and are generated simultaneously by a single laser pulse. Thus, there is no need for pixel registration within the focal plane assembly field of view, and simple alignment correction between a small local area in the inspection field of view and an equivalent area in the reference field of view is accurate over the entire focal plane field of view. Thus, in the present invention, residual misregistration is negligible, and enables improved defect detection sensitivity.

In substep (b) of step 8, following image alignment correction, the pixel intensities of each field of view that are equal to the field of view of the inspection wafer die starting from the first and the closest field of view pixel intensity of the wafer starting starting from the first By comparing the differences between them, we identify the presence of potential wafer defects. In this defect identification step, a standard algorithm of defect detection is used, which is based on comparative analysis of pixel intensities of an image obtained in the same fields of view of adjacent neighboring dies of the same pattern. Defect detection is based on a statistical approach where the probability of random defects at the same location of adjacent wafer dies is very low. An exemplary standard algorithm for finding non-uniformity of pixel intensities of different images is a three-die comparison. The entire wafer inspection system is programmed to inspect patterns of wafer die or field of view, commonly referred to as inspection patterns. Then, a comparison is made with a pattern assumed to be the same pattern on the same wafer serving as a reference pattern. To eliminate ambiguities that may exist when an inspection pattern is compared to just one pattern, the pattern under inspection is also compared with the same located pattern of other adjacent wafer dies. In order to maintain symmetry in the second comparison, the test pattern is referenced.

This image comparison procedure performed by the defect detection unit 96 (FIG. 2) is shown in FIG. Each pixel intensity in the image of the first field of view 24A of the first inspection wafer die 14A is compared with the pixel intensity in the image of the equally located first field of view 24J of the adjacent neighboring wafer die 14B.

In substep (c), the wafer die (in accordance with predetermined comparison criteria, such as a predetermined difference or non-uniformity of the threshold level) is further processed by the determining step (step 10) of identifying or clearing the presence and location of the defect. The difference or non-uniformity in the intensity of the corresponding two pixels in the first fields of view 24A, 24J, which are equally located in the wafer die 14B as reference 14A, is stored in the wafer defect file 98, respectively.

In substep (d), unnecessary image data of the first field of view 24A of the first inspection wafer die 14A is deleted from the image buffer 94. The first field of view 24A of the first inspection wafer die 14A is stored as comparison data of the first fields of view 24A, 24J, which are equally located in each of the first inspection wafer die 14A and the second inspection wafer die 14B, is stored. Image data is not required to image the next wafer die 14 within the wafer 12.

In step 9, steps 7 and 8 are repeated to obtain successive fields of view within the second inspection wafer die 14B until it includes image processing of the first field of view 24N of the third inspection wafer die 14C. Steps 7 and 8 are performed in parallel. The image acquisition step in step 7 is performed for each field of view in strip 110, and the comparison and image processing of the preceding field of view in strip 110 is performed according to step 8.

Step 10 is a determination and confirmation step performed by the defect detection unit 96, starting with the field of view 24J of the wafer die 14B and detecting wafer defects in each field of view initially processed according to step 8 Determine and confirm. In order to confirm or clear the presence of a defect located in the field of view 24J of the wafer die 14B, the first fields of view 24A, which are equally positioned in each of the first inspection wafer die 14A and the second inspection wafer die 14B, The presence of the difference or non-uniformity between 24J) involves the next comparison between the first field of view 24J, 24N, which is equally located in each of the second inspection wafer die 14B and the third inspection wafer die 14C.

Confirmed wafer defect information including the location of the wafer defect identified in substep (a) of step 10 is suitably stored in the defect file 98 for the possibility of use in feedback control in the wafer fabrication process.                     

In step 11, steps 7 to 10 are sequentially repeated for inspection of each field of view in the field of view strips 110 in the same wafer. In Fig. 6, for example, the wafer field of view 24K in the wafer die 14B becomes the next inspection field to be subjected to the image processing in steps 7-10.

Starting from the field of view 24K in the wafer die 14B, the images of consecutive fields of view in the second wafer die 14B are compared with images of fields of view at the same location in the wafer dies 14A and 14C. The field of view 24K in the wafer die 14B is compared with the field of view 24B at the same location in the wafer die 14A, with the field of view 24B serving as a reference. The field of view 24K in the wafer die 14B is compared with the field of view 24P at the same position in the wafer die 14C, with the field of view 24K serving as a reference. In this case, each image of each set of successive fields of view in strip 110 is compared once with an equivalent field of view on the wafer die preceding it in the strip, and once with an equivalent field of view on the wafer die subsequent to it in the strip. . Each of the compared visual fields serves as the reference visual field once in the comparison and once as the inspection visual field. Until the defects are checked for all the wafer dies 14 of the wafer 12 in synchronization with the sinuous movement of the wafer 12, the image of successive fields of view of the successive wafer dies is in turn the entire wafer ( 12) selection, illumination, imaging, acquisition, and processing from the wafer die to the wafer die.

