JPH1050589A - Alignment method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a high-accuracy alignment method not only taking the positional deviation of a projection optical system from alignment light having a wavelength which is different from that of exposure light caused by a chromatic aberration, but also the plotting error of marks on a reference plate, etc., into considerations. SOLUTION: A base line amount is found by means of a TTR alignment system 14 for exposed light wavelength and another TTR alignment system 12 for different wavelength. In addition, a reference mark (for example, Fu) for different wavelength is divided into a plurality of divisions, and the correcting value of the base line amount is computed by using the manufacturing error of the mark Fu at every division, in according to the irradiating position of an alignment beam having a different wavelength on a reference plate FP. After computing the correcting value, the positional deviation between a mark Au on a mask R and a mark on a photosensitive substrate W corresponding to the mark Au is detected, by using the alignment system 12 and the mask R is aligned with the substrate W by using the base line amount and correcting value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an alignment method, and more particularly to an alignment method of a mask and a photosensitive substrate used in an exposure apparatus having an exposure wavelength TTR alignment system and another wavelength TTR alignment system.
The alignment method according to the present invention provides a semiconductor device (IC)
And a projection exposure apparatus used for manufacturing a liquid crystal substrate, a thin film magnetic head, and the like.

[0002]

2. Description of the Related Art With the increase in the capacity of semiconductor devices and the miniaturization of patterns, the demand for higher resolution of projection exposure apparatuses has become more stringent each year. To meet such demands, the field of projection optical systems has been widened and the resolution has been increased. Furthermore, development of a high-precision alignment device is indispensable to meet such demands.

As one of the highly accurate alignment methods,
A so-called die-by-die alignment method has been proposed in which a mark on a reticle and a mark in each shot area on a photosensitive substrate such as a wafer or a glass plate (hereinafter, referred to as a “wafer”) are sequentially detected to perform alignment. ing.
For example, a method of simultaneously detecting a reticle mark and a wafer mark using illumination light having a wavelength different from that of exposure light, that is, a so-called another wavelength TTR alignment method has been proposed.

A conventional projection exposure apparatus of this type includes an exposure wavelength TTR alignment system for detecting the position of a mark provided at a predetermined position in a field of view of a projection optical system outside a circuit pattern area on a reticle by exposure light. A different wavelength TTR alignment system for detecting the position of a mark provided around the circuit pattern area on the reticle with light having a wavelength different from the exposure light.

When the reference plate fixed on the wafer stage is positioned below the projection optical system, the exposure wavelength TT
An R alignment system detects a displacement between a mark on the reticle and a mark on the reference plate (corresponding to a mark on the reticle). Further, the misalignment between the mark on the reticle and the mark on the reference plate is detected by another wavelength TTR alignment system. At this time, the difference between the displacement amount detected by the exposure wavelength TTR alignment system and the displacement amount detected by the other wavelength TTR alignment system is the difference TT
This is the amount of positional deviation due to chromatic aberration at the position of the mark on the reticle detected by the R alignment system.

Therefore, when aligning a reticle and a wafer using a different wavelength TTR alignment system, if the alignment position is corrected based on the detected positional deviation amount, alignment caused by chromatic aberration at the alignment wavelength (different wavelength) can be achieved. Accurate alignment can be performed without any error. An exposure apparatus that has an exposure wavelength TTR alignment system and another wavelength TTR alignment system and that performs alignment by correcting the amount of misalignment between a reticle and a wafer due to chromatic aberration is disclosed in
-45512.

[0007]

However, according to the above-mentioned technique, the mark on the reticle for the exposure wavelength TTR alignment system and the mark on the reticle for the exposure wavelength TTR alignment system are caused by, for example, a drawing error of the mark on the reference plate or a detection error of the exposure wavelength TTR alignment system. If an error occurs in the detected value of the positional relationship with the mark on the plate, an error such as offset or rotation occurs between the reticle and the wafer. Replacement) according to different wavelength T
When the position of the TR alignment system is changed, there is an inconvenience that the offset, rotation, and the like change depending on the position of the alignment system.

The present invention has been made in view of the inconveniences of the prior art, and has as its object to provide only a position shift caused by chromatic aberration of a projection optical system with respect to light (alignment light) having a wavelength different from that of exposure light. Another object of the present invention is to provide a high-precision alignment method that also takes into account drawing errors of marks on a reference plate.

[0009]

Means for Solving the Problems Normally, a reference plate is prepared by forming a mask with an electron beam lithography apparatus, a photo repeater, or the like.
The mask is made by reducing and projecting the mask on a substrate. By the way, the mark on the reference plate in question here is a sensor of another wavelength TTR alignment system (alignment microscope)
Is formed over a very wide range because the entire movable range of the above is covered. When creating a mask, the electron beam lithography system divides the inside of the mask into several sections and draws them one by one in order, but the mark does not fit in one section and is drawn in several sections. You. Also, in the photo repeater, several types of parent masks are created for each pattern on the mask, and the parent masks are reduced and projected onto necessary portions, respectively, to create masks. However, since the marks do not fit in the parent masks, When the mask is created by reducing and projecting the parent mask, several shots are added to form one mark. As described above, since the mark cannot be collectively described by any of the methods, the mark is divided into several sections (see FIG. 7), and discontinuity of the mark occurs at the boundary between the sections. For this reason,
When the manufacturing error of the reference plate is measured, the error gradually changes in one section, and changes abruptly in the discontinuous portion of the seam between the sections.

The present invention has been made by paying attention to the process of forming such marks and the characteristics of the marks, and employs the following method.

