CA1193030A - Alignment device for integrated circuit manufacturing machines - Google Patents

Abstract

AN ALIGNMENT DEVICE FOR INTEGRATED CIRCUIT
MANUFACTURING MACHINES

ABSTRACT OF THE DISCLOSURE

The alignment device , usable in wafer steppers, allows a reference pattern (7) carried by a reticle (1) to be aligned with an alignment pattern carried by the sili-con wafer on which the design of the reticle is to be repeat-ed. This device comprises a microscope movable in two conjugate directions x and y of the directions of movement of the reticle and of the wafer, and a sighting mirror (15).
This mirror is movable between a position where it frees the path of the beam coming from an illuminator (3) and a position in which it directs an illuminating light pencil from a source (17) through the reference pattern towards the wafer and brings the reflected light back towards the microscope while causing the mean ray to pass through the object focus (F) of the photoreducing lens (4).

Description

~3~3~

An alignment devi~e for integrated circuit manufa~t:~ring nn~hines BACKGRO~IND AND SUMMARY OF THE INVENrlON
The present i~vention rela-tes to ~he fiel~ oF integrated circuit manu~ac-turing and particularly to -those machines called "wafer steppers".
It relates more particularly to a device sui-table for aliyning a reticle, tha-t is to say a mask reproducing on a large scale one of the successive patterns to be formed on each elementary chip of the wa~er, with appro-priate positioning on the wafer before exposure of a photoresist on the latter.
Numerous machines of this type have already been described and reference may for instance be made to European Pa~ent Publication No.
OO 17759. They generally comprise a ~eticle having means for moving it in its plane, at least in two orthogol1al directions x and y corresponding to the two main axes of the reticle ;
a table for receiving the semiconductor waFer, typically silicon, having at least one alignment pattern, said table being equipped with means for mGving it in two col~ gate iirections of ciirections x and y ; a photore-lucing lens dis-posed between the tables ; means for illuminating the retic].e

2~ with actinic light for exposing the semiconductor wafer, anci means for simultaneous observation of an aliynment pattern on the wafer and a reference pattern on the reticle, compris-ing a microscope and detectors.
It is essential that the manufacturin(~ machine allows the image of the reticle to be reproduced in reduced form with respect to the wafer with high accuracy, of the order of O.l micron, on the wafer. The constructior1 of an aliqnment device complying with this condition raises numerous prob-lems, the resolution of which is to a crrLairl r.~xtent cnntra-dictory. Durin[~ aliqnment, an arnoul1t of` rlCt iO.ir ~ llt Wtl.iChimpresses an imaqe on i:l)e res.il1 wh(re it must sub~e~uently be exposed must not be appliec~ to the zones ol` the photo-sensitive resin layer wl1ich cover thr.~ wafer ; if we attempt to resolve thi.s problem by using, for alicJnrnrr1t, light of a sufficiently high waveler1gtl1 so as not to impress an image on the photosensitive resin, we come up against the problem of

3~

chromatic aberration. If For example, the aligrlment is carried out with a fir~t wavelength (line e of mercury For example) whereas the exposure is carried out with a second wavelength (line g oF mercury For example), it is very 5 difficult to construct a lens having exactly the same Focal length for both wavelengths.
To resolve this problem, alignment devices have been provided in which tlle reticle is aligne(i with respect to the wafer in two steps, by lJsing an intermediate reference : the lO alignment takes place for a fixed position of the wafer with respect to the field of the photoreducing lens, then the table receiving the wafer is caused to move with a trans-lational movement whose amplitude and direction ~orrespond to the distance between the alignment position anci the exposure 15 position. A description of machines oF this kind may be found in numerous documents, such as patents US 4 103 99 iR 2 40~ ~47 and rR 2 4~2 750.
It has also been proposed to carry out the alignment using a narrow pencil of ligllt from the exposurc source, 20 deFined by a diaphragm which masks all the l)eam other than the part corresponding to the zone oF the alignment pattern.
This masl< is then removed for exposu~e. In this way, we end u~ with a very complex arran~ement (IJ~-A-4 1 s3 371 ), It is an object of the invention to improve upon the prior art alignment devices; it is a more specific obiect to allow simple and straightforward alignment of a wafer with respect tû a reticle while the wafer is under the cond-itions and in the position in which it will be subsequently exposed; it is another object to remove limi-tations on the positioning of the alignment marks in the circ~ t to be forrned or a cut-out lane of the wafer, between adjacent chip locations.
To this end, there is provided a device in which-the observa-tion means, for finding coincidence between the alignment pattern on the wafer and the reference pattern on the reticle comprise a microscope hav-~a3~3~D

