JPH08274018A - Position detector and manufacturing method of semiconductor device using the same - Google Patents

Abstract

PURPOSE: To detect the positional data on alignment marks with high precision by detecting the positional data on the elements using the first measurement value given by the light flux through the intermediary of a mark A and the second measurement value given by the light flux through the intermediary of another mark B formed between a mark and adjacent marks. CONSTITUTION: Marks in width W and recessed sectional shape are arranged at a pitch of P. Out of these marks, six marks 01, I1, I2, I3, I4, 02 in receesed sectional shape are treated as the alignment marks A. Besides, five marks (Marks B) J1-J5 in width P-W and sectional shape in a convex type inverse to the marks A formed between adjacent marks A are treated as the alignment marks B. Accordingly, one alignment mark AA is treated as two alignment marks i.e., the alignment mark A in a recessed sectional shape and the alignment mark B in a convex sectional shape inverse to that of the alignment mark A.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a position detecting device and a method of manufacturing a semiconductor device using the position detecting device, and more particularly to a projection lens for forming a fine electronic circuit pattern such as IC or LSI formed on a reticle (mask) surface. When the system (projection optical system) projects and exposes on the wafer surface, the state (alignment mark) on the reticle surface or the wafer surface is observed, and the reticle and the wafer are aligned with each other to achieve high integration. It is suitable for manufacturing semiconductor devices.

[0002]

2. Description of the Related Art Conventionally, in a reduction projection type exposure apparatus for manufacturing semiconductor devices, a circuit pattern of a reticle as a first object is projected by a projection lens system onto a wafer as a second object to be exposed. At this time, the alignment mark on the wafer is detected by observing the wafer surface using an observation device (detection means) prior to the projection exposure, and the so-called alignment is performed on the reticle and the wafer based on the detection result. ing.

The alignment accuracy at this time largely depends on the optical performance of the observation apparatus. Therefore, the performance of the observation device is an important factor in the exposure device. In particular, recently, various kinds of exposure apparatuses for manufacturing highly integrated semiconductor devices by using a phase shift mask, modified illumination, etc. have been proposed, and higher alignment accuracy is required for such exposure apparatuses.

Conventionally, in an exposure apparatus, the following three methods are mainly used as a method of observing a wafer alignment mark (wafer mark) for obtaining position information on the wafer surface. (1-a) Method that uses non-exposure light and does not pass through the projection lens system (OFF-AXIS method) (1-b) Method that uses exposure light and passes through the projection lens system (exposure light TTL method) (1 -C) A method that uses non-exposure light and passes through the projection lens system (non-exposure light TTL method).

Among these methods, for example, the applicant of the present invention has proposed an alignment system using a non-exposure light TTL type observation apparatus in Japanese Patent Laid-Open No. 3-61802. In this publication, an optical image of an alignment mark (wafer mark) on the wafer surface is formed on an image pickup device such as a CCD camera,
The image information obtained from the image pickup device is processed to detect the position of the wafer mark.

Further, the present applicant has filed Japanese Patent Application Laid-Open No. 62-232504.
In the publication, an optical image of a wafer mark is formed by a CCD camera, image information obtained by the CCD camera is binarized,
A position detection device has been proposed which detects the position of a wafer mark by performing template matching processing using a template on the position coordinates of a specific image pattern in the binarized image.

[0007]

In recent years, the miniaturization of semiconductor devices has advanced, and in response to this, there is a demand for more accurate relative alignment between the reticle and the wafer. As the alignment method, the non-exposure light TT described above is used.
In the L method, an optical image of an alignment mark (hereinafter referred to as a wafer mark) formed on a wafer is imaged on an image pickup device such as a CCD camera through a projection optical system (TTL), and an electric signal of the image is processed to perform image processing of the wafer mark. The position is being detected. Generally, a major factor that deteriorates the alignment accuracy when aligning a reticle and a wafer is (a) non-uniformity of film thickness near the alignment mark of the resist (b) asymmetry of the shape of the alignment mark.