According to the first aspect of the present invention, there is provided a semiconductor device comprising:
Exposure light is applied to the first mark area where the mark is formed and the second mark area where the second mark is formed on the photosensitive substrate or reference plate corresponding to the first mark via the first mark area and the projection optical system. A first detection method of irradiating an exposure wavelength alignment beam having the same wavelength and photoelectrically detecting reflected light generated from the first mark area and the second mark area to detect a positional shift between the first mark and the second mark. A third mark area on the mask on which the third mark is formed, and a fourth mark on the photosensitive substrate or reference plate corresponding to the third mark via the third mark area and the projection optical system. The fourth mark area is irradiated with another wavelength alignment beam having a different wavelength from the exposure light, and the third mark area and the fourth
The mask and the photosensitive substrate are used in an exposure apparatus having a second detection system that detects a displacement between the third mark and the fourth mark by photoelectrically detecting reflected light generated from a mark area. An alignment method, wherein after positioning the reference plate below the projection optical system, the first detection system positions the first mark on the mask and the corresponding second mark on the reference plate. A first step of detecting a shift amount; a second step of detecting, by the second detection system, a position shift amount between a third mark on the mask and a corresponding fourth mark on the reference plate. A third step of calculating a difference between the displacement amounts obtained in the first and second steps as a displacement amount due to chromatic aberration; and dividing the fourth mark on the reference plate into a plurality of sections, The criteria in the second step The correction value of the displacement amount due to chromatic aberration calculated in the third step is calculated using the manufacturing error of the fourth mark for each section according to the irradiation position of the different wavelength alignment beam on the fourth mark. A fourth step of calculating; and then, using the second detection system, a third mark on the mask and a fourth mark on the photosensitive substrate corresponding thereto.
And a fifth step of detecting a positional deviation from the mark and performing alignment between the mask and the photosensitive substrate using the detection result and the values obtained in the third and fourth steps.

According to this, in the third step, the first,
The difference between the displacement amounts obtained in the second step is calculated as the displacement amount due to chromatic aberration. That is, a so-called baseline amount of the second detection system is obtained. In the fourth step, the fourth mark on the reference plate is divided into a plurality of sections, and the fourth mark for each section is divided according to the irradiation position of the different wavelength alignment beam on the fourth mark on the reference plate in the second step. A correction value of the amount of misregistration due to chromatic aberration calculated in the third step is calculated using the mark manufacturing error. Thereafter, in the fifth step, the third mark on the mask and the third mark on the mask are detected using the second detection system. Is detected, and the mask and the photosensitive substrate are aligned using the detection result and the values obtained in the third and fourth steps.

Therefore, the manufacturing error of the fourth mark (reference mark) on the reference plate used in the baseline measurement performed prior to exposure is corrected, and the position of the third mark (alignment mark) on the mask is corrected. Even when the position of the exposure light is changed, it is possible to suppress the change of the overlay error, thereby not only the positional shift due to the chromatic aberration of the projection optical system with respect to light (alignment light) having a wavelength different from the exposure light, but also the mark on the reference plate. High-precision alignment that also takes into account the drawing error and the like.

According to a second aspect of the present invention, in the alignment method according to the first aspect, the plurality of sections are determined according to a manufacturing step of a fourth mark on the reference plate. According to this, if the manufacturing process of the reference plate is known, it is not necessary to measure the number and arrangement of the sections of the fourth mark on the reference plate.

[0015] The third aspect of the present invention is the first or second aspect.
3. The alignment method according to claim 1, wherein the manufacturing error of each section of the fourth mark on the reference plate is a plurality of pieces having a predetermined positional relationship with the reference point selected based on the predetermined reference point of each section. Is an error obtained by calculation based on the deviation of the measurement point from the design value.

According to this, the manufacturing error of each section of the fourth mark on the reference plate is determined by a plurality of measurement marks having a predetermined positional relationship with the reference point selected based on the predetermined reference point of each section. Since it is an error calculated by the calculation based on the deviation of the point from the design value, for example, if at least three measurement points including the reference point are set in one section, the reference point in the section is set for each section. The rotation error, magnification error, and offset can be determined based on the reference mark, and the manufacturing error can be calculated based on less data. Since it is hardly affected by drift of a sensor for measurement at the time, and furthermore, it is based on the characteristic of the manufacturing error of the reference mark, high-precision correction is possible even with a small amount of data.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an exposure apparatus according to an embodiment to which the alignment method according to the present invention is applied will be described.
A description will be given based on FIGS. 1 to 9.

FIG. 1 schematically shows a configuration of a reduction projection type exposure apparatus (stepper) 10 as an exposure apparatus according to one embodiment. The reduced projection type exposure apparatus 10 includes an XY stage ST that moves in a two-dimensional direction (X, Y directions) in a reference plane orthogonal to the paper surface of FIG. 1, and an optical axis AX that is disposed above the XY stage ST. A projection lens PL as a projection optical system substantially orthogonal to the reference plane, a reticle R as a mask disposed above the projection lens PL, a fine movement adjustment system for the reticle R, and an XY stage movement control system. , Another wavelength TTR as the second detection system
(Through the reticle) An alignment system 12, an exposure wavelength TTR alignment system 14 as a first detection system, a dichroic mirror DM inclined at 45 ° above the reticle R, and a main controller 16 are provided. I have.

The XY stage ST is moved in the X and Y directions in a reference plane by a motor 18. On the XY stage ST, a wafer W as a photosensitive substrate and a reference plate FP (for this, , Which will be described later). The coordinate position of the XY stage ST is sequentially measured by the laser interferometer 20. The positioning of the XY stage ST is performed by monitoring a measurement value of the laser interferometer 20 and driving a motor 18 by a stage driver circuit 22.
Done by That is, in the present embodiment, the motor 1
8, laser interferometer 20, and stage driver circuit 22
Thus, an XY stage movement control system is configured.

The stage driver circuit 22 includes the main controller 1
6 to control the movement and positioning of the XY stage ST.

When the exposure light from the light source of the exposure illumination system (not shown) is reflected by the dichroic mirror DM and travels along the optical axis AX, the projection lens PL illuminates the reticle R with the exposure light. Is projected onto the wafer W. The dichroic mirror DM reflects the exposure light by about 90% or more and directs the exposure light to the pattern area PA of the reticle R (see FIG. 5). Here, as the exposure light, for example, an i-line (wavelength 365 nm), an excimer laser (wavelength 248 nm, 193 nm, etc.), a harmonic of a YAG laser (200 nm or less), or the like is used.

The reticle R is held on a reticle stage RS that moves minutely in two dimensions (X, Y, and θ directions), and the reticle stage RS is positioned by a driving unit 24. That is, in the present embodiment, the reticle stage RS and the drive unit 24 constitute a reticle fine movement adjustment system. Drive unit 24 also controls the movement of reticle stage RS based on a command from main controller 16.

Next, another wavelength TTR alignment system 12 will be described with reference to FIG. This different wavelength TTR alignment system 12 is actually constructed with four eyes (the number of objective lenses) above the dichroic mirror DM (see FIG. 5), but in FIG. 1, one of them (the objective lens OB)
Only the system corresponding to Ju) is shown.