ing an optlcal system switcl)able between ~ state in whicil itfrees the path of the exposure beam and a state in which it directs an illuminatirlg penci] of light from a sou~ce incorp-orated in the microscope and llaving the same wa~elength as 5 the illumination means, tllrough the reference pattern, to-wards the wafer anc~ in which it brings the light reflected by the reticle and the light reflected by the wafer back to the microscope while causing the mean ray to pass through the object focus of the photoreducing lens.
Such a device imposes no limitation on the size of the elementary ci~cuit to be reproduced, that is to say the chips to be formed from the wafer. The optical system associated with the microscope ~ay comprise a reflecting mirror movable, or "flippable", between a retracted position and an alignment 15 position. So that the mean incident and reflected rays on the wafer ha~e the same path, which involves the incident ray passing through tlle object focus of the photoreducing lens, whatever the point of the field of this lens which is aimed at, there will generally be disposed, ~etweerl the 20 illumination means and the reticle, and in close proximity to tllis latter, a field lens whose focal length is substant-ially equal to the distance between the reticle and the object focus. Thus, if the rays from the illumillation means striking tlle field lens are parallel to the optical axis, 25 the refracted rays will always pass tllrough the object focus, whatever the posit~on of the microscope. The field lens will have very little influence on the enlargement and quality of the image, since it is close to tlle retic]e. Itsonly effect is to turn back the ray towards the ol)jcct focus.
The invention will be be-tter unders-tood from the following descrip-tion of particular embodiments, given by way of examples.
SHORT DESCRIPTION OF T~IE ACCOMPANYING DRAWINGS
Figure 1 is a general perspec-tive diagram showing the optical elements of an integrated circuit manufacturing machine comprising an alignment device according to the invention, the distances along the optical axis no-t being ~3~

respec-ted for the sake of clari-ty;
Figure la is a detail view on a larger scale showing the positioning of a chip ~elemen-tary circuit) and of its alignment patterns on -the wafer;
Figure 2 is a general diagram showing the op-tical elements of the microscope of Figure 1 and its associated elements;
Figures 3 and 4 are diagrams showing the trend of the signals analyzed;
Figure 5 is a diagram showing a variation of Figure 2;
Figure 6 is a diagram illustrating possible oper-ation of the machine of Figure 1 with polarised light for decreasing the background level due to back scattering;
Figure 7 is a schematic representation of an image analyzing system which may be embodied in the device of Figures 1 and 2;
Figures 8 and 9 are schematic illustration of -the possible use of an alignment mark consisting of two gratings and the dis-tribution of the diffracting light, for align-ment along one clirection only;
Figures 10 and 11, similar to Figures 8 and 9, illustrate the use of gratings for alignment along two orthogonal directions;
Figure 12, similar to Figure ~, is a schematic representation of an alignment system using gratings diffr-action;
Figure 13 is a schematic representation of an hybrid system allowing image analysis as well as diffraction analysis.
DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS
Referring to Figure 1, there is shown a machine oF
the type currently called "wafer s-tepper" or "step and repeat" machine suitable for reproducing at even intervals on the photosensitive resin covering a semiconductor wafer 2 (in general silicon)~ a pattern drawn on a re-ticle or mask 1. The general construction of the machine shown in Figure 1 is conventional. It comprises an illuminator 3 able to supply, for a well defined period, a parallel beam of actinic light, a reticle carrying table 5, a projec-tion lens 4 for collecting the light which has passed through the reticle and forming the image thereof on the silicon wafer 2, carried by a table 6. The mechanical components will not be described in detail, since they may be of any cor,ventional construc-tion, for example that described in U.S. Patent No.

