The subject invention is an alignment system for use with a lithography system used to expose resist coated semiconductor wafers, and more particularly, is an optical alignment system for use with an X-ray lithography system for very precisely aligning alignment marks on the mask and on the wafer being exposed.

In the past, optical and X-ray lithography systems have been described for exposing a resist covered semicon¬ ductor wafer as one step in the process of fabrication semiconductor chips. For example, see United States Patent 4,870,668, granted September 26, 1989 in the name of Robert D. Frankel et al, entitled "Gap Sensing/Adjustment Appara¬ tus And Method For A Lithography Machine" and assigned to the assignee hereof, which describes an X-ray lithography system and United States Patent 4,444,492, granted April 24, 1984 in the name of Martin E. Lee and entitled, "Appar¬ atus for projecting a Series Of Images Onto Dies Of A Semi¬ conductor Wafer", which describes such an optical lithogra¬ phy system. In the process of fabricating semiconductor chips, the wafer, on which the chips are fabricated, is fabricated layer by layer to arrive at a final product. Modern chips may have many different such layers and during the fabrica¬ tion of each layer, that layer must be properly aligned with respect to the preceding layers, or the resulting chips will be inoperative. With the more recent introduc¬ tion of X-ray lithography systems, the features on the chips are designed to be one half micron, or less. Such small feature sizes require that the alignment system aligning each layer with the preceding layers be precise to about one tenth of a micron, or less. For example, when a particular feature is placed on a new layer, that feature may have to be precisely placed above a corresponding fea¬ ture of a preceding layer where the two features are to be connected together. If the two features are, for example.

lead lines of one half, or less micron, width, a slight layer to layer misregistration could result in poor or no contact.

In the past, layer to layer alignment has been done by placing indexing marks on both the mask and the wafer surface and then using the two alignment marks to align the mask and wafer. The alignment marks typically constitute a cross, or plus sign (+), and optical means are used to de¬ tect when the two marks are in vertical alignment. Each leg of these marks may have a relatively large width, in the order of two to three microns. In some instances, an open center cross is used as the alignment mark for one of the mask or wafer. In other instances, a technique of dark field scatter is used. The latter technique, for example, is utilized in the aforementioned Lee patent 4,444,492. However, in order for the dark field scatter technique to work, it is necessary for the light to illuminate both the mask mark and the wafer mark and this means that the light must travel through the mask to reach the wafer. With op- tical lithography systems such as shown in the aforemen¬ tioned Lee patent 4,444,492, providing the light through the mask is not a problem because the mask constitutes black marks on clear glass.

However, for X-ray lithography machines, such as exemplified by the aforementioned Frankel et al patent 4,870,668, the X-ray mask is fabricated of a material gen¬ erally opaque to light. For example, the X-ray mask may be a pattern of gold on a thin membrane of boron doped sili¬ con. Attempts to avoid the prior art limitation, of not being able to view the wafer mark through the mask, have included by placing the marks outside of the printing field and by providing openings through the mask membrane. Both of these prior art techniques have inherent limitations. Placing the marks outside the printing field below the mask requires either a large blind step to the mask area, re¬ sulting in inherent errors, or a relative alignment, re¬ sulting in a substantial amount of wasted real estate on