In FIG. 1, another wavelength TTR alignment system 12 includes a driving section 26, an objective lens OBJu, and a mirror M.
1, laser light source 28, 2 luminous flux frequency shifter unit 3
0, a beam splitter 32, a condenser lens (inverse Fourier transform lens) 34, a reference grating 36, a photoelectric element 38, a beam splitter 40, a condenser lens system 42, a beam splitter 43, a reticle reference grating plate 44, a field stop 46, Photoelectric element 48, mirror 50, condenser lens 52, photoelectric element 5
4 and a phase difference measurement unit 56.

Here, the different wavelength TTR alignment system 12 will be described together with the operation of each component.

The telecentric objective lens OBJu and the mirror M 1 are held by a hardware 58 and moved in the X and Y directions by the drive unit 26. The objective lens OBJu is arranged so as not to interfere with the dichroic mirror DM and so that its optical axis is perpendicular to the reticle R.

Illumination light for another wavelength alignment (another wavelength alignment beam) is emitted as a linearly polarized beam LB from a laser light source 28 such as helium-neon or argon ion, and is incident on a two-beam frequency shifter unit 30. The two beam splitting frequency shifter unit 30 splits the beam LB into two, and applies two high-frequency modulations (frequency shifts) to each of the two split beams.
OM (acousto-optic modulators) 60A and 60B, small lenses 62A and 62B for converging the frequency-shifted beams, and the like are provided. The AOM 60A is driven at a frequency f1 (for example, 80 MHz) and its first-order diffraction beam LB
1 is taken out through the small lens 62A and the AOM 60B
Is driven at a frequency f2 (for example, 80.03 MHz), and a first-order diffracted beam LB2 is extracted through a small lens 62B. The principal rays of the two beams LB1 and LB2 are determined in parallel and symmetrically with the optical axis AXa of the alignment system, and are split into two by the beam splitter 32. Split beam Lr that has passed through beam splitter 32
1 and Lr2 are converted by the condenser lens 34 into parallel light beams that intersect at the rear focal plane. A reference grating 36 is disposed on the rear focal plane, where interference fringes having a pitch corresponding to the intersection angle and wavelength of the two beams Lr1 and Lr2 are formed. This interference fringe has a frequency difference Δf (3) between the two beams Lr1 and Lr2.
0KHz) in one direction on the reference grid 36. On the reference grating 36, a transmission type diffraction grating having a pitch equal to the pitch of the interference fringes is provided in parallel with the fringes. Accordingly, from the reference grating 36, the 0th-order light of the beam Lr1 traveling in the same direction as the beam Lr1 and the 1
An interference light DR1 with the next light and an interference light DR2 with the zero-order light of the beam Lr2 traveling in the same direction as the beam Lr2 and the primary light generated from the beam Lr1 are generated. , DR2. Here, the interference light DR1
, DR2 change in intensity in the form of a sine wave at the frequency difference Δf (30 KHz), that is, the beat frequency, and the output signal SR from the photoelectric element 38 becomes an AC signal of 30 KHz. This signal SR becomes a reference signal for phase comparison at the time of alignment. From the beams LB1, LB2 and the lens 34, the photoelectric device 3
The state of each beam up to 8 is described in JP-A-4-455.
Since the contents corresponding to FIG. 3 of Japanese Patent Publication No. 12 and the description thereof are disclosed in detail, detailed description is omitted here.

On the other hand, the two beams Lm1 and Lm2 reflected by the beam splitter 32 are further added to the beam splitter 4
0 and is reflected by the objective lens OBJu via the mirror M1.
Incident on. The objective lens OBJu has two beams Lm1,
Lm2 is converted into a parallel light beam that intersects at plane IP in the space.
The surface IP is separated from the reticle R in the direction of the optical axis AX of the projection lens PL by an amount of axial chromatic aberration ΔL, and is conjugate with the wafer W or the reference plate FP under the wavelength (λZ) of the different wavelength alignment beam LB. Plane. The two beams Lm1 and Lm2 intersecting at the plane IP form the third mark Au on the reticle R.
In the third mark area where is formed, the light passes through the areas of the marks Aua and Aub separated from each other (see FIGS. 2 and 4), becomes a beam waist at the pupil EP of the projection lens PL, and then becomes the reference plate FP (or the wafer W). ) Above, the beams again become parallel light beams and intersect.

The interference light BT is generated in the vertical direction from the area on the reference plate FP where the reference mark Fu as the second mark is formed (see FIG. 4). The interference light BT is vertically generated from the reference mark Fu by irradiation of the beam Lm1.
The second order diffracted light interferes with the first order diffracted light generated perpendicularly from the reference mark Fu by the irradiation of the beam Lm2,
In the image space, it is a parallel light beam. After the interference light BT converges on the beam waist at the center of the pupil EP of the projection lens PL, it passes through the central transparent portion (between the marks Aua and Aub) in the mark area of the reticle R as a parallel light flux, and It moves backward along the optical axis AXa of the lens OBJu. The interference light BT converges again at the beam waist at the front focal plane of the objective lens OBJu, that is, at the center of the plane conjugate with the pupil EP of the projection lens PL, and goes to the light receiving system in the order of the mirror M1 and the beam splitter 40 in FIG. Return.

Here, the structure of the mark area on the reticle R where the mark Au is formed will be described with reference to FIGS. FIG. 2 is a diagram illustrating a mark area on the reticle R, and FIG. 3 is a diagram illustrating a mark on the wafer W. As shown in FIG. 2, a light-shielding band ESB having a fixed width is formed on the outer periphery of the circuit pattern area PA of the reticle R. A transparent window is provided in a part of the light-shielding band ESB, and two grid marks Aua and Aub are formed there. They are arranged with the light-shielding band ESB 'interposed therebetween.
The pitches of the marks Aua and Aub are equal to each other, and the width in the direction orthogonal to the pitch direction is set to about half of the transparent window.
The width of the light-shielding portion ESB 'is also set to about half of the transparent window, and the lattice mark WMu on the wafer W is located at the portion of the mark area WMU' under the wavelength for alignment. 2
The beams Lm1 and Lm2 individually illuminate each of the grid marks Aua and Aub in a rectangular shape, but the beams Lm1 transmitted through the transparent portion adjacent to the mark Aua on the pattern area PA side.
Irradiates the wafer mark WMu as shown in FIG. 3, and the beam Lm2 that has passed through the transparent portion adjacent to the mark Aub on the pattern area PA side irradiates the wafer mark WMu. The interference light BT generated vertically from the mark WMu returns to the objective lens OBJu on the reticle R through the portion WMu 'in FIG.

