Aligning Apparatus

1. Field of the Invention

The present invention relates to an aligning apparatus for mutually aligning a first substrate with a second substrate, and more particularly to an aligning apparatus adapted for use, for example, in an exposure apparatus for semiconductor device manufacture.

2. Related Background Art

For use for example in an exposure apparatus, the U.S. Pat. No. 4,251,160 discloses an aligning apparatus utilizing grating marks both on a mask constituting a first substrate, and a wafer constituting a second substrate. Also there is known an aligning apparatus in which the mask and the wafer have grating marks of mutually different pitches, and the alignment of the mask and wafer is achieved by detecting the position of a moire fringe pattern, generated by optically superposing said grating marks, with an image sensor. Though such known apparatus can detect a very small displacement with a high precision, fetching of the image information of the moire fringe pattern requires a certain time due to the limitation in the clock frequency of the image sensor, so that, in case of an aligning operation of an object in motion, the precision of positional measurement is deteriorated due to the movement of said object during said time.

In consideration of the foregoing, the object of the present invention is to provide an aligning apparatus in which the aligning precision is not affected, in the course of aligning a first substrate with a second substrate, by movement or vibration of a table supporting said second substrate.

The aligning apparatus of the present invention for mutual alignment of a first substrate with a second substrate, which are capable of mutual displacement, is featured by a movable table for moving at least either of said first and second substrates; first alignment mark means formed in predetermined position on said first substrate and composed of a pair of mutually different grating marks; second alignment mark means formed in predetermined position on the second substrate and composed of pair of grating marks constructed in opposite direction to said first alignment marks in such a manner that two moire fringe patterns, formed by optically superposing said second alignment marks with said first alignment marks, move in mutually opposite directions; illumination means for instantaneously illuminating (pulse illumination) said first and second alignment mark means; position reading means for reading the relative position of said first and second substrates in synchronization with said instantaneous illumination by said illumination means; moire fringe pattern detecting means containing a charge-accumulating image sensor for detecting the position of said two moire fringe patterns; an alignment optical system for guiding the images of said two moire fringe patterns to the image sensor; calculation means for calculating the phase difference of said two moire fringe patterns from the information on said moire fringe patterns accumulated in said moire fringe pattern detecting means and further calculating a positional deviation between the first and second substrates from thus calculated phase difference and from the value read by said position reading means; and driving means for moving said movable table according to the result of said calculation.

In the above-explained apparatus of the present invention, the alignment mark means on the first substrate is superposed with that on the second substrate by the movement of the movable table. Two moire fringe patterns, formed by a pair of alignment marks on the first substrate and a pair of alignment marks on the second substrate, instantaneously illuminated by the illuminating means, move in mutually opposite directions in response to the movement of the object. Said two moire patterns are guided through the alignment optical system to the charge-accumulating image sensor and stored therein as the moire fringe pattern information. On the other hand, the position of the movable table, or the relative position of the first and second substrates, is read by the position reading means in synchronization with the instantaneous illumination by the illuminating means. The stored moire fringe pattern information is supplied to the calculating means which calculates the phase difference of two moire fringe patterns and further calculates the positional difference between the first and second substrates, from the value read by the position reading means and from the calculated phase difference. The first and second substrates can be mutually aligned by moving the movable table by thus determined positional difference.

Thus, according to the present invention, the precision of measurement is not affected at all by an eventual positional error of the image sensor, since paired alignment marks on the first substrate and those on the second substrate are so formed that two moire fringe patterns obtained by respectively superposing said alignment marks move in mutually opposite directions and said moire fringe patterns are detected by the image sensor. Also the relative position of the first and second substrates is determined in synchronization with the instantaneous illumination to enable measurements of said mutual position in succession during mutual movement thereof, thereby allowing rapid and highly precise alignment.

Other objects of the present invention and the advantages thereof will become fully apparent from the following description, which is to be taken in conjunction with the attached drawings.

