GB2112933A - Positioning apparatus - Google Patents

Abstract

A semiconductor wafer (11) carrying a fresnel zone target (17) is precisely positioned in a projection printing system by illuminating the target with coherent light from a laser (21) thereby to generate first order point images (PA and PB) above and below the wafer surface. These points are re-imaged by a re-imaging system of mirrors (27, 39) lasers (31, 35, 37, 38) and a beam splitter (33), at essentially equal magnification, onto a pair of photoelectric detectors (41, 43) thereby to generate output signals each of which includes an error component indicating the relative position of the respective projected point image (PA'', PB''). The position of the wafer is then controlled by X and Y axis micro-positioners (47 etc.) in response to the sum of these signals computed by summing and servo electronics (45) to seek a position where the error components are equal and opposite. As translational mispositioning of a wafer produces additive components in the detector signals and tilting of the wafer surface produces offsetting components, positioning errors due to tilt are reduced. <IMAGE>

Description

SPECIFICATION Positioning apparatus The present invention relates to positioning apparatus and concerns more particularly, although not exclusively, positioning apparatus for aligning semiconductor wafers for exposure in a projection imaging system.

In the manufacture of integrated circuit, it is typically necessary to produce a succession of patterns on a semiconductor wafer using photolithographic processes. These patterns must be produced in precise registration with one another. As the integrated circuit art has moved to higher density circuitry requiring finer and finer resolution, the requirements for registration have tightened correspondingly.

It has previously been proposed to facilitate registration or alignment of the successive photolithographic patterns by creating, on the surface of the semiconductor wafer, an alignment mark or indicia which is in the form of a fresnel zone target. Such a target can effectively function as a form of diffraction lens and can create images of useful brightness. One such alignment system is for example described in U.S. Patent No. 4,037,969. In the alignment system described in that patent, the fresnel zone target is illuminated with coherent light and the resultant point image formed above the wafer surface is imaged on a photodetector. The signals generated by the photodetector in turn control the relative positioning of a semiconductor and a pattern or reticle which is to be projected onto the wafer.

While this system is capable of great precision under certain circumstances, it has now been discovered that the system is highly susceptible to any tilt in the wafer surface and that there apparently exist a number of sources of localized tilting in the water surface. The problem arises since a tilt of the fresnel zone target will displace the point image in a manner which is indistinguishable from displacement caused by lateral translation of the wafer.

Briefly, apparatus according to the present invention is operative to position or align an article such as a semiconductor wafer, e.g. for exposure in a projection imaging system. The apparatus utilizes fresnel zone target indicia on the article surface which is illuminated by coherent light thereby to form, by diffraction, respective first order point images above and below the article surface. The point images so formed are re-imaged, at approximately equal magnifications, onto respective photoelectric detector means thereby to generate, from each detector, an output signal which includes an error component indicative of the relative position of the projected point image. The position of the article is then controlled in response to the sum of the signals generated by the detector means to seek a position where the error components of the signals are equal and opposite.Since the tilt produces error components which are opposing while displacement of the target produces additive signal components, positioning errors due to localized tilt are minimized.

A specific embodiment of the present invention will now be described by way of example and not by way of limitation with reference to the accompanying drawings in which:~ Fig. 1 is a diagram illustrating the operation of an alignment system constructed in accordance with the present invention; Fig. 2 is a plan view of a fresnel zone target pattern which is useful in conjunction with the apparatus of the present invention when formed on the surface of a semiconductor wafer; Fig. 3 illustrates a physical arrangement of optical components useful with a commercial semiconductor exposing apparatus; Fig. 4 illustrates the manner in which lateral displacement of the wafer displaces the pair of point images; and Fig. 5 illustrates the manner in which local slope or tilt displaces the pair of point images.

Corresponding reference characters indicate corresponding parts throughout the several views of the drawings.

With reference to the accompanying drawings, Fig. 1 illustrates positioning apparatus of the invention in conjunction with a projection imaging system for forming patterns on a semiconductor wafer. The wafer to be exposed is represented by reference character 11 and the projection system includes a main imaging lens 13 for forming an image of a reticle 1 5 on a selection portion of the wafer's surface. The projection apparatus illustrated is assumed to be of the step and repeat type in which an image of the reticle 1 5 is projected in a repeated pattern covering the surface of the wafer.

Associated with each projection site on the wafer surface is a fresnel zone target. Such a target is indicated in Fig. 1 by reference character 1 7 but it should be understood that the size of the pattern has been greatly exaggerated for the purpose of illustration and that in actual practice the size of the total diameter of the target may be in the order of 50 microns.