Here, the generation of the first-order diffracted light from the reticle grating marks Aua and Aub will be described with reference to FIG. In FIG. 4, beams Lm1 and Lm2 irradiate reticle grating marks Aua and Aub, respectively, and beams Lm1 (Lm
The mark Aua (Aub) such that the diffraction angle of the first-order diffracted light with respect to the zero-order light D01 (D02) becomes exactly 2θ 'with respect to the angle of incidence θ' on the grating mark Aua (Aub) of m2).
Determine the pitch of As a result, the optical path of the beam Lm1 is directly transmitted from the grating mark Aua to the reverse first-order diffracted light D11.
(Frequency shift f1) occurs, and the first-order diffracted light D1 that travels backward from the grating mark Aub along the optical path of the beam Lm2 as it is.
2 (frequency shift f2) occurs. Therefore, these first-order diffracted lights D11 and D12 also return to the light receiving system via the objective lens OBJu together with the interference light BT.

Returning to FIG. 1, the interference light BT and the first-order diffracted lights D11 and D12 are converted by the beam splitter 4
0 is transmitted to the condenser lens system 42. Lens system 4
Reference numeral 2 denotes an inverse Fourier transform lens, which converts both the interference light BT and the first-order diffracted lights D11 and D12 into parallel light beams, and crosses them at the focal plane (image conjugate plane).
Each beam passing through the lens system 42 is converted into a beam splitter 43
The transmitted light reaches the reticle reference grating plate 44, and the reflected light reaches the field stop 46. The reference grating plate 44 and the field stop 46 are both in the plane IP (reference plate F
P or a conjugate plane with the wafer W). Therefore,
In the reference grating plate 44, the first-order diffracted lights D11 and D12 intersect, and interference fringes are formed in the intersection area. The interference fringes naturally flow one-dimensionally at the beat frequency Δf (30 KHz).

A transmission type diffraction grating 44A is provided on the reference grating plate 44 covered with the chromium layer. This grid 4
The photoelectric element 48 receives the interference light BTr of the diffracted light generated coaxially from 4A (each zero-order light BT0 of the first-order diffracted lights D11 and D12 advances so as to deviate from the photoelectric element 48). At this time, the interference light B from the wafer mark WMu can be obtained simply by matching the position and size of the grating 44A on the reference grating plate 44 with the size of the mark Aua or Aub.
T can be shielded from light. The interference light BTr thus received by the photoelectric element 48 has a beat frequency Δf (30 kHz).
z), the output signal Sm of the photoelectric element 48 has a linear phase difference with respect to the reference signal SR according to the amount of displacement of the reticle marks Aua and Aub in the pitch direction with respect to the reference grating 36. The AC signal changes to

The interference light BT that has passed through the field stop 46 reaches the photoelectric element 54 via the mirror 50 and the condenser lens 52. The photoelectric element 54 generates an output signal Sw corresponding to a change in the intensity of the interference light BT. This signal Sw also has a beat frequency Δf
The signal becomes a sine wave-like AC signal, and the wafer W
Upper mark WMU or reference mark Fu on reference plate FP
Signal SR in proportion to the amount of displacement of the reference
Changes with respect to.

The above-described reference signal SR and output signals Sm, Sw
Are input to the phase difference measurement unit 56, and the phase difference measurement unit 56 outputs the output signal Sm with respect to the reference signal SR.
And the phase difference φw of the output signal Sw with respect to the reference signal SR are obtained, and the difference Δφ = φm−φw is calculated. In the case of the present embodiment, the pitch of the diffraction grating on the reference plate FP (or the wafer W) is twice as large as the pitch of the interference fringes formed thereon.
180 °) corresponds to 回 折 of the diffraction grating pitch (± １ ／ pitch). Based on the phase difference Δφ, the phase difference measurement unit 56 calculates the position correction amounts (position shift amounts) ΔX and ΔY of the wafer stage ST or the reticle stage RS, and sends the values to the main controller 16.

Next, the exposure wavelength TTR alignment system 14
Will be described. The exposure wavelength TTR alignment system 14 is actually provided with two systems corresponding to the objective lenses OBr and OBI (see FIG. 5), but in FIG. 1, a system corresponding to the objective lens OBr (the first system on the reticle R). Only a system for detecting a reticle mark RMr as a mark) is shown.

This exposure wavelength TTR alignment system 14
Consists of a mirror M2, an objective lens OBr, a beam splitter 64, a lens system 66, an illumination field stop 68, a condenser lens 70, a fiber 72, an imaging lens 74, a beam splitter 76, and CCD image sensors 78A and 78B.

This will be described in more detail.
Emits illumination light of an exposure wavelength (exposure wavelength alignment beam), and uniformly irradiates an aperture 68 via a condenser lens 70. Illumination light passing through the aperture of the aperture 68 is passed through a lens system 66 and a beam splitter 64 to an objective lens OBr.
Then, after being vertically bent by the mirror M2, the local area including the third mark RMr of the reticle R is irradiated. The mirror M2 is fixedly arranged so as to be located outside the size of the largest pattern area PA assumed on the reticle R.

The aperture 68 is conjugate with the reticle R, and an aperture image of the aperture 68 is formed on the reticle R. Further, the beam splitter 64 serves as a front focal plane of the telecentric objective lens OBr, that is, the pupil E of the projection lens PL.
It is arranged near a plane conjugate with P, and reflects a part of the light returning from the objective lens OBr toward the imaging lens 74. The light receiving surfaces of the CCD imaging devices 78A and 78B are conjugated to the reticle R by the objective lens OBr and the imaging lens 74, and are also conjugated to the wafer W or the reference plate FP via the projection lens PL. Note that the exit end of the fiber 72 is arranged conjugate with the pupil EP of the projection lens PL to perform Koehler illumination.

The CCD elements 78A and 78B are rotated by 90 ° with respect to each other so that the horizontal scanning lines are in the X direction and the Y direction, respectively.
The position measurement in the direction and the position measurement in the Y direction are separately performed. This is because ordinary CCD elements have different pixel resolutions in the horizontal and vertical directions, so that a single CCD element avoids a difference in resolution when detecting a shift in both the horizontal and vertical mark images. is there. The image processing unit 80
An image signal (video signal) is input from each of the CCD elements 78A and 78B, and a reference mark FMr as a second mark on the substrate plate FP and a mark RM on the reticle R
The position shift amount with respect to r is detected, and information on the position shift amount is sent to the main controller 16.