the wafer. Placing holes through the mask membrane compli¬ cates an already difficult mask fabrication process and can lead to damage of the thin (typically, one micron) mem¬ brane. Other techniques of alignment of the masks and wa¬ fer in X-ray lithography machines include United States Patent 4,238,685 in the name of Peter Tischer and entitled, "Arrangement For The Production Of Electronic Semiconductor Components"; United States Patent 4,335,313 in the name of Justin L. Kreuzer et al and entitled, "Method And Apparatus For Aligning An Opaque Mask With An Integrated Circuit Wa¬ fer"; United States Patent 4,385,434 in the name of Theo¬ dore F. Zehnpfennig et al and entitled, "Alignment System"; United States Patent 4,472,824 in the name of W. Derek Buckley and entitled, "Apparatus For Effecting Alignment And Spacing Control Of A Mask And Wafer For Use In X-Ray Lithography"; United States Patent 4,513,203 in the name of Harald Bohlen et al and entitled, "Mask And System For Mu¬ tually Aligning Objects In Ray Exposure Systems"; United States Patents 4,514,858 and 4,516,253, both in the name of W. Thomas Novak and both entitled, "Lithography System"; United States Patent 4,539,695 in the name of Carlo La Fi- andra and entitled, "X-Ray Lithography System"; United States Patent 4,595,295 in the name of Janusz S. Wilczynski and entitled "Alignment System For Lithographic Proximity Printing"; United States Patent 4,613,981 in the name of Graham J. Siddall et al and entitled, "Method and Apparatus For Lithographic Rotate and Repeat Processing"; United States Patent 4,641,921 in the name of Heinz Beneking and entitled, "Optical Adjusting Process"; United States Patent 4,698,834 in the name of Ronnie Northrup et al and enti¬ tled, "X-Ray Mask Membrane Deflection Compensator And Meth¬ od"; United States Patent 4,777,641 in the name of Akira Inagake et al and entitled, "Method and Apparatus For Alignment"; and United States Patent 4,856,037 in the name

of Karl-Heinz Mueller et al and entitled, "Arrangement For Exposing Semiconductor Wafers By Means Of A Synchrotron Radiation In Lithographic Equipment".

None of the aforementioned prior art systems solve all of the problems necessary to obtain an alignment system useful with a modern X-ray lithography machine. What is needed is an alignment system which fits in the limited space of an X-ray lithography machine and which can align reference marks placed on both the mask and wafer without significant lateral displacement between the marks and which is accurate to one tenth of a micron, or less.

In accordance with one aspect of this invention, there is provided a lithography system, including a mask, a wafer spaced from the mask and means for moving the wafer relative to the mask. The mask and wafer each have an alignment mark thereon and the system further includes alignment means for detecting when the mask mark and the wafer mark are properly positioned relative to one another. The alignment means is characterized by means for directing light from a skewed direction along a plurality of discrete paths towards the marks such that the edge of only one of the alignment marks on each of the mask and wafer is illu¬ minated by the light from each path. The system is further characterized by means for imaging the illuminated mask mark edges and the illuminated wafer mark edges and means for determining the relative alignment of the mask mark and the wafer mark in response to the imaged mask mark and wafer mark edges.

One preferred embodiment of the subject invention is hereafter described with specific reference being made to the following Figures, in which;

Figure 1 illustrates one type of X-ray lithography device with which the alignment system of the subject in¬ vention may be used; Figure 2 illustrates the manner in which the wafer sections and mask are aligned;

Figure 3 illustrates one type of alignment mark which may be used in the subject alignment system;

Figure 4 illustrates schematically illustrates the optical system of the alignment system of the subject in- vention;

Figure 5 illustrates the manner in which light is reflected from the edges of the alignment mark shown in Figure 3;

Figure 6 illustrates the pattern on the video de- tector used with the alignment system of the subject inven¬ tion;

Figure 7 illustrates the electrical signals derived from scanning the video detector used with the alignment system of the subject invention; Figure 8 illustrates a side view of one preferred embodiment of the alignment system of the subject inven¬ tion;

Figure 9 illustrates a top view of the preferred embodiment of the alignment system of the subject inven- tion; and

Figure 10 illustrates a more detailed view of the objective lens and a portion of the light directing means of the alignment system of the subject invention.

Referring now to Figure 1, an X-ray lithography machine 10 is generally shown and includes a high peak pow¬ er and repetition rate pulsed laser 12, an X-ray source 14, and a wafer handling mechanism 16 as the principal compo¬ nents of machine 10. Laser 12 provides a laser beam 18 which is directed by mirrors 20 and 22 and focused by lens 24 and finally directed by mirror 26 towards X-ray source 14. Laser beam 18 provided by laser 12 should be powerful enough to cause an X-ray emitting plasma to be formed when beam 18 is focused on a metal target, such as target 28.

X-ray source 14 includes a cassette target 28 con- tained within a low pressure helium chamber 30. Laser beam

18 is provided through laser beam port 32, which is a part of low pressure helium chamber 30, and focused by lens 24

to a focal point on target 28. The intensity of laser beam 18 is sufficient to create an X-ray emitting plasma at the focal point on target 28 and the plasma, in turn, emits X- rays 34 into an exposure column 36. The broad concept of exposure column 36 is shown and described in United States Patent 4,484,339 in the name of Phillip J. Mallozzi et al.