FIG. 6 is an explanatory view of an alignment mark in this alignment method. FIG. 6A shows the wafer W.
6B shows the X-direction detection mark F1 and FIG. 6C shows the Y-direction detection mark F2. In many cases, these alignment marks F1 and F2 are arranged in the scribe line in each direction. Like this alignment mark (in this case, the number of marks forming the alignment mark is 6)
By increasing the number of lines, it is possible to perform highly accurate alignment by the statistical effect.

Next, the deterioration of the alignment accuracy due to the non-uniformity of the film thickness in the vicinity of the formation of the alignment mark on the resist will be described. Out of a plurality of marks, the outermost mark 0 is used to obtain the alignment signal for image processing.
Inner marks I1 to I excluding 1,02,03,04
4, I5 to I8. The vicinity of the outer mark is affected by uneven coating of the resist, and if an alignment signal is obtained from the outer mark, the alignment accuracy will deteriorate. The inner mark is less affected by the nonuniformity of the resist due to the so-called "breakwater effect" (described below).

The effect of this breakwater is that when a large wave crosses the breakwater in the sea, the height of the wave becomes various levels near the breakwater like turbulence, but after passing the breakwater. The height of the wave is constant. It is said that this phenomenon similarly occurs when the resist is applied.

In general, when a resist is applied to the wafer surface, the resist flows from a certain direction to the outer mark 0.
It hits one of 1,02,03,04 first. At that time, the resist film thickness changes in the vicinity of the outer mark, but the film thickness of the inner mark flowing thereafter becomes uniform.

When the film thickness changes, the alignment signal obtained from the mark in the vicinity of the film becomes as shown in FIG.
As shown in the figure, this is the resist surface S1 and the substrate surface S2.
When the phase difference of the reflected light is λ + (λ / 4) (where n is an integer), the blackened part (condition to become dark) due to the interference of the reflected light from is there.

The whited portion (brightening condition) is when the phase of the reflected light becomes n'λ + (λ / 2) (where n'is an integer). As described above, conventionally, the inner mark of the plurality of marks forming the alignment mark is used to prevent the deterioration of the alignment accuracy.

Next, the deterioration of the alignment accuracy due to the asymmetry of the mark shape of the alignment mark will be described.

FIG. 8A shows the case where the cross-sectional shape of one mark M1 is asymmetric, and FIG. 8B shows the alignment signals se1 and se2 obtained at that time. FIG.
As shown in (A), no distortion is generated at the dented edge e1 of the mark M1, but the protruding edge e2 is tapered as shown in the drawing, and the left and right cross-sectional shapes are asymmetric. As shown in FIG. 8B, the signal se1 from the edge e1 is weaker than the signal se2 from the edge e2.

The alignment accuracy is deteriorated due to the asymmetry of the signal. For example, when such asymmetry is similarly generated on the wafer surface, the alignment shift occurs. Further, as shown in FIGS. 9A and 9B, when the signal asymmetry direction is reversed on the left and right of the wafer, a wafer magnification measurement error occurs in the global alignment, which causes deterioration of 3σ.

The applicant of the present invention discloses in Japanese Patent Application Laid-Open No. 62-58625 that the alignment signals are opposite when the mark has a concave cross section and when the cross section has a convex cross section, and the alignment accuracy is prevented from deteriorating based on the signal error at this time. We are proposing a positioning device. In this publication, marks having concave and convex cross sections are arranged close to each other to obtain a highly accurate alignment signal.

According to the present invention, the positional information of the alignment mark composed of a plurality of marks provided on the surface of the reticle and the wafer is detected with high accuracy, and the relative alignment between the reticle and the wafer is performed with high accuracy, so that a highly integrated semiconductor can be obtained. An object of the present invention is to provide a position detection device that facilitates device manufacturing and a semiconductor device manufacturing method using the position detection device.

[0019]

The position detecting device of the present invention is a positioning mark in which N (A) marks A having a width (1-1) are arranged in a one-dimensional direction at a pitch P (N is an integer of 2 or more). Is formed on the object surface to be aligned, and the first measurement value obtained from the light flux passing through M (M ≦ N) marks A among the N marks A, the mark and the mark adjacent thereto. The position information of the object is detected using the second measurement value obtained from the light flux passing through the M-1 marks B among the marks B having the width P-W formed between I am trying.