In addition to the above configuration, an off-axis global mark detection system 82 for detecting a global alignment mark on the wafer W, a latent image in the resist layer, or each reference mark on the reference plate FP, A processing unit 84 is provided. Further, an input unit 17 such as a keyboard and a bar code reader for inputting information on a mark position (mark position for another wavelength TTR alignment) on the reticle R to the main controller 16 is provided. Objective lenses OBJu, OBJ
Main controller 16 controls drive unit 24 based on positional information from input unit 17 so that reticle alignment marks Au, Ar, Ad, and Al as third marks are included in the fields of view of r, OBJd, and OBJl. .

Incidentally, another wavelength TTR alignment system 1
2 and the exposure wavelength TTR alignment system 14 are preferably arranged as schematically shown in FIG. In the reticle R, as shown in FIG. 6, reticle lattice marks Au, Ar, Ad, and Al as third marks are formed on each piece of the light-shielding band ESB around the pattern area PA. The marks Au and Ad are used for alignment in the X direction in FIG. 6, and the marks Ar and Al are used for alignment in the Y direction. Further, a similar reticle mark RMl is provided outside the light-shielding band ESB at a position symmetric to the reticle mark RMr as the first mark. Therefore, two exposure wavelengths T
When the objective lenses OBr and OBI of the TR alignment system 14 have the marks RMr and RMl under the dichroic mirror DM.
OBJu, OBJd, OBJ of the four different wavelength TTR alignment systems 12
r and OBJl are arranged so as to detect marks Au, Ar, Ad and Al via the dichroic mirror DM, respectively.

Next, the mark arrangement of the reticle R shown in FIG. 5 and the mark arrangement on the reference plate FP suitable for this embodiment will be described with reference to FIGS. FIG. 6 shows a pattern arrangement of a preferred example of the reticle R. This figure 6
As shown in FIG. 7, two cross-shaped reticle marks RMr and RMl are provided on a line passing through the center Rcc of the pattern area PA of the reticle R and parallel to the X axis. In the reticle R shown in FIG. 6, four marks A
u, Ar, Ad, and Al are all disposed at positions substantially the same distance from the reticle center Rcc, and rectangular regions Su, Sr, Sd, and Sl shown by broken lines in FIG. Objective lens OBJu,
An example of the movable range of the optical axis (detection center) of OBJr, OBJd, and OBJl is shown.

In FIG. 6, marks Au, Ar, Ad,
The positions of Al are indicated by Pud0, Prc0, Pdd0, Plc0, respectively, and the corresponding points on the reference plate FP corresponding to the positions Pud0, Prc0, Pdd0, Plc0 are Fudo, Frc0, Fdd0, Fl0, respectively.
This is indicated by c0 (FIG. 7).

FIG. 7 shows the arrangement of each mark on the reference plate FP. Reference marks FMr and FMl as double cross-shaped second marks are provided on the left and right in the X direction. The distance between the centers of the two marks FMr and FMl is determined to be equal to a value obtained by multiplying the distance between the centers of the marks RMr and RM1 on the reticle by the projection magnification (1 / M). Therefore, the center F of the reference plate FP
When cc is matched with the projection point of the reticle center Rcc, with the reticle mark RMr positioned between the double lines of the mark FMr, they are simultaneously observed by the objective lens OBr of the exposure wavelength TTR alignment system, and the mark FM
With the reticle marks RMl located between the l double lines, they are simultaneously observed by the objective lens OBl. Further, when the center Fcc and the reticle center Rcc coincide with each other on the reference plate FP, diffraction is performed so that the position and size correspond to each of the movable ranges Su, Sr, Sd, and Sl on the reticle R. Reference marks Fu, Fr, Fd, and Fl are provided as fourth marks inscribed in the lattice. For example, the reference mark Fu is the mark WM on the wafer W.
Like u, it is formed from a group of gratings engraved at a constant pitch in the X direction, and is used to detect a relative positional deviation in the X direction from the mark Au on the reticle R. Other fiducial marks Fr, F
The same applies to d and Fl. Therefore, even if the size of the pattern area PA (light-shielding band ESB) changes due to the exchange of the reticle R, each mark Au, A on the reticle.
As long as r, Ad, and Al exist inside the movable ranges Su, Sr, Sd, and Sl, the X-direction shift between the mark Au and the reference mark Fu, the X-direction shift between the mark area Ad and the reference mark Fd, and the mark Ar The displacement of the reference mark Fr in the Y direction and the displacement of the mark Al and the reference mark Fl in the Y direction can be detected.

In the following description, the reference marks Fu, Fr, Fd, and F1 on the reference plate FP indicate that the reference marks Fu, Fr, Fd, and F1 are 4 due to manufacturing problems as described above, as shown in the enlarged view in the circle P in FIG. It is assumed that it is divided into two sections.

Next, a description will be given of the baseline measurement (when it is not necessary to consider the replacement of marks) performed prior to the alignment between the reticle R and the wafer W in the present embodiment.

First, reticle alignment is performed, and two beams Lr1 and Tr2 of another wavelength TTR alignment system are used.
The deviation of the irradiation position of Lr2 and the telecentric error of the two beams Lm1 and Lm2 are checked. When the check is completed, the wafer stage ST is moved to position the reference plate FP below the projection lens PL. Two exposure wavelengths TT
The CCD element of the R alignment system 14 (78 in FIG. 1)
A, 78B only), using the reticle mark RM
r, RMl and the reference marks FMr, FMl are simultaneously detected, and at the same time, marks Au, Ar, Ad, and Al are arranged in each mark area on the reticle R using four different wavelength TTR alignment systems 12. And a relative displacement between each of the reference marks Fu, Fr, Fd, and Fl on the reference plate FP is detected. Marks Au and A on reticle R at this time
r, Ad, Al position (different wavelength TTR alignment system 1
2 objective lenses OBJu, OBJr, OBJd, OBJ
1 are detected as Pud0, Prc0, Pdd0, Pl, respectively.
Let it be c0 (see FIG. 6).

Specifically, the measuring unit 56 (different wavelength T
As a result of detecting four reticle marks Au (reference mark Fu) and Ad (reference mark Fd), Ar (reference mark Fr) and Al (reference mark Fl) by a TR alignment system, the phase difference Δφ between them is all zero. , The positional deviation amounts (ΔXr, ΔYr) detected by the image processing unit 80 (exposure wavelength TTR alignment system), (Δ
Xl, ΔYl) are stored in the main controller 16. Based on the stored result, main controller 16 causes positional deviation C (ΔMx,
ΔMy, ΔMθ) are calculated as follows.