X-ray mask 38, which is positioned at the bottom of exposure column 36, is of a type which blocks certain X- rays 34 and permits the remaining X-rays 34 to pass there- through, thereby providing a defined pattern of X-rays from the exposure column 36. Mask 38 may be of a type described in Published European Patent Application No. 244,246, enti¬ tled "X-Ray Mask and Structure" in the name of Irving Plot- nik, which application is owned by the assignee hereof. Briefly, mask 38 includes a support ring 39 and a membrane 37 and gold or tungsten patterning is placed over membrane 37 forming the image to be fabricated into a chip.

Wafer handling mechanism 16 is of the type in which a semiconductor wafer 40 is held in a chuck 42 and moved in steps so that one exposure section at a time of wafer 40 is positioned beneath exposure column 36 to be exposed by the pattern of X-rays 34. In fabricating a semiconductor chip, many various series of operations on each exposure section of wafer 40 are performed. Many of these operations include exposing the pattern formed on the membrane 37 of mask 38 on each exposure section followed by further processing that exposed pattern in a predetermined fashion. Except for the first layer, each pattern is exposed on top of a previ¬ ously exposed and processed layer of wafer 40. It is ex- tremely important that each new exposure section be prop¬ erly aligned with respect to the position of the corres¬ ponding exposure section in the prior layer in order that the fabricated chip operates properly.

In order to preform proper movement and alignment of each section of wafer 40, wafer handling mechanism 16 must be capable of moving wafer 40 with six degrees of freedom. These six degrees of freedom include three linear

directions, that is, the X direction, the Y direction, and the Z direction, and three angular directions. In Figure 1, the X direction may be right to left, the Y direction may be in and out of the paper and the Z direction may be up and down. Mechanism 16 rides on a flat granite base 44 and includes a Y stage 46 and an X stage 48. Affixed above and to X stage 48 is a substage 50 upon which the control mechanisms for Z movement, and the rotational movements are mounted. Substage 50 is affixed to X stage 48 and moves therewith in the X and Y directions.

Y stage 46 moves over granite base 44 in a direc¬ tion determined by a guide 52 affixed to granite base 44. Guide 52 determines the Y direction and Y stage 46 moves back and forth along guide 52 in that determined direction. Y stage 46 has a guide 54 affixed thereto defining the X direction and X stage 48 moves along guide 54 over air bearings 55. Substage 50 is positioned above X stage 48 by three legs 56-1, 56-2 and 56-3, each of which includes a stepper motor assembly 58-1 through 58-3 and a sensor 60-1 through 60-3. The three motor assemblies 58-1 through 58-3 are independently controllable to raise or lower one trian¬ gular corner of substage 50 in the Z direction. Each assem¬ bly 58-1 through 58-3 shaft is precisely guided by means of flexure pivots (not shown) . By moving each of the three motor assemblies 58-1 through 58-3 together, small incre¬ mental movements in the Z direction can be obtained. By moving one or two of the motor assemblies 58-1 through 58-3 individually, the tip and tilt degrees of freedom referred to earlier can be obtained. Motor assemblies 58-1 through 58-3 are rigidly af¬ fixed in a vertical direction from the X stage 48. This causes a slight lateral movement of substage 50, relative to the fixed drive shafts of the motor assemblies 58-1 through 58-3, to occur during the tipping or tilting of substage 50. To permit this lateral movement, the shaft connecting motor assembly 58-1 to substage 50 includes a ball and socket coupling 62 designed to prevent any lateral

movement. However, the coupling of the shaft from motor assembly 58-2 to substage 50 is a ball and Vee grove cou¬ pling 64, designed to permit lateral movement in either the x- or y-direction only, and the coupling of the shaft from motor assembly 58-3 to substage 50 is a ball and flat cou¬ pling 66 (seen in Figure 3), designed to allow lateral movement in either of the X or Y directions.