4,153,371. It is sufficient to note that the reticle carrying table 5 is designed so as to allow the reticle to be moved in two orthogonal directions x and y and in rotation about an axis parallel to the optical axis of the machine. This table allows the orientation of reticle 1 to be adjusted so that the edges of the design -are parallel to axes x and y. ~ith table 6, movements may be imparted to wafer 2 in directions x and y corresponding to the repetition step of the design on the wafer.
Reticle 1 carries not only the circuit diagram to be reproduced, but also reference patterns 7 which will be caused to coincide with alignment marks 8 carried by the wafer so as to ensure alignment with an accuracy which must be of the order of 0.1 micron in both directions x and y. In Figures 1 and 2, several successive alignment patterns are shown, such as 8, for each chip 9. The reason is that, when a positive photo-sensitive resine is used, the resin layer which has been zexposed disappears during development, at the same as the alignment pattern which has just been used, so that several may be used in succession when manufact-ure of the chip requires several exposures under different reticles and several developments.

~33C~3~

The marks are placed either in the part of the design close to the edge or --so as to use to the maximum the available area o-f the wafer-- in cut-out lames 9a in wafer 2 (Figure 1a). The patterns may have very different shapes but, as a general rule, they comprise lines crossed at 45 from directions x and y for aligning in these tWO directions.
The alignment device properly speaking comprises an observation microscope lû. Microscope 10 is i-tself ùounted on a crossed movement table ll for moving it in two orthogonal directions conjugate of directions x and y of the reticle and of the wafer. Table ll may be associated wtih a movement control device pro-grammed to cause movements of well defined amplitude in both directions. The amplitude of the rectilinear movement in direction x, fixed once and for all for a chio of given dimensions, allows observation to pass from one chip to the next. Translation in direction y makes it possible to go from one alignment mark 8 on the wafer (and from the corresponding refer-ence mark 7 on the reticle) to the next.
The microscope is provided with a light source at the exposure light wave length, one particular embodiment of which will be described further on. This source illuminates a window 12 of the re-ticle contain-ing the reference mark 7 used. The photoreducing lens 4 forms an image-of this window on a zone 13 of the wafer containing the alignmen-t mark 8. The image of the alignment mark 8 is conversely formed in window 12 by the photoreducing lens 4. Thus, micro-scope lO receives the light reflected by window 12 and ~5a3~3~

20ne 13 for simultalleo-l31y obierving marks 7 and 8 ',o that tlle mean incide~lt arld reflected rays on waFer 2 follow merged paths, the inciderlt ray must pass throuyh the object focuc; F of tt-le photoreducinq lens 4 whatever the

5 point of the field aimeci at by the microscope. io cornply with this conclition, the alignmerlt devicr comprises, immedi-ately above reticle 1, a field lens 14 whose focal length is equal to the distance between reticle 1 and focus F. Thus, provided that the incident ray striking lens 14 is parallel 10 to the optical axis, the refracted ray always passes throuyh F, whatever the position of the microscope. Lens 14 has very little influence on the enlargement and quality of the image, since it is close to window 12. It will be sufficient to give it an optical quality comparable to that of illuminator 3.
15 So that the image is satisf`actory, the photoreducing lens 4 will be chosen so that its object focus F is in the vicinity oF its input pupil. In practice, this condition is generally fulfilled by the usual lenses~ which are telecentrically arranged. The li~ht beam for illuminatin~
20 window 12 must be parallel to the optical axis. Since micro-scope 10, on the other hand, is placed laterally, a sightiny mirror 15, reflecting the light at 90, is placed in the optical path between microscope 10 and lens 14. With the reference patterns to be inscribed on the resin covering 25 wafer 2, mirror 15 must be able to be rapidly retracted to allow exposure. This arrangernent furtt)er allows the image obtained by direct reflection on reticle 1 to be isolated by placing a diaphragm 16 of suitable shape. DiaphIagm 16 is placed at the image focus of the lens of tlle microscope where 30 the image of the source incorporalcd in the microscope is formed, as will be secn Further on.
- A single alignment device has been shown in figure 1 but two devices may of course be used which are aimed at marks placed respectively in the right-hand and left-hand 35 cut out lanes.
The provision oF a retractable mirror 15, or more gener-ally a switchable optical system, imposes no serious restric-3~