As mentioned above, although the present invention has been described in detail with reference to preferred embodiments, the present invention is not limited to the above embodiments, and various modifications can be made by those skilled in the art within the technical idea of the present invention. Is obvious.

 In accordance with the present invention, semiconductor wafers can be inspected to detect wafer die defects of wafers with smaller critical dimensions and larger sizes with higher yields than are currently possible and cost effective. In addition, it is possible to improve the defect detection sensitivity by minimizing the residual false registration of the image pixel position.

Continuously moving wafers are illuminated with laser pulses of short duration of, for example, 10 nanoseconds, which are considerably shorter than the image pixel stay time, resulting in little pixel corruption during wafer motion. Full focal plane assembly parallel processing comprising 24 CCD matrix photo-detectors provides a total pixel processing speed of close to 1.5 gigapixels per second. In addition, the entire wafer inspection system operates at substantially 100% efficiency, with a laser pulse rate of 30 pulses per second synchronized with a frame rate of 30 frames per second of each CCD matrix photo-detector, and the wafers being separated between consecutive fields of view. Moves at linear speed so that is covered in 1/30 seconds.

Claims (127)

At least two two-dimensional matrix lights having surfaces that sequentially acquire images of the subjects with at least one illuminated field of view and define focal planes while subjects are illuminated and move along a path of travel; An at least two two-dimensional matrix photo-detectors each comprising: an electro-optical camera focusing on at least partially different areas of the subject and providing a two-dimensional digital output; A transporter providing relative motion between the electro-optical camera and the subjects moving along the path of travel; A pulsed laser illuminator for illuminating the subjects with short light pulses as the subjects move along the movement path; And A digital image processing subsystem including a defect detection function, The duration of the short light pulse is shorter than the image pixel stay time, The speed of the short light pulse is synchronized with the frame rate of each of the fields of view. The system of claim 1, wherein the digital image processing subsystem comprises: An image buffer used by the defect detection function during operation of the defect detection function; And Further comprising multiple processing channels, At least a first multiple processing channel of the multiple processing channels is operative to process the digital output of at least one of the at least two two-dimensional matrix photo-detectors, wherein at least a second multiple processing channel of the multiple processing channels is And process the digital output of at least another one of the at least two two-dimensional matrix photo-detectors in parallel to the at least first multiple processing channel of the multiple processing channels. 3. Inspection system according to claim 1 or 2, wherein the inspection system further comprises a synchronizer for synchronizing the operation of the pulsed laser illuminator with the operation of the carrier. The method according to claim 1 or 2, And the at least two two-dimensional matrix photo-detectors comprise a plurality of two-dimensional matrix photo-detectors disposed in different physical planes and optically coupled. The method according to claim 1 or 2, The focal plane comprises a continuous surface of two-dimensional matrix photo-detectors generated by optically combining a plurality of two-dimensional matrix photo-detectors disposed in different physical planes. The digital image processing subsystem of claim 2, wherein the digital image processing subsystem comprises a plurality of image processing channels, Each of said at least two two-dimensional matrix photo-detectors is processed in a different image processing channel of said plurality of image processing channels. The method of claim 2 or 6, wherein the image buffer comprises a plurality of buffering regions, The digital image processing subsystem is configured to output the digital output of at least one of the at least two two-dimensional matrix photo-detectors in parallel with obtaining the digital output of at least another of the at least two two-dimensional matrix photo-detectors. And an image grabber operable to store the at least one of the at least one of the at least two two-dimensional matrix photo-detectors and the at least another one of the digital outputs in different ones of the plurality of buffering regions of the image buffer. Inspection system comprising a). 7. The digital image processing subsystem of claim 2 or 6, wherein the digital image processing subsystem outputs a defect file, the at least one of the at least two two-dimensional matrix photo-detectors and the at least two two-dimensional matrix photo-detectors. At least one of said processed digital outputs is written to said defect file. 3. The speed of the motion provided by the carrier and the pulse length of the pulsed laser illuminator are selected such that smearing of the image is less than one tenth of the pixels. Inspection system. The inspection system according to claim 1 or 2, wherein the speed of movement provided by the carrier and the pulse length of the pulsed laser illuminator are selected such that the spots of the image have an order of one hundredth of the pixel. . The method according to claim 1 or 2, wherein the inspection system, Further comprising an optical subsystem comprising at least one beam splitter and at least one highly reflective optical member, The at least one beam splitter and the at least one highly reflective optical member work together to split an image of the focal plane between the at least two two dimensional matrix photo-detectors. The method of claim 11, The at least one highly reflective optical member comprises at least two highly reflective optical members, wherein the image is split into portions by the beam splitter and subsequently the respective portions of the at least two highly reflective optical members Split by one, And the at least two highly reflective members and the at least two two dimensional matrix photo-detectors are arranged to provide a generally continuous optical surface. The method of claim 11, The at least two two-dimensional matrix photo-detectors comprise at least four detector ensembles, each of the ensembles comprising at least one two-dimensional matrix photo-detector, wherein the image of the focal plane is Split between at least four detector ensembles such that each of the at least four two-dimensional matrix photo-detectors in the at least four detector ensembles receives about 33% of the illumination and the focal plane is the at least four detector ensembles A continuous surface on adjacent ones of the detectors included in the device. 