In particular, the marks A and B are characterized in that their cross-sectional shapes are opposite to each other.

In the projection exposure apparatus of the present invention, (2-1) the relative position between the first object and the second object is adjusted by the position detection apparatus, and then the pattern on the first object is adjusted by the projection optical system. When projecting and exposing on the second object, the position detecting device should align the alignment marks in which the marks A having the width W are arranged in the one-dimensional direction at N (N is an integer of 2 or more) pitch P. The marks A are formed on the object surface, and M (M ≦ M) of the N marks A are formed.
N) The first measurement value obtained from the light flux passing through the mark A and the width P formed between the mark and the adjacent mark.
The position information of the object is detected using the second measurement value obtained from the light flux passing through the M−1 marks B among the −B marks B.

The semiconductor device manufacturing method of the present invention is
(3-1) After the relative position between the reticle and the wafer is adjusted by a position detecting device, the pattern on the reticle surface is projected and exposed on the wafer surface through a projection lens, and the wafer is subjected to a developing step. When manufacturing a semiconductor device via the, the position detecting apparatus uses N marks A having a width W (N is 2
Alignment marks arranged in a one-dimensional direction at the above pitch P) are formed on the object surface to be aligned, and N
Among the marks A, the first measurement value obtained from the light flux passing through M (M ≦ N) marks A and the mark B having the width P−W formed between the marks and the adjacent marks M
It is characterized in that the positional information of the object is detected by using the second measurement value obtained from the light flux passing through one mark B.

[0023]

FIG. 1 is a schematic view of an essential part of an optical system of an embodiment when the position detecting apparatus of the present invention is applied to a projection exposure apparatus for manufacturing a semiconductor device. In the figure, numeral 8 is a reticle as a first object, which is placed on the reticle stage 10. A wafer 3 as a second object has alignment marks (AA marks) 4 as shown in FIG. 2 on its surface. A projection optical system 5 is composed of a projection lens system and projects a circuit pattern or the like on the reticle (mask) 8 surface onto the wafer 3 surface. The projection lens system 5 is a telecentric system on both the reticle 8 side and the wafer 3 side.

An illuminating system 9 illuminates the reticle 8 with exposure light. Reference numeral 2 denotes a θ, Z stage on which the wafer 3 is placed, and θ rotation and focus adjustment of the wafer 3, that is, Z
We are adjusting the direction. The θ, Z stage 2 is mounted on the XY stage 1 for performing the step operation with high accuracy. An optical square (bar mirror) 6 that serves as a reference for measuring the stage position is placed on the XY stage 1.
This optical square 6 is monitored by a laser interferometer.

The alignment between the reticle 8 and the wafer 3 in this embodiment is performed indirectly by aligning the reference marks 17, which will be described later, with the positional relationship. Or, actually align the resist image pattern etc. and perform exposure,
The error (offset) is measured, and offset processing is performed in consideration of the value thereafter.

Next, each element of the detecting means 101 for detecting the position of the mark 4 on the surface of the wafer 3 will be described.

Reference numeral 14 denotes a light source for detecting the mark 4, which emits a light flux (non-exposure light) having a wavelength different from that of the exposure light.
-It consists of a Ne laser. Reference numeral 15 denotes a lens, which collects the light flux (detection light) from the light source 14 and makes it enter the beam splitter 16 for illumination. Reference numeral 12 is an objective lens. Lens 15 reflected by beam splitter 16
The detection light from is collected by the objective lens 12, reflected by the mirror 11, and guided to the mark 4 on the surface of the wafer 3 via the projection lens system 5 to illuminate it.

Reference numeral 19 is a light source for illuminating the reference mark 17
It consists of ED etc. Reference numeral 18 is a lens. Light source 19
The light flux from is condensed by the lens 18 and guided to the reference mark 17 for illumination. Reference numeral 20 is a beam splitter for the reference mark, which reflects the light beam from the reference mark 17 and makes it enter the erector lens 21. The erector lens 21 images the reference mark 17 and the mark 4 on the surface of the wafer 3 on the image pickup surface of the CCD camera 22, respectively. The detecting means 101 in this embodiment has each of the above elements.