tion on the machine : in fact, since m;rror 15 may be veryligi~t, thr- positioning ancl reiraction timts are very snort (e.g. 50 ms and 40 ms). The t;-me required fGr carrying out the alignment with the reguircd accuracy is also very short, 5 oF the order of 1~0 ms. Exposure under a light intensity much greater than that used during alignment will generally be of the order of lnO ms also.
There will now be described, with reference to fiyure 2, the general construction of the microscope and of the 10 associated parts (source and analysis system).
The source incorporated in the microscope is formed, in the embodiment shown in figures 1 and 2, by an optical fiber 17 which is sufficiently flexible to follow the mDvements of the microscope. This fiber takes a fraction of 15 the light frnm the exposure lamp incorporated in illuminator 3 (in general a mercury lamp). The light exiting From the endmost face of fiber 17 forming the source passes throutlh a field diaphragm 1~ and an aperture diaphragm 19 al a loc-ation where a llollaston prism 20 may be placed when it 20 is desired to form an interferential contrast. A semi-transparent mirror 21 reflects the illuminating light beam to the lens of microscope 10.
The light reflected by the wafer, which will form images of marks 7 and a is collectrd by the lens of the 25 microscope~ passes through thr semi-transparent mirror 21 and is spl it into two paths, each analyzing one of the directions of the patterns. A separator ~semi-transparent separating mirror or separating cube, possibly polarizing) divides the light intensity betwetrl thtl two paths whicll have 30 similar construction. Path l for txample compriscs succes~ivt-ly a spatial filter 23, reflrcting mirrors 24 and a detector 25. Spatial filter 23, Formed generally by a diaphragm having a slit perpendicular to the analysis direction, is intended to block the image coming from the direct reflection from 35 reticle 1. It is placed in the image focus oF the lens, corresponding to the plane of the pupil, that is to say ac the only position where in practice the two images are 3~