FIG. 1 is a block diagram of an embodiment of the present invention applied to a projection exposure apparatus;

FIG. 2A is a magnified plan view of alignment marks formed on a mask shown in FIG. 1;

FIG. 2B is a magnified plan view of alignment marks formed on a wafer shown in FIG. 1;

FIG. 3 is a schematic view showing the light path of a moire fringe pattern detecting unit shown in FIG. 1, seen from a direction A;

FIG. 4 is a schematic view showing the separation of diffracted light by a spatial filter;

FIG. 5 is a plan view showing moire fringe patterns on an image sensor shown in FIG. 1;

FIG. 6 is a plan view showing the arrangement of alignment marks on a mask;

FIG. 7 is a plan view showing the arrangement of alignment marks on a wafer; and o FIG. 8 is a plan view showing a moire fringe pattern formed with alignment marks different from those shown in FIGS. 2A and 2B.

FIG. 1 shows an embodiment of the present invention, applied to a projection exposure apparatus. A light beam I, for illuminating the alignment marks, emitted from a pulse light source 2 activated by a signal from a timing circuit 1 to be explained later, is guided through a field stop 3, a lens 19, a mirror M1, a half mirror 4, an objective lens 5 and a mirror M2 and illuminates alignment marks MY on a mask 20. The field stop 3 limits the areas illuminated on the mask 20. The images of the mask marks MY are projected onto a wafer 14 by a projection lens 11. Said mask marks MY are formed as grating patterns consisting of mutually parallel plural reflecting portions and transmitting portions (slits), and is composed of a grating mark MM1 having a grating pitch P1 in a measuring direction Y when projected onto the wafer 14, and a grating mark MM2 having a grating pitch P2 in a similar manner. The alignment marks WY on the wafer are composed of a pair of marks WM1, WM2 each consisting of plural rectangular projections arranged in a matrix as shown in FIG. 2B, wherein the mark WM2 has a grating pitch P2 in the measuring direction Y while the mark WM1 has a grating pitch P1 in said direction. The above-mentioned marks are so arranged that the mask mark MM1 superposes with the wafer mark WM2 through the projection lens 11, and the mask mark MM2 simultaneously superposes with the wafer mark WM1. The light projecting the images of the mask marks onto the wafer 14 is reflected by the surface thereof, and generates, upon illuminating the rectangular projections of the wafer marks, diffracted light in a direction X perpendicular to the measuring direction Y, depending on the arrangement of said rectangular projections in said direction X. The reflected and diffracted light is again transmitted through the projection lens and the slits of the mask marks MY, then guided through the mirror M2, objective lens 5 and half mirror 4, and is focused through a spatial filter 8 and an imaging lens 6 onto an image sensor 7. Simultaneously the images of the mask marks on the mask 20 are formed on the image sensor 7. However the spatial filter 8 positioned conjugate with the pupil Pl of the projection lens 11 intercepts the normal reflected light from the wafer 14 and from the mask 20, thus allowing the diffracted light only, generated by the wafer marks WY in the direction X, to reach the image sensor 7 as will be explained further in the following with reference to FIG. 4.

The spatial filter 8 intercepts the normal reflected light from the wafer and a light portion D.sub.x0 entering the mask from the mirror M2 and diffracted by the mask marks MY, so that only the light beams diffracted by the wafer marks WY in the direction X and further diffracted by the mask marks MY, thus forming moire fringe patterns, are transmitted by apertures 8a, 8b and reaches the image sensor 7 through the imaging lens 6. Thus a moire fringe pattern, consisting of light and dark portions as shown in FIG. 5, is formed by the difference in grating pitch (p.sub.1 -p.sub.2) of the mask marks MY and the wafer marks WY in the measuring direction Y, on the image sensor 7. The positional deviation between the mask 20 and the wafer 14 is determined by measuring the position of said moire fringe pattern by means to be explained later.