In this embodiment, the non-exposure light (detection light) emitted from the light source 14 is the lens 15, the beam splitter 16, the objective lens 12, the mirror 11, and the projection lens system 5 in order.
After that, the AA mark 4 on the wafer 3 is illuminated. The observation image information of the AA mark 4 on the wafer 3 is the projection lens 5,
Mirror 11, objective lens 12, beam splitter 1
6, an image is formed on the CCD camera 22 through the beam splitter 20 and the erector lens 21.

The projection lens system 5 is satisfactorily corrected for the exposure light in order to project the pattern of the electronic circuit drawn on the reticle 8 onto the wafer 3. Therefore, when the non-exposure light passes through the projection lens system 5, various aberrations occur.

In this embodiment, the objective lens 12 corrects various aberrations generated in the projection lens system 5. As a result, a good observation image of the AA mark 4 is formed on the CCD camera 22.

On the other hand, the reference mark 17 has a lens 18 that emits light emitted from a light source 19 formed of an LED for illuminating the reference mark.
The light is condensed and illuminated by, and forms an image on the CCD camera 22 through the beam splitter 20 and the erector lens 21. A on the wafer 3 imaged on the CCD camera 22
Position information of the observed image of the A mark 4 on the CCD camera 22 is measured by a signal processing device (not shown) by a method described later, and is imaged on the CCD camera 22 also measured by the signal processing device,
Accurate position information of the XY stage 1 is obtained by comparing the position information of the projection image of the reference mark 17 which is always fixed, and thereby highly accurate alignment is performed.

Next, a method of detecting the position information of the AA mark 4 on the wafer 3 in this embodiment will be described.

FIG. 2 shows the alignment mark (A
It is explanatory drawing of (A mark) 4. FIG. 2A is a plan view of a main part, and FIGS. 2B and 2C are cross-sectional views of the main part. In the figure, a plurality of marks having a width W (4 μm) and a concave sectional shape are arranged on the substrate at a pitch P (20 μm). Of these, FIG. 2B shows six marks (marks A) 01, I1, I having a concave cross section with a width W as alignment marks A.
2, I3, I4, 02 are shown.
In FIG. 2C, five marks (marks B) J1 to J5 each having a width P-W formed between two adjacent marks A as an alignment mark B and having a convex cross section opposite to the mark A (marks B) are shown.
Shows the case of handling as. Of the alignment marks in FIG. 2B, the outer marks 01 and 05 are not used as marks that serve as a breakwater. FIG. 2
Of the alignment marks of (C), the outer mark J1,
Similarly, J5 is not used.

When using the alignment mark A of FIG. 2B, the edges a1 and a2 of the mark I1, the edges b1 and b2 of the mark I2, and the edge c of the mark I3 are used.
1 and the edge c2, and the scattered light obtained from the edge of the mark I4 having a width of 4 μm and the edge d1 and the edge d2 of the mark I4 are detected by the CCD camera 22 as the detection means, and the detected values (statistics such as abnormal value splashes) are detected. The position information (first measurement value) of the alignment mark A is obtained from the average value of (after processing). As signal processing at this time, position detection of a concave mark having a width of 4 μm such as edge a1 and edge a2 is performed.

On the other hand, when the alignment mark B shown in FIG. 6C is used, the edges a2 and b1 of the mark J2, the edges b2 and c1 of the mark J3, and the mark J4 are used.
The scattered light obtained from the edges c2 and d1 having a width of 16 μm (P−W) is detected by the detecting means, and the position information (second measurement value) of the alignment mark B is obtained from the average value of the detected values. ing.

As described above, in the present embodiment, one alignment mark AA is replaced by an alignment mark A having a concave sectional shape.
And a convex alignment mark B whose cross-sectional shape is opposite to that of
The alignment marks A and B are processed by one image processing, and the position information as the alignment mark AA is obtained by using the respective position information obtained from the alignment marks A and B. ing.

As a result, the area for arranging the marks (the same area as the case where the occupied area is not doubled and is the same as one area) is prevented from being expanded, and the throughput is not lowered (image processing of one alignment mark and (Approximately the same time), the offset amount of the concave mark and the offset amount of the convex mark cancel each other as described later to prevent the occurrence of offset and perform highly accurate alignment.