14. The inspection system of claim 13, wherein said at least four detector ensembles comprise at least six detector ensembles. The inspection system of claim 1, wherein the at least two two dimensional matrix photo-detectors comprise a two dimensional array of two dimensional matrix photo-detectors. The inspection system of claim 15, wherein the two-dimensional arrays of two-dimensional matrix photo-detectors comprise at least three detectors in at least one of the vertical and horizontal axes of the two-dimensional array. Imaging an object using an electro-optical camera, wherein the electro-optic camera sequentially obtains images of the subjects with at least one illuminated field of view and has at least two two-dimensional matrices having surfaces defining a focal plane Imaging the subject comprising photo-detectors, the at least two two-dimensional matrix photo-detectors each focusing on at least partially different areas of the subject and providing a two-dimensional digital output; Using a carrier to provide relative motion between the electro-optic camera and the subjects along a path of travel; Pulsed laser illumination of at least one illuminated fields of each of the subjects by pulsed laser illuminators while the subjects are moving along the movement path; Obtaining an image of at least one illuminated field of view of each of the subjects with a short light pulse and providing a two-dimensional output, wherein the duration of the short light pulse is short compared to the image pixel stay time and the speed of the short light pulse Providing the two-dimensional output synchronized with each frame rate of the fields of view; And Using a digital image processing subsystem to process digital outputs of the electro-optical camera, And said digital image processing subsystem includes a defect detection function. The method of claim 17, Using the digital image processing subsystem includes using a digital image processing subsystem further comprising an image buffer and multiple processing channels used by the defect detection function during operation of the defect detection function, At least a first multiple processing channel of the multiple processing channels is operative to process the digital output of at least one of the two-dimensional matrix photo-detectors, and at least a second multiple processing channel of the multiple processing channels is configured to process the multiple processing. And process the digital output of at least another one of the second dimensional matrix photo-detectors in parallel with the at least first multiple processing channel of channels. The method of claim 18, The image buffer includes a plurality of buffering regions, and using the digital image processing subsystem, Obtaining the digital output of at least one of the two-dimensional matrix photo-detectors in parallel with grabbing the digital output of the at least another one of the two-dimensional matrix photo-detectors; And Storing outputs of one or more of the two-dimensional matrix photo-detectors in different ones of the plurality of buffering regions of the image buffer. The method of claim 17 or 18, Selecting the speed of movement provided by the carrier and the pulse length of the pulsed laser illuminator so that the image blob is less than 1/10 of the pixel. The method of claim 17 or 18, Further comprising using an optical subsystem comprising at least one beam splitter and at least one highly reflective optical member to split the image of the focal plane between any of the two-dimensional matrix photo-detectors Inspection method characterized in that. An electro-optical camera comprising two-dimensional matrix photo-detectors operative to image the subjects while the subjects are illuminated and moving along a path of movement, wherein the two-dimensional matrix photo-detectors sequentially obtain images of the subjects with at least one illuminated field of view; A carrier for providing relative motion between the electro-optical camera and the subjects moving along the path of travel; A pulsed laser illuminator for illuminating the subjects with short light pulses as the subjects move along the movement path; And A digital image processing subsystem including defect detection functionality, The duration of the short light pulse is shorter than the image pixel stay time, The speed of the short light pulse is synchronized with the frame rate of each of the fields of view. Pulsed laser illumination of at least one illuminated field of view of each of the subjects by pulsed laser illuminators while the subjects are moving along the movement path; Obtaining an image of at least one illuminated field of view of each of the subjects with a short light pulse and providing a two-dimensional output, wherein the duration of the short light pulse is short compared to the image pixel stay time and the speed of the short light pulse Providing the two-dimensional output synchronized with each frame rate of the fields of view; And Using a digital image processing subsystem to process the digital outputs of the camera, And said digital image processing subsystem includes a defect detection function. delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete

Images