When the mask 20 and the wafer 14 are mutually superposed with an error not exceeding .+-.1/2 of the grating pitch of the aligning grating marks MY, WY by an unrepresented crude aligning system such as a global alignment optical system, for example employed in a reduction projection exposure apparatus (stepper), a start signal for starting fine alignment is given to the timing circuit 1, which, in response, scans the image sensor 7 to dissipate the charge accumulated thereon. Then the timing circuit 1 sends a trigger signal to the light source 2 for pulse illumination and a hold circuit 17 which holds the value read by an interferometer 9 for measuring the position of a wafer table 15 movable with the wafer 14 placed thereon, thereby causing light emission from the pulse light source 2 and causing the hold circuit 17 to read the position of the wafer table 15. The aligning marks MY. WY illuminated by the pulse light generate moire fringe patterns by diffraction on the image sensor 7, and the information of said patterns is stored in the form of charges on the image sensor 7. In said patterns, the high-order pitch components are removed by the spatial filter 8 as shown in FIGS. 3 and 4. FIG. 5 shows the moire fringe patterns of the alignment marks, formed on the image sensor 7, in which MA indicates a moire fringe pattern obtained by superposing of the mask mark MM1 and the wafer mark WM2, while MB indicates a pattern obtained by superposing of the mask mark MM2 and the wafer mark WMl. If the alignment marks MM1, MM2, WM1 and WM2 have grating pitches satisfying a condition P1>P2, the moire fringe pattern MA moves in the same direction as the moving direction of the wafer 14, while the pattern MB moves in the opposite direction. On the other hand, if P1<P2, the pattern MB moves in the same direction as the moving direction of the wafer 14 while the pattern MA moves in the opposite direction. The positions of said moire fringe patterns MA, MB are read from the image sensor 7 to a frame memory 12, and a computer 16 executes a Fourrier analysis to extract the basic frequency components of the moire fringe pattern, and to determine the phase angles thereof. Said extraction of basic frequency components removes noises different from the frequency of the moire fringe patterns, such as caused by unevenness in the sensitivity of illumination of the image sensor, thus improving the precision of position reading.

The computer 16 determines the phase angle .phi..sub.A of the moire fringe pattern MA and that .phi..sub.B of the pattern MB as represented by:

wherein .phi..sub.0 is a phase angle with respect to a reference point on the image sensor 7 when the mask 20 and the wafer 14 are mutually aligned to a desired position; .DELTA..phi. is a change in the phase angle resulting from a movement of the image sensor 7; and .phi..sub.1, .phi..sub.2 are phase angles of the moire fringe patterns MA, MB corresponding to the movement of the wafer 14 with respect to the mask 20. Thus the amount of relative movement between the mask 20 and the wafer 14 can be determined from a difference of the measured phase angles of the patterns MA and MB, namely .phi.=.phi..sub.A -.phi..sub.B =.phi..sub.1 -.phi..sub.2, without depending on the position of the image sensor 7.

More specifically, the basic frequency component of the moire fringe pattern MA, when the wafer 14 is moved by Y.sub.0 with respect to the mask 20, is represented by: ##EQU1## Similarly the basic frequency component of the pattern MB is represented by: ##EQU2## so that said difference of phase angles is represented by: ##EQU3## Consequently the relative position Y.sub.0 of the wafer with respect to the mask is given by: ##EQU4##

An adder 18 calculates the position of the wafer table 15 where the wafer 14 is aligned to the mask 20 at a desired position, from thus determined value Y.sub.0 and the position of the wafer table 15 read by the hold circuit 17 from the interferometer 9 at the pulse illumination. Even during said calculation, the position of the wafer table 15 is continuously read by the driving circuit 10 through the interferometer 9, and said driving circuit controls a driving motor 13 so as to move the wafer table 15 to the desired alignment position.

As shown in FIGS. 6 and 7, alignment marks MX, MY, M0 and WX, WY, W0 are formed, respectively on the mask 20 and the wafer 14, in three positions around the exposure area of a step, and used respectively for measuring the position in the directions X, Y and the rotational direction 0 as indicated by suffixes.

The foregoing explanation relating the FIGS. 1 to 5 is limited to the detection of position in the direction Y, but the alignment can be achieved in the direction X and the rotational direction 0 by similar means. Also it can be confirmed that the mask and the wafer are mutually aligned with a desired precision, by repeating the above-explained procedure. The measurement of position can be conducted while the wafer table 15 is in motion or is stopped, and the mask can be aligned to a desired position of the wafer 14 with a high precision, without influence of the movement or vibration of said table, or of a positional error of the image sensor 7.

In place for the grating marks of different pitches employed in the foregoing embodiment, there may also be employed, as shown in FIG. 8, marks MM10, MM20, WM20 and WM10 of a same pitch, formed respectively on the mask and the wafer with mutually opposite slight inclinations to achieve a similar effect. In this case, when the wafer marks WM10, WM20 are moved in the direction Y together with the wafer, the two moire fringe patterns move in mutually opposite directions X perpendicular to said direction Y. Consequently the image sensor 7 can be placed along said pattern moving direction X.

In an exposure apparatus employing an excimer laser as the light source, a part of the excimer laser beam may be employed as the light source for alignment.