Y stage 46 moves along Y guide 52 by conventional driver means (not shown) in discrete steps and X stage 48 moves along guide 54 by similar conventional drive means (not shown) . X stage 48 is held above Y stage 46 by four air bearings 55 extending down from the corners thereof. Air bearings 55 glide along granite base 44 in the direc¬ tion defined by guide 54. By spacing the four air bearings 55 as far apart as possible, the slight variations in the plane of granite base 44 cause relatively small and repeat- able tip and tilt changes in the wafer 40 held by chuck 42. Adjustments to the tip and tilt position of wafer 40 can then be made by the apparatus and techniques hereafter de- scribed.

Chuck 42 includes a chuck plate 68 for holding wa¬ fer 40 and beneath chuck plate 68 is wafer lifting mecha¬ nism 70. The Z movement is controlled by motor assemblies 58-1 through 58-3, which have incremental steps of 0.12 microns and a maximum range of 400 microns. In order to properly position wafer 40 with respect to mask 38, a chuck sensor 74 and X-ray chamber sensor 76 are provided, each of which is coupled to the respective chuck 42 and exposure column 36 by respective brackets 78 and 80. The detailed construction and operation of sensor 74 and 76 are de¬ scribed in the aforementioned U.S. Patent 4,870,668 to Frankel et al.

The precise X and Y positions of wafer handling mechanism 16 can be determined by a interferometer device. Such a device is well known in the art and includes a light transmitting device (not shown) , one of the Y mirror 82 or the X mirror 84, and a light receiving device (not shown).

The interferometer device measures the accumulated Doppler shift between transmitted and received light beam and thereby determines very precise displacements. For exam¬ ple, the interferometer devices may measure relative X and Y distances in lithography system 10 in the order of 0.02 micron. The Y mirror 82 arid X mirror 84 associated with the two interferometer devices are mounted on substage plate 50 very close to the same height as wafer 40. By mounting mir¬ rors 82 and 84 in this position, any corresponding X and Y lateral movement at the wafer 40 exposure plane due to the tip and tilt adjustment by fine Z motor assemblies 58-1 through 58-3 is monitored by the interferometer.

In addition, there is associated with chuck plate 68 means (not shown) for rotating the plate 68 to provide a rotational degree of freedom movement. Such means are well known in the art and are described in United States Patent 4,444,492 in the name of Martin E. Lee.

Finally, in Figure 1, the alignment system 86 is shown as being movable into and out of exposure column 16 to permit each section of wafer 40 to be aligned with re¬ spect to mask 38. Such alignment must be very precise, preferably within a fraction of the minimum line width be¬ ing printed. In other words, the position at which a new feature is to be placed on a particular section of wafer 40 using the lithography machine 10 should be aligned within one fifth of the design rule limits of the previous layers fabricated on wafer 40. Thus, if the previous layer con¬ tained lines of one half of a micron, the alignment would need to be precise to within one tenth of a micron and smaller line widths would require sub-tenth of a micron alignment precision. After the alignment procedures using alignment system 86 are performed, alignment system 86 must be moved out of the exposure column path of X-rays 34. This is accomplished by moving bracket 88 in the direction of the arrow, such that bracket 88 is moved upward and outward by guide pins 92, which are affixed to bracket 88 being

guided upward and outward by slotted guide bracket 90. Alignment system 86 may be stored in tunnel 94, attached to exposure column 36, when not in use.

Referring now to Figure 2, there is shown the rela- tive positioning of wafer 40 and membrane 37 of mask 38, as wafer 40 is moved relative to mask 38. As seen in Figure 2, membrane 37 is typically circular in shape and it is approximately twenty-five millimeters in diameter. The patterned portion 96 of membrane 37 is generally square or rectangular in shaped and may include any pattern desired within defined size limits. Immediately adjacent to the actual patterned portion 96 is a mask alignment mark 98. Mark 98 may be in any position along the top or bottom edge of pattern portion 96 and is used for the dual purpose of aligning the mask 38 and the previously exposed and pro¬ cessed layer of wafer 40 and additionally for the purpose of exposing a new alignment mark on the level of wafer 40 to then be exposed. This new mark will be used in aligning the next level. Additional marks (not shown) may also be placed along the vertical sides of the pattern portion 96 for aligning the mask during installation.