ΔMx = (ΔXr + ΔXl) / 2 ΔMy = (ΔYr + ΔYl) / 2 ΔMθ = ΔYr-ΔYl As a result of aligning the reticle R with the reference plate FP by the different wavelength TTR alignment system 12, the reticle R and the reference plate FP are Has an error (amount of displacement C) in the X, Y, and θ directions according to the chromatic aberration of the projection lens PL under another wavelength. This error is caused by marks Au, Ar, A on reticle R.
As long as the positions of d and Al (positions Pud0, Prc0, Pdd0, Plc0) do not change, it can be considered as a constant positional deviation amount C. This displacement amount C is the baseline amount.

Thereafter, another wavelength TTR alignment system 1
When the shots on the wafer W are aligned in Step 2, the position shifted by the positional shift amount C is set to the true alignment achievement position (the phase difference from the measurement unit 56 is shifted by the positional shift amount C). Main controller 16 controls reticle stage RS or wafer stage ST.

However, even if the displacement amount C (base line amount) due to the chromatic aberration is corrected, if there is a drawing error of the reference pattern on the reference plate FP or a detection error of the exposure wavelength TTR alignment system, etc. When the pattern of the reticle R and the pattern on the wafer W are aligned, there is an alignment error O (offset), so it is necessary to correct the offset. The offset can be corrected by performing test burning and so-called vernier measurement. However, when any one of the different wavelength TTR alignment systems 12 is moved in the X and Y directions to change the alignment position, an error such as displacement or rotation occurs between the projected image of the reticle pattern and the shot area of the wafer. The amount of the error varies depending on the alignment position. This is because the difference between the designed pattern and the actual pattern on the reference plate FP due to the drawing error of the reference plate FP changes depending on the alignment position, and the offset component changes depending on the alignment position. Conceivable.

Therefore, in the case of the above vernier measurement, in order to correct the offset over the entire movable range of the different wavelength TTR alignment microscope, the microscope on the reference plate FP (object lenses OBJu, OBJr, OBJd, O
Difference between the design pattern and the actual pattern over the entire mark of the movable range of BJl) (manufacturing error of the reference plate)
Must be measured and a manufacturing error map of the reference plate FP needs to be created in advance, which is very complicated. Also,
In order to perform high-precision correction, it is necessary to increase the number of measurement points, it takes a long time for measurement, and it is considered that the measured value is easily affected by drift of the sensor and the like.

Therefore, in the present embodiment, the following method is adopted in order to improve such inconvenience. Here, four different wavelength TTR alignment systems 1 in the present embodiment
When the objective lens No. 2 is moved in the X and Y directions (the reticle mark position, that is, the alignment mark position is different between the reticle at the previous exposure and the reticle at the current exposure, or the reticle mark position is changed. The method of correcting the position error amount occurring in the case (2) will be described. Here, the mark Fd on the reference plate FP will be described as an example.

As described above, the mark Fd is divided into four sections A, B, C, and D as shown in FIG. At this time, for example, as shown in FIG. 8, five points Fda0, Fda1, Fda2, Fda3, Fda4
And the distances from Fda0 to the other four points are all equal and Lx (an integral multiple of the mark grid pitch) in the X direction and Ly in the Y direction.

First, deviation Eda of each point from the design value
1. Find Eda2, Eda3, and Eda4. Specifically, it is as follows.

First, the objective lens OBJd (detection center) is adjusted to the position Fda0. (Here, Fda0 may actually be anywhere, and the objective lens OBJd is set near the center of the section A from design data, and that point is considered as a position Fda0.) At that position, another wavelength TTR alignment detection system having OBJd 12, the position difference between the mark Ad and the mark Fd is measured, and the phase difference measuring unit 5 at that time is measured.
The phase difference from No. 6 is set as a reference (reference phase difference).

Next, the reference plate is moved by Lx in the + X direction and by Ly in the -Y direction. Then, the position difference between the mark Ad and the mark Fd is measured by another wavelength TTR alignment detection system 12, and the phase difference is calculated. To detect. Here, the mark Fd
Is manufactured (drawn) according to the design value, this phase difference should be the same as the reference phase difference. However, if there is a manufacturing error, a change occurs with respect to the reference phase difference. The change is the deviation Eda1 from the design value of the measurement point Fda1.

Similarly, the other measurement points Fda2, Fda3, Fda
Deviations Eda2, Eda3, and Eda4 from the design value of da4 can also be obtained. Here, the design value of each point is centered on the center Fcc of the reference plate FP obtained using the measurement results of two points FMr and FMl, and passes through the X axis formed by connecting the two points FMr and FMl, and Fcc. What is necessary is just to make the coordinate system created by the X-axis and the Y-axis orthogonal to a reference a reference.

Next, Eda1, Eda2, Eda determined above
3. Using Eda4, the rotation error Rda of the section A and the magnification error Mda of the section A when Fda0 is used as a reference point based on the following equation:
An offset error Oda of the section A (the offset error Oda is an error in the X direction because the mark Fd is a mark for measuring in the X direction) is obtained.

Rda = {(Eda3-Eda1) + (Eda4-Eda2)} / (4 × Lx) Mda = {(Eda2-Eda1) + (Eda4-Eda3)} / (4 × Ly) Oda = (Eda1 + Eda2 + Eda3 + Eda4) / 4 Similarly, for the remaining sections B, C, and D, a rotation error, a magnification error, and an offset error with respect to a reference point in each section are obtained.

The remaining reference marks Fu, Fr, Fl
Similarly, the rotation error, magnification error, and offset error (the X direction for the mark Fd, and the Y direction for the marks Fr and Fl) with respect to the reference point for each section are obtained in advance.
The rotation error, magnification error, and offset error for each section of the marks Fd, Fu, Fr, and Fl thus obtained are stored in the internal memory of the main control device 16.

If the process of manufacturing the reference plate is known, the number and arrangement of the sections on the mark can be known without measuring. Also, if you do not know it, measure it once and check it,
It can be used for other reference plates made in the same process.

Next, an actual method of aligning the reticle R with the wafer W in this embodiment will be described.
Here, a case in which a mark on the reticle is replaced will be described.

First, the marks Fd, Fu, F
Baseline measurement is performed using a reference plate in which the rotation error, magnification error, and offset error of each section of r and Fl are measured. This baseline measurement is performed for each of the four different wavelength TTL alignment systems 12 in consideration of mark replacement.