Although the non-exposure light TTL method is shown in the embodiment of FIG. 1, the present invention is not limited to the exposure light TTL method or OFF.
The same applies to the AXIS system.

Next, the features of the alignment mark detecting method according to the present invention will be described. First, a description will be given of a case where the shapes of the alignment mark and the photoresist are symmetrical and there is no error in drawing the reticle.

FIG. 3 shows two illumination lights IL and IR which illuminate one mark M1 which constitutes the alignment mark. When the intensity of two illumination lights IL and IR that illuminate one mark M1 with the illumination lights IL and IR from the left and right with respect to the normal line of the surface affects the detection accuracy of the alignment detection system. Come on.

Now, suppose that the illumination lights IL and IR are not equal in intensity, and the illumination light IL is stronger than the illumination light IR, for example. An alignment signal when detecting the alignment position with the wafer when the illumination light is scattered by the step structure of the cross section of the mark M1 and when the scattering angle has directivity will be described.

(A) First, the case where the illumination light is scattered by the step structure of the mark M1 and the scattering angle has no directivity will be described. FIG. 4 shows a case where the illumination light is scattered by the step structure (edge) of the mark M1 and the scattering angle has no directivity. 4A shows the case where the mark M1 is illuminated from the left side only with the illumination light IL, and FIG. 4B shows the case where the mark M1 is illuminated from the right side only with the illumination light IR. Generally, when the illumination light is scattered by the step structure of the photoresist pattern, the scattering angle has no directivity. Therefore, in both FIGS. 4A and 4B, the scattered light from the edge a1 and the edge a2 are the same. Occurs.

Therefore, the intensity of the alignment signal is shown in FIG.
Compared to the cases of (A) and (B), signals SL and SR as shown in the lower part of the figure are obtained, which are proportional to the intensity ratio of the illumination lights IL and IR. The alignment signals SL and SR are larger when illuminated with the illumination light IL having a higher intensity than when illuminated with the illumination light IR.

However, the intensity difference between the scattered light from the edge a1 and the edge a2 is substantially the same in the case of the illumination light IL and the illumination light IR. Therefore, as shown in FIG.
When the mark M1 is illuminated with both L and the illumination light IR, the intensity ratios of the scattered light from the edge a1 and the edge a2 are substantially the same. When the illumination light is scattered by the step structure of the mark M1, when the scattered light has no directivity, the intensities of the illumination lights IL and IR are not equal, and the alignment detection system detects the alignment signal even when the detection accuracy is low. No asymmetry in the waveform. Therefore, the alignment detection accuracy does not deteriorate.

(B) Next, the case where the illumination light is scattered by the step structure of the mark M1 and the scattering angle has directivity will be described. FIG. 5A shows the illumination light I from the left side of the mark M1.
FIG. 5B shows a case of illuminating only with L, and mark M1
Is illuminated from the right with only the illumination light IR. When the illumination light is scattered by the step structure of the mark M1, if the scattering angle has directivity, the intensities of the scattered light from the edge a1 and the edge a2 differ.

The intensity ratio of the scattered light from the edge a1 and the edge a2 in FIG. 5A is the edge a2 and the edge a in FIG. 5B.
It becomes the same as the intensity ratio of 1. The edge a1 of FIG. 5A and the edge a2 of FIG. 5B (the edge a2 of FIG. 5A and the edge a2 of FIG.
The intensity ratio of the edge a1) in (A) is proportional to the intensity ratio of the illumination lights IL and IR. Therefore, when the mark M1 is illuminated by both the illumination light IL and the illumination light IR as shown in FIG. 5C, a difference in intensity of scattered light from the edge a1 and the edge a2 of the mark M1 occurs.

The sum of the alignment signal a1L by the illumination light IL based on the edge a1 and the alignment signal a1R by the illumination light IR and the illumination light IL based on the edge a2.
The intensity ratio does not occur only when the sum of the alignment signal a2L by the light source and the alignment signal a2R by the illumination light IR is equal, but this is stochastically very low.