Wafer 40 is divided into various sections 100 which contain a small space 102 therebetween. Each of the sec¬ tions 100 is the same size as the patterned portion 96 of mask 38 and each of the spaces 102 is sufficiently wide to contain one or more wafer alignment marks 104. Each wafer alignment mark 104 is patterned by the previous exposure as a result of the existence of a mask alignment mark on the previous layer mask. In order to maintain the best possi- ble wafer alignment mark 104 for each layer, the marks are positioned laterally along the edge of the patterned por¬ tion for each different mask 38 pattern for each layer. In Figure 2, the wafer alignment marks 104 of preceding layers are indicated by the dashed marks 104' . As seen, the wafer alignment marks 104 and 104' are positioned laterally along each section 100, with the marks of each succeeding upward layer being in the direction of

movement of the wafer 40. As wafer 40 is moved relative to mask 38, the mask alignment mark 98 and the most recent layer wafer alignment mark 104 become aligned with one an¬ other. Such alignment is not a vertical alignment, but rather a slightly spaced apart, or offset, relative align¬ ment. This is to permit space for the next layer alignment mark 104 to be exposed onto wafer 40. Further, it permits a side by side comparison of the signals from the two spaced marks 98 and 104, as will be described hereafter. It should be noted that the amount of offset between each successive wafer alignment mark 104, 104' is quite small and only slightly greater than the size of each alignment mark 98 and 104.

Referring now to Figure 3, each of the two align- ment marks 98 and 104 are generally in the shape of a plus sign (+) and each leg 106 of the marks 98 and 104 may be approximately two to three microns in width and sixty mi¬ crons in length. When aligning the mask 38 and each sec¬ tion 100 of wafer 40, it is desirable to align the center of each leg 106 of one of the alignment mark 98 and 104 with the center of the other of the alignment marks 98 and 104. In the past, alignment has been done by lining up the marks vertically until they appear as a single mark. How¬ ever, any slight processing variance can result in a mark variation, which could result in an error in the mark, and when the design goal is to maintain layer to layer align¬ ment at sub-tenths of a micron, such slights errors cannot be tolerated.

Referring now to Figure 4, a schematic diagram is shown illustrating the optical system used to align marks 98 and 104. As noted previously, mask alignment mark 98 is on mask 38 and wafer alignment mark 104 is on wafer 40. In the present invention, an alignment occurs when the center of the two marks 98 and 104 are offset from one another by a predetermined amount. It should also be recalled that mask 38 and wafer 40 are vertically separated by approxi¬ mately twenty microns. Each of the two marks 98 and 104

are illuminated by illumination means 108, schematically shown as a lamp in Figure 4, and described in detail here¬ after with respect to Figure 10. At this point, it is suf¬ ficient to state that illuminating means 108 is of a type which provides sufficient light through the membrane 37 of mask 38 so as to illuminate marks 98 and 104 and permit the light scattered therefrom to again traverse membrane 37. An objective lens 110 is positioned directly above both alignment marks 98 and 104 and projects an image with a large focal length towards a video pickup device, such as charge coupled device (CCD) 112.

Since CCD 112 is a planar element, the image of only one of the two alignment marks 98 and 104 can be fo¬ cused at the plane of CCD 112 due to the spacing between mask 38 and wafer 40. However, in order to facilitate a comparison of the two marks, it is necessary to focus im¬ ages of both marks 98 and 104 on CCD 112. In order to ac¬ complish this, mirrors 114, 116, 118 and 120 are provided to compensate for the extended focal length required for the more distant image of mask alignment mark 98. Mirror 114 may be positioned to intersect the image from lens 110 corresponding to the image of mask alignment mark 98. The image is then reflected towards and by mirror 116 towards mirror 118 and finally reflected by mirror 120 against CCD 112. The spacing between mirrors 114, 116, 118 and 120 is selected to compensate for the difference in the image fo¬ cal length for mask alignment mark 98. Alternatively, mir¬ ror 114 may be replaced by a partially transmissive device and a portion of the entire image may be deflected and only that portion corresponding to the image of mask alignment mark 98 may be directed to CCD 112. By adjusting the trans- missive/reflective ratio of this device, differences in the scattering efficiencies between the mask and wafer align¬ ment marks 98 and 104 can be compensated. Scanner and detector circuitry 122 is connected to

CCD 112 and scans each horizontal and/or vertical line of the face of CCD 112. Device 122 further detects and pro-

cesses the signals obtained by such scanning. In response to these signals and the processing thereof, circuitry 122 provides signals to control the movement of wafer 40 by wafer handling means 16. The detected signals and manner of processing the detected signals by circuitry 122 will be described hereafter with respect to Figures 6 and 7.