A suitable pattern for the target is represented in Fig. 2. As will be understood by those skilled in the art, this target would typically be formed as part of the first overall pattern formed on the surface of the wafer and so that it can be referenced during successive steps in the integrated surface manufacturing process.

Preferably the fresnel zone targets 17 are located away from the portion of the wafer surface which will form the actual integrated surfaces, e.g. in the scribe lanes between the portions which are to form the individual integrated circuit dies.

A helium neon laser 21 is provided for illuminating the fresnel zone target with coherent light. Laser 21 is preferably of the single mode type generating polarized light at a wavelength substantially different from that used for exposing the photoresist coated wafer. For example, in one embodiment of the invention the light source which exposes the photoresist through the reticle generates light at a wavelength of 436 nanometers and the laser 21 operates at a wavelength of 633 nanometers.

Light from the laser 21 is directed, through a polarizing beam splitter 23 and a quarter-wave plate 25, to a mirror 27 located at the side edge of reticle 15. Mirror 27 directs the laser light down, through the main projection lens 13, onto the surface of the semiconductor wafer 11 at the fresnel zone target 17. Diffraction of the incident coherent light by the fresnel zone target 17 generates two point images, a real image located above the wafer surface, and designated by reference character PA, and a virtual image, designated by reference character PB, which is below the surface of the semiconductor wafer. At least a portion of the light diffracted by the fresnel zone target 17 will be collected by the projection lines 13 and the point images PA and PB will be re-imaged back up in the optical column.

As will be understood by those skilled in the art, the projection lens 13 is highly corrected for the wavelength used in exposing the photoresist but its focal length for other wavelengths may be significantly different. In the particular embodiment illustrated, the focal length for the wavelength of the laser 21 is assumed to be substantially longer than for the wavelength of the exposing light and thus the re-imaged points occur outside of the projection columns. These reimaged points are designated PA' and PB' respectively.

As will be understood by those skilled in the art, the portion of the laser light scattered back from the wafer surface and returning to the mirror 27 through lens 13 will, after again passing through the quarter wave plate 25, be polarized so that most of the returned light will pass straight through the polarized beam splitter 23 rather than being turned back toward the laser 21.

A relay lens 31 directs light from the point images PA' and PB' into a beam splitter 33. The separation between PA' and lens 31 is selected to be about equal to the focal length of the lens 31 so that the light entering the beam splitter from point image PA' is essentially collimated. The portion of this light directed downward is refocused to form a point image PA". This point image is, in turn, re-imaged by a lens 37 on a photoelectric quadrant detector 41.

The focal length of lens 37 is chosen to provide an additional magnification of about 10 to 1, giving a total magnification including the main projection lens 13 of about 100 to 1. This is the effective magnification with reference to lateral movement to the fresnel zone target 17.

Rather than collimating the light obtained from point image PB', the lens 31 effectively refocuses this light to form a further point of image PB", the beam path in the meantime having been directed downwardly by a mirror 39. Point image PB" is in turn further re-imaged by a lens 38 onto a photoelectric quadrant detector 43. The focal length of lens 38 is selected, in relation to the focal length of lens 31, so that the total magnification provided to the photodetector 43 is essentially the same as that provided to the photodetector 41. Exact equality however, is not necessary since the sensitivities of the two quadrant detectors with respect to lateral movement of the wafer are preferably matched in subsequent signal processing.

The optical system is aligned so that, when the wafer 11 is properly positioned, the reformed images of the point images PA and PB will be centered on the respective photodetectors, assuming that there is no tilt to the surface of the wafer. As will be understood by those skilled in the art the quadrant detectors 41 and 43 are devices which comprise four photoelectric detectors, i.e. photodiodes, and thus detect displacement of a spot away from the center point along either of two transverse axes. The output signals from the two photoelectric quadrant detectors 41 and 43 are applied to suitable summing und servo electronics 45 which drive X-axis and Y-axis micropositioners to provide control of the lateral position of the wafer 11 relative to the projection system. The X-axis micro-positioner is indicated by reference character 47.The Y-axis micropositioner is not shown in the drawing but is oriented, as will be understood by those skilled in the art, so as to provide vernier movement n the transverse axis. If the wafer, though perfectly flat, is displaced laterally, both relayed images will be displaced in the same direction on the respect photoelectric quadrant detectors. Thus, the signals obtained from both photodetectors provide information which may be utilized as an error signal for positioning. As will be understood, the summing and servo electronics may be either analog or digital and, if digital, may be implemented as part of an overall computer control system as is becoming increasingly common in the overall systems of this sort.