separate. A Wollaston prism (not shown) may also be placed in the vicinity of the Focus so as to provide interferential contrast when only 3 phase imrl9e iS available.
~lirrDrs 24 in path I and 26 in path II give to the 5 image a constant lateral position when the microscope is moved in direction y (figure 2) in the field of the reducing lens 4. The lateral translational movements in direction y are thus transformed into longitudinal translational move-ments, in the overa~ moving direction x of the analysis 10 device as a whole. The focussing defect which results there-from is in practice compensated for by the variation of the thickness passed through in lens 14, when the enlargement of microsocpe 10 is optimized for this purpose.
Detectors 25 may be of different types. When, more 15 especially, the patterns are of the kind shown in figures 1 and la, the image analysis may be effected for example by means of slit scanning. The representation obtained is of the kind shown in figure 3. The position of mark 1 8 on the wafer is calculated after digitalization of the signal with 20 respect to the center of the analysis window 12 and center-ing is provided by moving table 6. Before the signal is proc-essed, it is advan~a~eous to differentiate it,then to rectify the negative peaks so as to obtain the form sl-own in figure 4.
From the deviation measurement made on the signal the align-25 ment may then be performed by moving table 5 along x and/oralong y.
Figure 5, where the parts corresponding to those show in figure 2 bear the same reference number9 describes anoth-er embodiment providing servo-control. The source consists 30 of the endmost face of an optical fiber 17 placed in a conjugate plane of tlle pupil of the photoreducing lens. The field illuminated is limited by field diaphragm 18. The dif-fraction figurcs obtained by projection of mark 7 of the reticle on mdrk 8 of the wafer yives rise, through 35 mdrk 7, to a new image which the microscope 10 (lenses 27 and 28) forms on a detector. This detector may be formed by the endmost faces, evenly spaced apart along concentric circle~, of a burldle of optical fibers 29. These endmost faces are placed in a conjugate plane of the input pupil 30 of the photoreducing lens 4. Thus, analysis may be made in the Fourier plane. The faces at one end of the fibers are 5 situated at the location of the di~fraction figure formed by the patterns of the reticle and of the wafer, formed aclvant-ageously this time by linear constant pitch gratings. The photoelectric analysis rneans situated at the other end of the fibers are fixed with respect to the machine.
The alignment device ~hich has just been described in which the rays always travel along the same path in micro-scope 10, allows an analysis to be made, giving an error signal, in the plane of the pupil of the photoreducing lens at any point in the field.
Whatever the embodiment adopted, the switchable optical means (retractab~e or "flipping" mirror 15 for e~ample) allows the resin to be given only a very small exposure during alignment, all the more so since alignement may be carried out in a very short time. No movement of the wafer is 20 necessary between alignment and exposure, which increases the operating rate and gets round the problems of accurate move-ment of the wafer over short distances. The fact of having to place the observation microscope on a table for moving it in both directions does not constitute in practice a real 25 restriction, nn more than the addition of the field lens 14.
Numerous other variations are possible and selection among the possible embodiments will be made depending upon the particular condi-tions of uses.
Referring to Figure 6, -there is shown a modified embodiment for operation With polarized light. The use of polarized light overcomes the problem of the direct back scattering of light by the optical surfaces of the micro-scope. For simplicity, those elements in Figure 6 whiCh correspond to elements already shown in Figures 1 and 2 are designated by the same reference numerals.
The incident ligh-t from optical fiber 17 is polarized by a polarizing element 31 which reflects that light whiCh ~3~3~1 is polarized perpendicularly to the direction o~ polariz-ation impressed by -the elemen-t. The direct light (indicated by single arrows) passes -through a quarter wave strip and has a circular righ-t hand polarization at the output of strip 32. Reflection on wafer 2 results in a left hand cir-cular polariza-tion of the reflected light indicated by double arrows in Figure 6, wi-th polarization of the reflect-ed light being indicated in dashes. After the reflec-ted ligl1t has passed through the quarter wave strip 32, it has a linear polarization directed in the plane of incidence on mirror 15. That polarization direction is passed by the element 31 to the alignment device 33, which may have the same construction as in Figure 2.
The ~uarter wave strip 32 may be placed before the microscope 10. Then, that light which is back scattered by the optical surfaces, which have a perpendicular polar-ization, is reflected toward fiber 17. ~ quarter wave s-trip 34 may as well be located between reticle 1 and projection objective lens 4. Then, that light which is back scattered by the reticle, which is quite troublesome for alignment, is eliminated.
That approach provides an increased overall gain and increases the contras-t offered by certain types of details, which are more clearly seen in polarized ligh-t 25 than in natural light.
Referring to Figure 7, an image analysing system which may be implemented in the device of Figure 2 will be described. That system used slit scanning. It comprises a scanning mirror 35 which is re~lecting on both surfaces.
The angular deviation o~ mirror 35 is con-trolled by a mechanism 36, which may be of the type used in ammeters.
That mechanism receives a saw tooth signal from a generator 37. On each channel I or Il, there is located an image rotation element 38 or 39 and a pass-banding mirror 40 or 41 for image displacement. Elements 38 and 39 may for instance be Dove or Wollaston prisms. The double-face rotating mirror 35 moves the image of each channel over a corresponding slit which is perpendicular to the analysis 3~