Referring now to Figure 5, the illuminating means 108 is shown placed above and perpendicular to one of the legs 106 of alignment marks 98 or 104. Further, illumina- tion means 108 is placed below, or at least to the side of, lens 110 so that the light therefrom does not pass directly through lens 110. With this positioning, the light from illuminating means 110 strikes leg 106 at approximately a 30° angle relative to the horizontal. Thus, that portion of the light from illumination means 108 striking the flat horizontal top of leg 106 is reflected away from lens 110; similarly the light striking the approximately vertical side of leg 106 is also reflected away from lens 110. In other words, the only light reflected towards lens 110 will be that light scattered as a result of striking the edge 124 of leg 106. By placing an illuminating means similar to illuminating means 108 on the other side of lens 110, as seen in Figure 5, light reflected from the other edge 125 could also be seen through lens 110. By knowing the posi- tion of the two edges 124 and 125, the center of leg 106 can be easily determined.

Referring again to Figure 3, four orthogonally po¬ sitioned illuminating means 126, 128, 130 and 132 are shown. Each of the illuminating means 126, 128, 130 and 132 is aligned with the mark 98 or 104 so that it is generally above and perpendicular to two of the legs 106. In this manner, the longitudinal edges 134 are illuminated by illu¬ minating means 126, the longitudinal edges 136 are illu¬ minated by illuminating means 128, the longitudinal edges 138 are illuminated by illuminating means 130, and the lon¬ gitudinal edges 140 are illuminated by illuminating means 132. It should be noted that the end edges 142 of leg 106

are also illuminated by the illuminating means 126, 128, 130 and 132, although the amount of light scattered to lens 110 will be substantially less. By turning on each of the illuminating means 126, 128, 130 and 132, one at a time, each of the individual edges 134, 136, 138 and 140 can be individually seen and detected. Alignment may also be done by turning on each of the illuminating means 126, 128, 130 and 132 together, although interference and spurious sig¬ nals may be present. Further, by turning on the four illu- mination means 126, 128, 130 and 132 one at a time, the amount of heat generated by the alignment system is re¬ duced, and thus has a less negative effect on the thermal stability of the system.

In practice, each of the illuminating means 126, 128, 130 and 132 may be laser diodes which provide near infrared light at a wavelength of approximately 800 nanome¬ ters. For example, a solid stare galliun arsenide type of laser diode may be used. Further, the laser diode generally will be positioned well away from the lens 110 and the near infrared light is transmitted to the positions around lens 110, as illustrated schematically in Figures 3 and 5, by conventional fiber optics and other light directing means, such as mirrors and lens. The details of the structure for directing the laser diode light around lens 110 will be described hereafter with respect to Figures 8, 9 and 10. The reason that near infrared light having a wavelength of approximately 800 nanometers is used is because light at this wavelength is a good compromise between absorption in the boron doped silicon material of membrane 37 and spacial resolution. Further, the scattered signal portion of the light must again pass through membrane 37 in order to be properly focused by lens 110.

Referring to Figures 4, 6 and 7, the manner of de¬ tecting the center of each set of alignment marks will now be described. As previously noted, the structure illus¬ trated schematically in Figure 4 causes an image of the light collected by objective lens 110 to appear on the face

of CCD 112. Figure 6 illustrates how this image appears after all four illuminating means 126, 128, 130 and 132 are operated. If the four illuminating means 126, 128, 130 and 132 are operated one at a time, the images reflected by 5 each illuminating means 126, 128, 130 and 132 need to be stored in a memory associated with CCD 112 in order to continue displaying the image after operation ends for each particular illuminating means 126, 128, 130 and 132.