In accordance with the practice of the present invention, a signal representing the sum of the two detector signals is utilized to control the positioning of the wafer. As illustrated in Fig. 4, a displacement of the wafer 11 to the right moves both of the first order point images PA and PB to the right also, i.e. both move in the same direction relative to the nominal desired position. Thus, it can be seen that it is appropriate to utilize the sum of the output signals from the two photodetectors which respond to displacements of the relayed or reformed point images. On the other hand, local tilting of the wafer surface, as illustrated in Fig. 5, moves the real image PA (above the wafer surface) to the left and moves the virtual image (below the wafer surface) to the right.It can thus be seen that, if only the real image were used for position control in a servo system, there would be introduced a translation error sufficient to bring the image into the desired position, rather than the center of the fresnel zone target as desired. However, since the virtual image below the wafer surface moves in the opposite direction in response to local slope or tilt, a servo error signal responsive to the sum of the displacements, appropriately weighted, will be essentially insensitive to small changes in local slope but will still be responsive to translation of the fresnel zone target.

If the coherent light incident on the fresnel zone target were collimated, the height of the real image PA above the wafer surrace would be equal to the depth of the virtual image PB below the surface and their lateral movements in response to local slope would be equal and opposite. This equality is not a requirement of the invention, however, since the sensitivities of the two photodetectors will typically be normalized for other reasons and the same normalization or equalization can take into account the different vertical separations of the point images from the fresnel zone target. In fact, in the embodiment illustrated, the incident coherent light will not be collimated due to the presence of the projection lens 13.While this effect could be neutralized by utilizing a compensating lens prior to introducing the laser beam into the projection column, there is essentially no reason to introduce this complication since a compensating sensitivity adjustment can be made in conjunction with normalizing the signals from the photodetectors.

Fig. 3 illustrates a practical arrangement of the presention invention as applied to a Model 4800 step and repeat microlithographic imaging system as made and sold by the Burlington Division of GCA Corporation of Bedford, Massachusetts. This commercial system utilizes a ten-to-one projection ratio between the recticle 15 and the wafer 11. In the arrangement shown in Fig. 3, radiation from the laser 21 was redirected by a pair of mirrors 61 and 63 before entering the polarizing beam splitter 23 and, after traversing the quarter-wave plate 25, an additional mirror 65 was utilized to direct the beam upwardly into the mirror 27 mounted adjacent to reticle 15. The returning radiation, after emerging from the polarizing beam splitter 23, is redirected by mirrors 67 and 69 on either side of the lens 31.

The remainder of the system was arranged essentially as shown in Fig. 1 except that the lens 38 preceded the mirror 39 rather than following it in the optical path. In this particular physical implementation, the focal lengths of the various lenses were as follows:~ Lens No. Focal length 13 49.1 mm 31 40 mum 35 10 mm 37 1.6 mm 38 10 mm

Claims (3)

Claims

1. Apparatus for positioning an article utilizing fresnel zone target indicia on the article, said apparatus comprising: electrically controllable positioning means for adjusting the position of the article; first and second photoelectric detector means; means for projecting coherent light onto said indicia thereby to form respective point images above and below the surface of the article; means utilizing light scattered from said indicia for projecting each of said point images on a respective one of said detector means; thereby to generate, from each detector, an output signal including an error component which indicates the position of the respective projected point image relative to the detector; means responsive to a sum of signals generated by said detector means for controlling said positioning means to seek a position where the error components of said signals are equal and opposite.

2. Apparatus for aligning an article for exposure in a projection imaging system utilizing fresnel zone target indicia on the surface of the article, said apparatus comprising: first and second electrically controllable positioning means for adjusting the position of the article in the projection imaging system along respective transverse axes; first and second quadrant detectors; means for projecting coherent light onto said indicia thereby to form by diffraction respective first order point images above and below the surface of the article; means utilizing a light diffracted from said indicia for projecting each of said point images on a respective one of said quadrant detectors at approximately equal magnifications; thereby to generate, from each quadrant detector, output signals including error components which indicate the position of the respective projected point image relative to the detector; means responsive to a sum of signals generated by both of said detector means for controlling each of said positioning means to seek a position along the respective axis where the error components of said signals are equal and opposite.

3. Apparatus for positioning an article substantially as hereinbefore described with reference to the accompanying drawings.