First, reticle alignment is performed.
And two beams Lr of another wavelength TTR alignment system
1, the displacement of the irradiation position of Lr2 and the two beams Lm1 and Lm2
Check for telecentricity error. When the check is completed, the wafer stage ST is moved and the reference plate FP
Is positioned below the projection lens PL. The CCD elements of the two exposure wavelength TTR alignment systems 14 (7 in FIG. 1)
8A and 78B only), using the reticle mark RM
r, RMl and the reference marks FMr, FMl are simultaneously detected, and at the same time, marks Au, Ar, Ad, and Al are arranged in each mark area on the reticle R using four different wavelength TTR alignment systems 12. And a relative displacement between each of the reference marks Fu, Fr, Fd, and Fl on the reference plate FP is detected.

As a result, the amount of misalignment detected by the exposure wavelength TTR alignment system 14 and each wavelength TTR
The base line amount of each wavelength alignment detection system 12 is obtained from the difference from the displacement amount detected by each of the alignment systems 12.

As a result, for example, the reference marks Fu, Fr,
It is assumed that the baseline amounts of the alignment microscopes (object lenses OBJu, OBJr, OBJd, OBJl) corresponding to Fd and Fl are calculated as BLu, BLr, BLd, and BLl. These baseline amounts BLu, BLr, B
Ld and BLl include a manufacturing error of the reference plate FP. In the following, another wavelength TTR having a mark Au, a reference mark Fu, and an objective lens OBJd for detecting these marks will be described.
Alignment detection system (hereinafter, the microscope of this detection system is referred to as “A
Md ”) as an example.

Next, the baseline amount B measured above
The correction value of Ld (that is, the manufacturing error EFd of the reference plate) is calculated. Specifically, it is as follows.

Assume that the coordinates of the observation position of the microscope AMd are based on the shot center (AMdx, AMdy), and the size of the alignment beam is ABdx in the X direction and ABdy in the Y direction. Also, the coordinates of the reference point Fua0 in the coordinate system of the reference plate FP are (Fdax, Fday),
Similarly, the coordinates of the reference points Fdb0, Fdc0, and Fdd0 in the coordinate system of the reference plate FP are represented by (Fdbx, Fdby), (Fdcx, Fd
cy) and (Fddx, Fddy). Further, as shown in FIG. 9, here, the boundary between the section A and the section B and the boundary between the section C and the section D are both on X = α.
It is assumed that both the boundary line between section B and section C and the boundary line between section B and section D are on Y = β.

A. AMdx + ABdx / 2 ≦ α, AMdy + AB
When dy / 2 ≦ β, in other words, when the alignment beam is within the section A, the manufacturing error of the mark Fd may be considered only for the section A, and the error EFd in this case is EFd = Oda + (AMdx -Fdax) × Mda + (AMdy-Fday) ×
It is expressed as Rda.

B. α <AMdx + ABdx / 2 <α + ABdx /
2. When AMdy + ABdy / 2 ≦ β, in other words, when the alignment beam is on the boundary between the section A and the section B, the manufacturing error of the mark Fd is determined by aligning the respective errors of the section A and the section B.・ It is only necessary to sort according to the ratio of the area that the beam covers each section, for example,

EFd = {Oda + (AMdx−Fdax) × Mda + (AMdy−Fday) × Rda} × {α− (AMdx−ABdx / 2)} / ABdx + {Odb + (AMdx−Fdbx) × Mdb + (AMdy−Fdby ) × Rdb} × {(AMdx + ABdx / 2) -α} / ABdx.

C. AMdx + ABdx / 2 ≦ α, β <AMdy +
When ABdy / 2 <β + ABdy / 2, in other words, when the alignment beam is on the boundary between the section A and the section C, the manufacturing error of the mark Fd is b. As in the case of
For example,

EFd = {Oda + (AMdx-Fdax) × Mda + (AMdy-Fday) × Rda} × {β- (AMdy-ABdy / 2)} / ABdy + {Odc + (AMdx-Fdcx) × Mdc + (AMdy-Fdcy ) × Rdc} × {(AMdy + ABdy / 2) -β} / ABdy.

D. α <AMdx + ABdx / 2 <α + ABdx /
2. When β <AMdy + ABdy / 2 <β + ABdy / 2, in other words, the alignment beam is divided into sections A, B, C, and D.
, Depending on the ratio of the area of each section of the alignment beam, for example

EFd = {Oda + (AMdx−Fdax) × Mda + (AMdy−Fday) × Rda} × {α− (AMdx−ABdx / 2)} / ABdx × {β− (AMdy−ABdy / 2)} / ABdy + {Odb + (AMdx-Fdbx) × Mdb + (AMdy-Fdby) × Rdb} × {(AMdx + ABdx / 2) -α} / ABdx × {β- (AMdy-ABdy / 2)} / ABdy + {Odc + ( AMdx-Fdcx) × Mdc + (AMdy-Fdcy) × Rdc} × {α- (AMdx-ABdx / 2)} / ABdx × {(AMdy + ABdy / 2) -β} / ABdy + {Odd + (AMdx-Fddx) × Mdd + (AMdy−Fddy) × Rdd} × {(AMdx + ABdx / 2) -α} / ABdx × {(AMdy + ABdy / 2) -β} / ABdy.

The alignment beam is divided into sections C,
The error EFd can be calculated in exactly the same manner when it falls within any of D and E, or when it is on the boundary line between them.

Mark Au (reference mark Fu), mark A
r (reference mark Fr), mark Al (reference mark Fl)
Regarding another wavelength TTR alignment detection system having the objective lenses OBJu, OBJr, and OBJl to be detected (the microscope of this detection system is referred to as “AMu, AMr, AM1”), the production error of the reference mark can be calculated in the same manner. .

Actually another wavelength TTR alignment system 1
When the shots on the wafer W are aligned in step 2, the manufacturing error (correction value) EFd obtained in step 2 is used to correct the base line amount BLd of the alignment microscope AMd obtained in step 2 to obtain a true baseline. Quantity B
Ld ′ is calculated by the following equation: BLd ′ = BLD−EFd The position shifted by the base line amount BLd ′ is set as the true alignment achievement position (the phase difference from the measurement unit 56 is the base line amount BLd Main controller 16 controls reticle stage RS or wafer stage ST so that the phase difference corresponds to a position shifted by ').