As seen in Figure 6, the two sets of images 144 and _Q 146 directed to £CD 112 represent the edges 128, 130, 132 and 134 of the two alignment marks 98 and 104. In using the alignment system 10, alignment between the mask 38 and each section 100 of wafer 40 is when the two marks are on the same Y line, or displaced by a specific distance in the 5 Y direction, and displaced by a specific distance in the X direction. Thus, at alignment the two sets of images 144 and 146 appearing on CCD 112 will appear vertically at the same, or similar, level, but horizontally displaced, as seen in Figure 6. Alignment can be detected using scanning _Q and detecting circuitry 122 by comparing sequences of scan lines taken across the face of CCD, such as scan lines 148 and 150 in Figure 6. In practice, an entire series of scan lines similar to lines 148 and 150 will be obtained and provided to circuitry for processing. In addition, verti- 5 cal scan lines may also be utilized.

The electrical signals 152 and 154 derived from the scan lines 148 and 150 are shown in Figure 7. Each signal 152 and 154 include four pulses 156, 158, 160 and 162 cor¬ responding to the presence of image light in the image sets _Q 144 and 146. Further, each pair of pulses 156, 158 and 160, 162 are separated by a slight duration of reference value signal 164 and 166 in the signals 152 and 154. Circuitry 122 compares all of the scan lines, such as 148 and 150, as represented by the electrical signals, such as signals 152 5 and 154. By so doing, a determination by circuitry 122 can be made first when the signals are aligned in the Y direc¬ tion by assuring that each scan line includes the two sets

of pulse pairs 156, 158 and 160, 162 and second by assuring that the two sets of pulse pairs 156, 158 and 160, 162 are separated by the proper distance. The proper distance is determined by detecting and measuring the time, or distance in scan lines, between reference values 164 and 166. The reference values 164 and 166 can further be compared from scan line to scan line to assure proper orientation of the alignment marks 98 and 104. Alternatively, other conven¬ tional wave comparing techniques ^* may be used to make the determination of alignment.

In any alignment system, the critical part of the aligning process is placing one alignment mark in the prop¬ er position relative to the other alignment mark. Conven¬ tional techniques have done this by placing one entire mark over, or relative to, the other entire mark. Because the two marks are relatively wide (e.g. two to three microns) extreme precision, to one tenth of a micron, or less, is not possible by the conventional techniques. However, by only detecting the edges of the alignment marks 98 and 104, as herein described, the precise center of each alignment mark 98 and 104 can be found. The center will be the center of the reference value points 164 and 166. Because these points 164 and 166 are so precisely identifiable, as seen in Figure 7, precision to better than a tenth of a micron becomes possible. Now, all that needs to be done is detect when the points 164 and 166 are separated by a fixed dura¬ tion. Further, by noting whether the duration between points 164 and 166 is greater or less than the desired du¬ ration, control signals can be provided to wafer handling mechanism 16 to bring system 10 into alignment.

Referring now to Figures 8 and 9, a specific embod¬ iment of an alignment system 168, incorporating the con¬ cepts of the more generalized alignment system discussed above, will be described. Where appropriate, like numeri- cal designations are used for like component parts in the Figure 8 and 9 specific embodiment relative to the more generalized discussion above. Alignment system 168 shown in

Figures 8 and 9 includes objective lens system 110 and four light deflectors 170, 172, 174 and 176, which are shown in detail with in Figure 10. Light deflectors 170, 172, 174 and 176 each receive 800 nanometers light from respective fiber optic elements 178, 180, 182 and 184 coupled thereto. Light deflectors 170, 172, 174 and 176 are mounted on ob¬ jective lens 110 so that light at a wavelength of 800 na¬ nometers is orthogonally directed in the area of the align¬ ment mark 98 on mask membrane 37. This light passes through membrane 37 and also illuminates the surface of wafer 40, including wafer alignment mark 104 thereon as it is moved generally beneath mark 98. As previously noted, the two alignment marks 98 and 104 are slightly offset from one another when proper alignment is found, but the amount of such offset is too small to be shown in Figures 8-10.