direction. That sli-t constitutes ~he input opening of an associated photomultiplier tube 42 or 43. The ammeter 36 may be provided with a position transducer 44 connected to a conventional sampling and processing circuit 45.
The image analyzing provides satisfactory results when the wafer has a substantial con-trast, providing a large phase signal. On the other hand, difficulties may be encoun-tered in other situations. Then, the grating analysing system which will now be described with reference to Figures 8-9 may be preferred.
Figure 8 illustrates -the approach which is used in a diffraction grating analysis device. It will first be assumed that alignment is to be made along a single direct-ion x. That alignment is made by equilizing the amounts of light received by two zones o-f the semiconductive wafer coded with gratings having different directions for differentiating between them. Figure 8 shows two adjacent zones A and B of the silicone wafer, having a common edge 46. gratings consisting of lines at an angle with respect to the common edge 46 are printed on zones A and B~ When the gratings receive a parallel light beam, each grating exhibits a diffraction pattern along two directions in a plane perpendicular to the direction of the gra-ting lines, and particularly along the first order directions indicated by IA for~zone A. It the first order diffracted ligh-t beams are collected by an optical system, there is obtained a series of light spots distributed about a central spot corresponding to the zero order. The relative posi-tions of the spots are conditioned by the inclination o-f the grating lines (Figure 9).
If a transparent slot 47 is formed in the reticle and the image of tha-t slot is projected on zones A and B
parallel to the direction of the common edge 46, the amounts of light IB and IA collected will be in direct relation with the light areas. In the situation illustrated in Figure 8, IB> IA- Then, the wafer will be moved until the amounts of light are equal.
If alignment is to be achieved along two mutually ~ 3~

orthogonal directions, four di-rfraction zones should be provided and they may be distributed as indicated in Figure 10. The diffraction diagram will then be of the type illustrated in Figure 11. Rather than a slot, -the reticle 1 should then include a -transparent cross 48. The signal IB-IA will provide the alignment error along direction x while the signal IC-ID will provide the alignment error along direction y.
That diffraction grating alignment sys-tem may be located in the "step and repeat" device as schematically indicated in Figure 12. In Figure 12, those elements which correspond to those in Figure 5 are designated by the same reference numerals. The image of the end surface of fiber 17 is formed at the object focus of the projection objective lens 4 by an additional objective lens 27 and a lens 14. The optical system is designed for the image of the field diaphragm 18 to be formed on reticle 1 and lighting the reference mark 7, consisting of a transparent cross. The image of -the cross is formed on the wafer (not shown) by the projection objective lens 4. The diffrac-ted orders are collected at the focus of lens 4 and light is again collected, through reticle 1, by lens 14 and objective lens 27 and formed on diaphragm 49. The last lens 28 forms a new image at an infinite distance and, at the same time, forms an image of -the focal plane of the project-ing objective lens 4 on the end surfaces oF the optical fibers of a bundle 29. The end surfaces are distributed for corresponding to the diffraction diagram of Figure 11. For instance, the lines of gratings A and B may be printed at angles of 22.5 and -22.5 with respect to direction x. The lines of grating C and D may be a-t 67.5 and -67.5 with respect to the same direction. A second set of fibers may be provided for receiving the diffracted beams corresponding to orders +3 and -3, for anticipating a situation where the interferences could be destructive for -the first order.
The detection of the devia-tion error may be made by sensing the amounts of light delivered by those fibers which receive IA and IB, with a photomul-tiplier tube 50 through a chopper disk 52. The output signal of the pho-tomul-tiplier tube 50 is applied to a phase lock demodulator 53 which also receives the references signal from the motor which drives the disk. The al-terna-ting signal delivered by the photo-multiplier tube, which has a value proportional to thealignment error, is subjected to phase lock detection and gives way to an analog error signal which may be injected into a servoloop for table 6. The signals corresponding IC and ID may be subjected to the same processing for obtaining an error signal for control along direction y.
The alignement system may be designed for making it possible to use image analysis or grating analysis at will.
For that purpose, those elements which are necessary to the two operating modes and switch means should be provided.
Referring to Figure 13, a possible hybrid arrangement is shown: a quarter wave strip 54 may be moved between a position where analysis occurs by diffraction and a position where there is image analysis, using a polarizing element 55.