Other alignment microscopes AMu, AMr,
Regarding AM1, main controller 16 controls reticle stage RS or wafer stage ST in consideration of the true baseline amount obtained by correcting the baseline amount based on the reference mark manufacturing error calculated above.

In the above-described alignment method according to the present embodiment, the reference marks corresponding to each of the four different wavelength TTR alignment systems 12 are measured, so that the reference marks Fu, Fr, Fd, and Fl are measured. The relative positional relationship between them is not exactly known. However, by correcting the alignment position using the above-described manufacturing error (correction value), the four reference marks Fu and Fr based on the positions of the reference marks FMr and FMl corresponding to the microscope position of the exposure wavelength TTR alignment system 14 are used. , Fd, and Fl are corrected for relative positional deviation.

As described above, according to the alignment method according to the present embodiment, even when the four different wavelength TTR alignment systems 12 are moved to change the alignment position, the base newly measured at the displaced alignment position is obtained. Based on the line amount BL and the respective manufacturing errors EF of the reference marks Fu, Fd, Fr, and Fl at the alignment position after the displacement, the actual alignment of the reticle R and the wafer W by the different wavelength TTR alignment system 12 is performed. By correcting each alignment position, an offset such as a mounting error of the reference plate and a drawing error of a pattern on the reference plate can be corrected, and alignment accuracy is dramatically improved. Also, in this case, highly accurate alignment can be performed even when the microscope is anywhere within the movable range.

As a method of correcting a reference mark manufacturing error, the correction value is calculated based on less data based on the characteristics of the error generated in the manufacturing stage. The measurement does not take a long time, it is hardly affected by the drift of the measurement sensor at the time of error measurement, and it is highly accurate even with a small amount of data because it is based on the characteristic of the manufacturing error of the reference mark. Correction is possible.

In the above embodiment, an exposure apparatus having a two-lens four-axis exposure light TTR alignment system and a four-lens four-axis different-wavelength TTR alignment system has been described. Such an alignment method can be applied. Further, in the above embodiment, one set of correction values (offset, rotation, magnification) is obtained for one section, but one section may be further divided finely and similar correction may be performed.

[0086]

As described above, according to the present invention,
The mask and the photosensitive substrate can be precisely aligned by taking into account not only positional deviation due to chromatic aberration of the projection optical system with respect to light (alignment light) having a wavelength different from the exposure light, but also drawing errors of marks on the reference plate. There is an unprecedented excellent effect that alignment can be performed.

In particular, according to the third aspect of the present invention, it does not take a long time to measure the manufacturing error of the reference mark,
There is an effect that it is hardly affected by drift of a sensor for measurement at the time of error measurement, and that highly accurate correction can be performed even with a small amount of data.

[Brief description of the drawings]

FIG. 1 is a view schematically showing a configuration of a stepper according to one embodiment of the present invention.

FIG. 2 is a diagram showing an arrangement of reticle grating marks.

FIG. 3 is a diagram showing an arrangement of wafer grid marks.

FIG. 4 is a diagram illustrating an optical system of another wavelength TTR alignment system.

FIG. 5: Different wavelength TTR alignment system and exposure wavelength TTR
FIG. 3 is a diagram illustrating an arrangement with an alignment system.

FIG. 6 is a diagram showing a reticle pattern arrangement.

FIG. 7 is a diagram showing a pattern arrangement on a reference mark plate.

FIG. 8 is an enlarged view showing a reference mark Fd of FIG. 7;

FIG. 9 is a diagram showing a positional relationship between each section of the reference mark Fd and another wavelength alignment beam.

[Explanation of symbols]

Reference Signs List 10 reduction projection type exposure apparatus (exposure apparatus) 12 different wavelength TTR alignment system (second detection system) 14 exposure wavelength TTR alignment system (first detection system) R reticle (mask) W wafer (photosensitive substrate) PL projection lens (Projection optical system) FP reference plate RMr, RMl Reticle mark (first mark) FMr, FMl Reference mark (second mark) Au, Ad, Ar, Al Reticle alignment mark (third mark) Fu, Fd, Fr, Fl Reference mark (fourth mark)

Claims (3)

[Claims] A first mark formed on a mask and having a first mark formed thereon;
An exposure wavelength alignment beam having the same wavelength as the exposure light in the mark area and the second mark area where the second mark is formed on the photosensitive substrate or the reference plate corresponding to the first mark via the first mark area and the projection optical system. And a first detection system for detecting a displacement between the first mark and the second mark by photoelectrically detecting reflected light generated from the first mark area and the second mark area; The third mark area where the third mark is formed and the fourth mark area where the fourth mark is formed on the photosensitive substrate or the reference plate corresponding to the third mark via the third mark area and the projection optical system. The third mark and the fourth mark are irradiated by irradiating another wavelength alignment beam having a wavelength different from that of the light, and photoelectrically detecting reflected light generated from the third mark area and the fourth mark area.
A method for aligning the mask and the photosensitive substrate, which is used in an exposure apparatus having a second detection system for detecting a positional deviation from a mark; after positioning the reference plate below the projection optical system A first mark on a mask and a second mark on the reference plate corresponding to the first mark on the mask by the first detection system.
A first step of detecting an amount of displacement from a mark;
A second step of detecting the amount of misalignment between the third mark on the mask and the corresponding fourth mark on the reference plate by the detection system of the above (1) and (2); A third step of calculating a difference in the amount of positional shift as a positional shift amount due to chromatic aberration; dividing the fourth mark on the reference plate into a plurality of sections, and the fourth mark on the reference plate in the second step A fourth step of calculating a correction value of a positional shift amount due to chromatic aberration calculated in the third step using a manufacturing error of the fourth mark for each section according to the irradiation position of the different wavelength alignment beam above; Then, using the second detection system, detect a position shift between the third mark on the mask and the corresponding fourth mark on the photosensitive substrate, and detect the detection result with the third and third marks. Use the values obtained in each of the four steps Alignment method comprising a fifth step of performing positioning between the photosensitive substrate and the mask Te. 2. The alignment method according to claim 1, wherein the plurality of sections are determined according to a step of manufacturing a fourth mark on the reference plate. 3. The manufacturing error of each section of the fourth mark on the reference plate is obtained by measuring a plurality of measurement errors having a predetermined positional relationship with the reference point selected based on a predetermined reference point of each section. 3. The alignment method according to claim 1, wherein the error is an error calculated by a calculation based on a deviation of a point from a design value.