Referring now to Figure 10, each of the four light deflectors 170, 172, 174 and 176 are mounted from four or¬ thogonal directions on the casing of objective lens system 110. Objective lens system 110 may be any commercially available objective lens system with a magnification of 25X and it may include more than one actual lens element. Be¬ cause of the limited space available in exposure column 36, objective lens system 110 should also be as small as possi¬ ble. The case containing objective lens system 110 has four orthogonally positioned slots formed therein for re¬ ceiving the four light deflectors 170, 172, 174 and 176. Each of the light deflectors 170, 172, 174 and 176 receives one of the fiber optic elements 178, 180, 182 and 184 at¬ tached thereto, as shown with respect to light deflector 170 and fiber optic element 178. As seen in light deflec¬ tor 170, there is a lens system 186 for receiving the light from fiber 178 and directing it through opening 188 against a mirror 190. From mirror 190, the light is reflected to¬ wards mask alignment mark 98 and through membrane 37 to- wards wafer 40. Each of the other light deflectors 172, 174 and 176 is identical to deflector 170.

Referring again to Figures 8 and 9, the four fiber optic elements 178, 180, 182 and 184 are connected to four laser diode devices 210, 212, 214 and 216. Each of laser diode devices 210, 212, 214 and 216 are selected to provide laser light of 800 nanometers to one of the fiber optic elements .178, 180, 182 and 184. Further, each of the laser diode devices 210, 212, 214 and 216 may be controlled so that only one at a time is operated to provide such light. Fiber optic elements 178, 180, 182 and 184 are then bundled and provided to respective light deflectors 170, 172, 174 and 176. The light scattered towards objective lens system 110 from the alignment marks 98 and 104 is magnified by 25X. The image from objective lens system 110 is reflected by mirrors 192, 194 and 196 and further magnified by 4X by divergent lens 198. In this manner, only one at a time of the edges 134, 136, 138 and 140 of the alignment marks 98 and 104 is illuminated.

Alignment system 168 is generally constructed on two levels. The upper level is contained on a plate 200 ^an£ i the lower system is held by bracket 88 and the associ¬ ated mechanisms 202 for moving bracket 88 upward and out¬ ward from mask membrane 37 when alignment system 168 is not in use. It should be noted that objective lens system and the components affixed thereto must be moved away from the X-ray exposure area so that it does not block the X-rays 34. This movement is accomplished by mechanism 202 merely moving bracket 88 laterally (to the left in Figure 8) in a manner determined by the pins 92 and 92' sliding in the slots of brackets 90 and 90*. Pins 92 and 92* are a part of bracket 88 and brackets 90 and 90' are affixed from plate 200. Mechanically, mirror 192 is affixed to objective lens system 110 directly above lens system 110 and mirror 194 is connected to the lower side of plate 200 at the same level as mirror 192. The light is reflected through an opening 203 in plate 200.

Because of the 10OX total magnification provided by objective lens system 110 and divergent lens 198, it is not possible to focus both the mask alignment mark 98 and the wafer alignment mark 104 at the same image plane due to the approximately twenty micron gap between mask membrane 37 and the upper surface of wafer 40. To solve this problem, a portion of the image from lens 198 is diverted using re¬ flectors 114, 116, 118 and 120, mounted on plate 200, as best seen in Figure 9. Reflector 114 is actually a prism which reflects seventy percent of the light applied thereto and passes thirty percent of the light applied thereto. Re¬ flector 120 is a mirror and is positioned to chop one half of the image transmitted through prism reflector 114. By making mirror 114 a partially reflecting prism, compensa- tion is provided due to the presence of membrane 37, when imaging the light reflected from wafer alignment mark 104.

Mirrors 116 and 118 are mounted on a bracket 204 which is adjustably mounted in slot 206 of plate 200. Bracket 204 may be adjusted to compensate for variations in the gap between membrane 37 and wafer 40, as it will vary under certain circumstances. While not shown, automatic means may be included to provide proper focus of the marks, based upon the variations of the gap. A lens 208 is placed between mirrors 114 and 116 to assure that the two images focused on CCD 112 are the same size.

With the components shown in Figures 8, 9 and 10, the images focused on CCD 112 are as shown in Figure 6. These images can then be detected and processed as previ¬ ously described and proper alignment between mask 38 and an appropriate section 100 of wafer 40 can be obtained.