Claims (12)

- 15 -

1. In an apparatus for printing integrated circuits, an alignment device having:
a holder for holding a reticle provided with a refer-ence mark, means for moving said holder along coordinate mutually orthogonal axes, a table for receiving an object provided with at least one alignment mark, a projection lens between said holder and table for projecting an image of said reticle on said object, means for illuminating the reticle with a beam of actinic light for exposure of said object, means for moving said table along directions conjugate of said coordinate axes, and means for simultaneous observation of said alignment mark and reference mark and determination of lack of coincidence between the marks, comprising a light source at the same wavelength as said illuminating means and optical means switchable between a first condition where it clears said beam and a second condition where it directs a narrow light beam from said light source to said reference mark and object and it brings light reflected by the object back to detection means.

2. An alignment device according to claim 1, wherein said optical means is arranged to pass a central ray of said light beam through the object focus of said projection lens in said second condition.

3. An alignment device according to claim 2, wherein said optical means include a sighting mirror displaceable between a position inside said beam from said illuminating means and a position outside said beam.

4. An alignment device according to claim 2, further comprising a field lens interposed between the illuminating means and said reticle, close to said reticle, the focal length of said field lens being substantially equal to the distance between said reticle and object focus.

5.An alignment device according to claim 2, wherein said detection means includes a microscope and holding means carrying said microscope for movement of said micro-scope along directions conjugate of said coordinate axes.

6. An alignment device according to claim 5, wherein said light source is a first end face of a flexible optical fiber, said first end face being connected to said micro-scope for movement therewith while the other end face of said fiber is arranged to receive light from said illuminat-ing means.

7. An alignment device according to claim 2, wherein said light source is an end face of an optical guide located to collect light from said illuminating means.

8. An alignment device according to claim 2, wherein said optical means comprises a beam separator for reflecting said narrow light beam from said light source to said switchable optical means and transmitting light reflected by said object to analysing means.

9. A device according to claim 6, wherein said analys-ing means comprises a separator for splitting the reflected beam into two paths, each comprising a spatial filter and a detector, each path being assigned to a particular direction of the patterns.

10. A device according to claim 2, wherein said alignment mark comprises at least a pair of adjacent gratings whose lines are symetrically directed with respect to a common edge of said two gratings, said reference mark comprises a transparent slot whose direction is conjugate of the direction of said common edge, and said detection means comprises means for collecting the individual amounts of light received by said gratings from said light source and diffracted by said gratings and for comparing said individual amounts.

11. A method for aligning a semiconductive wafer provided with an alignment mark and a reticle provided with a reference mark prior to photoprinting a circuit pattern on said wafer, comprising the steps of:
(a) locating said reticle on a holder, (b) locating said wafer on a table, (c) illuminating a zone of said reticle including said pattern and said reference mark with a beam of actinic light of low intensity, (d) forming a reduced image of said reticle on said wafer with a photoreduction projection lens, (e) locating an optical element in said beam for forming an observation path for simultaneous visual obser-vation of said alignment mark through said photoreduction projection lens and of said reference mark, (f) moving said holder along coordinate directions for coincidence of the observed images of said marks, (g) moving said optical element for clearing said path, (h) increasing the intensity of said actinic light for printing said pattern.

12. A method according to claim 11,further comprising repeating the steps (c)-(h) after said wafer has been advanced by steps each corresponding to the distance bet-ween adjacent integrated circuits to be printed on said wafer.