Wafer alignment system using parallel imaging detection

The invention generally relates to optical exposure systems for semiconductor wafers and more particularly to an improved system and method that reduces the amount of time needed to align the wafer in the optical exposure system.

With the everspeed of wafer exposure systems, the time spent on alignment is becoming the limiting factor in the total throughput of steppers or stepsystems. Typical time spent for a single wafer alignment can reach as much as 10 seconds, which represents a large portion of the total time that is needed for single wafer exposure. For example, if an exposure process is able to produce 60 wafers per hour, each wafer would take one minute to expose. Similarly, if the exposure process is able to produce 90 wafers per hour, each wafer is exposed in 40 seconds. Therefore, the 10 seconds required to align each wafer becomes a significant factor.

One conventional system attempts to reduce the time required to align the wafer using a “twin stage” system. In such a twin stage system, a single machine includes two wafer stages. While one said wafer stages is being aligned, the other said wafer stages can expose a wafer. This directly eliminates the alignment time; however, such devices are substantially larger (30% in size and weight than the single stage devices and are more expensive than single stage devices. An additional drawback of twin stage systems is that if one of the stages experiences a defect, the machine must stop production, which dramatically increases inefficiency for the single defect.

Therefore, if the savings in alignment time can be achieved using single stage systems, such benefits are produced with fewer drawbacks when compared to twin stage systems. The invention described below reduces the alignment time without having to resort to twin stage systems.

To overcome the problems discussed above, the invention provides a system for aligning a wafer in an exposure apparatus. The inventive system includes a holder adapted to hold a wafer (the wafer includes alignment marks), a coarse alignment system, and a fine alignment system having a higher precision than the coarse alignment system. The fine alignment system includes optical detectors. Each of the optical detectors is positioned to detect a corresponding alignment mark on the wafer. An alignment processor is connected to and controls the optical detectors and the holder. The optical detectors are controlled by the alignment processor to simultaneously detect the alignment marks in parallel operations. Further, the alignment processor simultaneously processes signals from the optical detectors in parallel operations.

In addition, in a second embodiment, the optical detectors can be controlled by the alignment processor to simultaneously detect the alignment marks multiple times while the holder moves the wafer up and/or down in a direction perpendicular to the surface of the wafer, which allows the alignment processor to pick the bestimage for each detector from a sequentially acquired series of images through focus alignment position determination.

The alignment marks can comprise twomarks. In one embodiment, the number of alignment marks is equal to the number of optical detectors, such that each optical detector is dedicated to a specific alignment mark. If the number of alignment marks are greater than the number of optical detectors, at least some of the optical detectors detect multiple alignment marks. If there are more alignment marks than optical detectors, the invention would detect as many alignment marks as possible in a first detection process and then detect the remaining alignment marks that were missed during the first detection process in a second detection process. Each of these detection processes performs parallel simultaneous detections of multiple alignment marks using the multiple optical detectors. The alignment marks are preferably positioned so that the invention only needs to shift the wafer a single time to allow the remaining alignment marks to be detected during the subsequent detection process.

Bases of the optical detectors are positioned around a periphery of the wafer in an approximate equally spaced distribution. Alternatively, the invention can include common shafts supporting the optical detectors that surround the wafer. In this embodiment, groups of the optical detectors are connected to each of the common shafts.

The invention also provides a method of aligning the wafer in the exposure apparatus that first mounts a wafer in a holder, then performs the coarse alignment and the fine alignment. The fine alignment has a higher precision than the coarse alignment. The process of performing said fine alignment simultaneously detects multiple alignment marks using a plurality of optical detectors. Each of the optical detectors is positioned to detect a corresponding alignment mark. The invention also simultaneously processes the signals from the optical detectors using the alignment processor. When the holder is moved up and/or down, multiple observations are made during the moving process, again to allow the alignment processor to processes multiple focal regions within the topography of the wafer. This process allows the invention to align the wafer with the optical exposure device based upon signals received from the optical detectors.

The invention also provides a method of calibrating the system used to align the wafer in the optical exposure apparatus. In this methodology, the invention first locates a primary alignment mark on a wafer using a preoptical detector and calibrates the position of the alignment mark and then uses this mark to sequentially calibrate the positions of the remaining optical detectors through the stepping of the wafer stage. Placement of the primary alignment mark and the remaining alignment marks is restricted in a design of the wafer to prevent spatial conflict in the array of optical detectors. This calibration process is generally performed for the first wafer in a batch of wafers that includes a new pattern of alignment marks. Once the system is calibrated, the remaining wafers are simply aligned with parallel detection with the calibrated system.

By simultaneously detecting multiple alignment marks during the finealignment process, the invention substantially reduces the time required to complete the finealignment. This time savings is important because the time required to align the wafer is a large portion of the overall exposure processing time associated with the wafer. Further, the invention is also useful with twin stage systems because, as such twin stage systems move toward the use of faster exposure systems, the alignment process will become the bottleneck of the processing. Therefore, by reducing the time necessary to complete the finealignment, the invention improves many different forms of processing.

Another benefit produced by the invention results from performing multiple image detections as the wafer holder is moved up and/or down. With this feature, the invention solves the wafer topography problem, which changes the current per-mark autofocusing into a parallel process.

The invention will be better understood from the following detailed description of preferred embodiments of the invention with reference to the drawings, in which:

FIG. 1 is a schematic diagram of a wafer alignment system according to the invention;

FIG. 2 is a schematic diagram of a wafer alignment system according to the invention;

FIG. 3 is a schematic diagram of the invention being applied to a common pattern of wafer alignment marks;

FIG. 4 is a schematic diagram of the invention being applied to a common pattern of wafer alignment marks; and

FIG. 5 is a flow diagram illustrating a preferred method of the invention.

In order to overcome the problems mentioned above, the optical detectors are controlled by the alignment processor to simultaneously detect the alignment marks in parallel operations. By simultaneously detecting multiple alignment marks during the finealignment process, the invention substantially reduces the time required to complete the fine alignment.

As shown in FIG. 1 , the invention provides a system for aligning a wafer 126 in an exposure apparatus 150 . The inventive system includes a holder 124 such as a common wafer chuck or stage. The invention includes a conventional coarse alignment system (not illustrated) such as that disclosed in U.S. Pat. No. 6,225,012 to Nishi et al., which is fully incorporated herein by reference.

The invention also provides a finealignment system (sometimes referred to herein as a “fine alignment” system) having a higher precision than the coarse alignment system. The fine alignment system includes multiple optical units 134 . Each of the optical units 134 comprises a microscope that includes some form of optical detector 122 (such as a charge coupled device (CCD) camera, an aperture, lens, fiber optic cable, etc.), a connection arm 130 and a base 128 .

Each of the optical detectors 122 is positioned to detect a corresponding alignment mark on the wafer. In the views shown in FIGS. 1 and 2 , the alignment marks cannot be seen because they are directly below the optical detectors 122 and are blocked from view by the illustration of the optical detectors 122 . However, in FIGS. 3 and 4 , discussed below, the alignment marks 142 are represented as small white dots while the optical detectors 122 are represented as grey circles.

An alignment processor 120 is connected to and controls the optical units 134 and the holder 124 through electrical connections 132 . The alignment processor 120 can be a single computerized processor or can be multiple processors, depending upon the specific application of the invention. In addition, the wiring connections 132 between the alignment processor 120 and the optical units 134 does not need to include the specific wiring patterns/connections illustrated in FIG. 1 . To the contrary, any wiring patterns/connections that will allow the alignment processor 120 to control and receive signals from the optical units 134 (and the holder 124 ) will work acceptably with the invention. Note that the alignment processor 120 and associated wiring 132 are only illustrated in FIG. 1 and are intentionally omitted from the remaining drawings so as to not obscure the salient features of the invention being illustrated in the remaining drawings.

The optical detectors 122 are controlled by the alignment processor 120 to simultaneously detect the alignment marks 142 in parallel operations. To the contrary, conventional systems use a single fineoptical detector to detect individual alignment marks 142 one at a time in a serial fashion, moving the wafer and/or stage after each alignment mark is detected. Since the invention can detect all alignment marks 142 simultaneously by using multiple optical detectors 122 , the amount of time needed to perform the optical detection of alignment marks 142 is substantially reduced. Further, the alignment processor 120 simultaneously processes signals from the optical detectors 122 in parallel operations.

In addition, in a second embodiment, the optical detectors 122 can be controlled by the alignment processor 120 to simultaneously detect the alignment marks 142 multiple times while the holder 124 moves the wafer up and/or down in a direction perpendicular to the surface of the wafer, which allows the alignment processor 120 to processes multiple focal regions within the topography of the wafer. Some current systems perform focusing for individual images by adjusting the height of the holder (in the “Z” direction) prior to acquiring the single image of a single alignment mark. However, there are problems with such autofocusing systems. Since such systems usually require the holder to settle into a stationary position after focus scanning for each alignment mark, the holder settling time will add a substantial amount in the total alignment budget.

The topology of the wafer can vary dramatically (in a relative sense) between alignment marks 142 . If the images of the alignment marks 142 are captured simultaneously (as with the present invention) some of the marks will be in focus while others will be out of focus because of the topology height variations. To accommodate this situation, the invention provides a new autofocusing system that moves the height of the holder through an acceptable range of focus steps (preferably including all focus depths needed to allow all the optical detectors to acquire sharp images). During the processing of such images by the alignment processor 120 , only the most undistorted image from each optical unit 134 is used. Even though this requires that each optical unit 134 to acquire multiple images at the different focus depths, the images can be taken much faster than with conventional systems because the invention does not require the wafer holder to settle into stationary positions for individual marks, which saves settling time. Secondly, the acquisition of many images in a digital format provides opportunities for further image analysis to more accurately determine the bestpositions.

The alignment marks 142 can comprise twomarks. In one embodiment, the number of alignment marks 142 is equal to the number of optical detectors 122 , such that each optical detector is dedicated to a specific alignment mark. If the number of alignments marks are greater than the number of optical detectors 122 , at least some of the optical detectors 122 detect multiple alignment marks 142 . Therefore, in the embodiment where the number of alignment marks 142 exceeds the number or optical detectors 122 , the invention would detect as many alignment marks 142 as possible in a first alignment process and then detect the remaining alignment marks 142 that were missed during the first detection process in a second detection process. Each of these detection processes performs parallel simultaneous detections of multiple alignment marks 142 using the multiple optical detectors 122 . The alignment marks 142 are preferably positioned so that the invention only needs to shift the wafer a single time to allow the remaining alignment marks 142 to be detected during the subsequent detection process (or multiple subsequent detection processes).

Bases of the optical detectors 122 are positioned around a periphery of the wafer in an approximate equally spaced distribution. Alternatively, as shown in FIG. 2 , the invention can include common shafts 140 supporting the optical detectors 122 that surround the wafer. In this embodiment, groups of the optical detectors 122 are connected to each of the common shafts 140 .

FIGS. 3 and 4 shown the inventive features discussed above used with common alignment mark patterns. In FIGS. 3 and 4 , the alignment marks 142 are shown as small white dots, while the position of the optical detectors 122 is shown as gray circles. The alignment mark pattern shown in FIG. 3 is intersecting linear sets of alignment marks 142 that preferably intersect at the center of the wafer. The alignment mark pattern shown in FIG. 4 places the alignment marks 142 in an approximately equal distribution along the wafer to form an approximate circle of alignment marks 142 .

As shown in FIG. 5 , the invention also provides a method of calibrating the system used to align the wafer in the optical exposure apparatus. For the first wafer in a set of wafers that has a new alignment mark pattern, the invention first locates a primary alignment mark on the wafer using a preoptical detector 500 . In item 502 , the invention sequentially locates remaining alignment marks on the wafer. In item 504 , the invention positions the movable optical detectors. Once the movable optical detectors are moved, calibration is necessasry, as shown in item 505 . The primary alignment mark is preferably located at a known position (e.g., the very center) of the wafer. Further, the primary alignment mark would be detected using an optical detector that is permanently positioned (preto detect alignment marks at that known position (the center) of the wafer. In the illustrations shown in FIGS. 1 it is presumed that the preoptical detector is positioned to detect an alignment mark at the center of the wafer. Therefore, the figures do not illustrate an optical detector that can be adjusted over the center alignment mark. Thus, the primary alignment optical detector is distinguished from the remaining optical detectors because the position of the primary alignment optical detector cannot be changed while the positions of the remaining optical detectors can (e.g., by movement of the base 128 , connection arm 130 , and/or common shafts 140 ). The location of all alignment marks are determined in the process program (the user definable software code). The system then uses the wafer stage to step this alignment mark to the rest of the detectors to calibrate their positions since all position information originates from the position of the wafer stage/holder. Placement of the primary alignment mark and the remaining alignment marks is restricted in a design of the wafer to prevent overlapping of the optical detectors 122 . This calibration process is generally performed for the first wafer in a batch of wafers that includes a new pattern of alignment marks. Once the system is calibrated, the remaining wafers are simply aligned with the calibrated system.

As also shown in FIG. 5 , the invention also provides a method of aligning the wafer in the exposure apparatus that first mounts a wafer in a holder 506 , then performs the coarse alignment 508 and the fine alignment 510 The fine alignment has a higher precision than the coarse alignment. The process of performing the fine alignment simultaneously detects multiple alignment marks using a plurality of optical detectors 510 . If the holder is moved up and/or down, multiple observations are made during the moving process 512 , again to allow the alignment processor to processes multiple focal regions within the topography of the wafer. The invention also simultaneously processes the signals from the optical detectors 514 . This process allows the invention to align the wafer with the optical exposure device based upon signals received from the optical detectors.

FIG. 6 is a flowchart illustrating the invention. In item 600 , the invention loads the process program from the database for an incoming lithography level which includes the choice of the number and location of the alignment marks (both coarse and fine). Next, in item 601 , the invention positions the fine alignment detectors to accommodate the fine alignment mark distribution. The invention then loads the first wafer onto the holder, in item 602 . In item 603 , the invention performs the regular coarse alignment (using coarse alignment detectors) on the wafer, the function of which is to center the fine alignment marks to within the capture range of the fine alignment detectors. Next, in item 604 , the invention picks the primary fine alignment mark from the first wafer, and determines its fine position using the primary detector, which is pre In item 605 , the invention steps the wafer stage sequentially to position the primary fine alignment mark under each of the remaining moveable fine detectors to calibrate their positions. In item 606 , the invention positions the first wafer for fine alignment, scans the wafer stage/holder in the vertical axis, (z-plane) from all detectors in parallel, acquires the images through focus for each alignment mark, and picks the bestimage to determine alignment positions. In item 607 , the invention exposes the said first wafer. Lastly, in item 608 , the invention loads the next wafer and starts from step 603 and skips step 604 and 605 (calibration) until the end of the batch.

By simultaneously detecting multiple alignment marks during the finealignment process, the invention substantially reduces the time required to complete the fine alignment. This time savings is important because the time required to align the wafer is a large portion of the overall exposure processing time associated with the wafer. Further, the invention is useful with twin stage systems because, as such twin stage systems reduce exposure times, the alignment process will become the bottleneck of the processing. Therefore, by reducing the time necessary to complete the finealignment, the invention improves many different forms of processing.

Another benefit produced by the invention results from performing multiple image detections as the wafer holder 124 is moved up and/or down. With this feature, the invention does not need to rely upon conventional autofocusing (which the settling time slows down process). Instead, by acquiring multiple images at different focal lengths, the invention can simply select the best image and this image will be undistorted and more accurately reflect the alignment of the wafer. The invention is that the invention saves time for alignment by operating in parallel mode and using autofocus. The invention also improves the throughput for single stage wafer exposure systems and for twinwafer exposure systems.

While the invention has been described in terms of preferred embodiments, those skilled in the art will recognize that the invention can be practiced with modification within the spirit and scope of the appended claims.