When scattered by the stepped structure of the mark M1, the scattering angle has directivity, and if the intensities of the illumination lights IL and IR are not equal, the detection waveform SLR of the alignment signal becomes asymmetric and the alignment accuracy deteriorates. On the other hand, according to the present invention, even in such a case, the alignment can be performed with high accuracy without being affected by the intensity difference between the illumination lights IL and IR. Next, the reason will be described.

As the alignment mark AA, when the alignment mark A is targeted first, as shown in FIG. 5C, if the alignment signals sa1 and sa2 obtained from the edge a1 and the edge a2 are not equal to each other, the alignment accuracy is deteriorated. come. At this time, the misalignment direction of the alignment is the direction with the stronger alignment signal (alignment signal sa2, rightward direction). On the other hand, as shown in FIG. 5D, when the alignment mark B is used as the alignment mark AA and the edges a2 and b1 are used, the alignment signal sa2 is stronger than the alignment signal sb1. The direction is opposite to that when A is used (to the left).

In this embodiment, by using the average value of the position measurement values of the edge a1 and the edge a2 of the mark I1 and the position measurement values of the edge a2 and the edge b1 of the mark J2, the deviation directions are opposite and the absolute values are equal. Therefore, on average, they cancel each other out, whereby the intensity of the illumination light IL, IR is different, and even if the scattered light from the edge has a directivity in the scattering angle, good alignment is performed without degrading the alignment accuracy. ing.

Next, the case where there is an error on the alignment mark side will be described. In this case, the wafer is rotated by 180 degrees and measured, and the wafer-induced deviation (Wafer Induce
d Shift: WIS).

For example, “Δ” shown below is defined as the evaluation amount. As a name of a mark used for alignment, first, its pitch is called a line width. For example, the mark “20P4” is a mark having a pitch of 20 μm and a line width of 4 μm. One mark is 4 μm
Each mark element with a line width is composed of a plurality of marks. For image processing, the position of one mark element with a line width of 4 μm is detected and the statistical value is used as a measurement value. ing.

[0054] delta = [measured value of the 20P4] - [measured value of the 20P16] When the "delta" and that does not rotate the wafer 180 (delta _o)
And when it is rotated 180 degrees (Δ ₁₈₀ ). Half of the difference between “Δ ₀ ” and “Δ ₁₈₀ ” is the wafer factor, and half of the sum is the device factor (Tool Induced Shift: TIS) error. The present invention is to provide a method for eliminating the influence of TIS even if it is present.

In this embodiment, when the WIS amount is a problem as the required alignment accuracy,
It is better to change the alignment detection system at that time. Specifically, for example, the mark pitch, duty, illumination light width, center wavelength, detection method (bright or dark field detection, image detection or phase detection), processing method, etc. are changed. In this embodiment, the alignment detection system is optimized using the WIS as the evaluation amount in this way.

Although TIS changes as WIS changes, the present invention is to provide a detection method that is not affected by TIS, and the result of alignment accuracy is WI.
It is only the change of S. In addition, the definition of TIS in the automatic overlay measuring device of KLA, etc.
Is not defined (denoted as Δ ').

The definition of Δ'in the automatic overlay measuring device is as follows: Δ '= [measurement value of two inside 20P4]-[measurement value of two outside 20P4], and "Δ' ₀ " and " TIS is half of the sum of Δ _'180 "
Is the same. The difference in this definition arises from the nature of the device. Δ is intended to quantify the amount of deviation in the detection range, and Δ ′ is intended to quantify the non-linearity of the measurement within the detection range. The automatic overlay measuring apparatus obtains a relative deviation between one layer and another layer, and even if a uniform deviation occurs in both layers, it does not cause a problem.

[0058]

According to the present invention, by properly setting each element as described above, the positional information of the alignment mark composed of a plurality of marks provided on the reticle or / and the wafer surface can be detected with high accuracy. It is possible to achieve a position detection apparatus and a semiconductor device manufacturing method using the position detection apparatus, in which relative positioning between a reticle and a wafer is performed with high accuracy to facilitate manufacturing of a highly integrated semiconductor device.

[Brief description of drawings]

FIG. 1 is a schematic view of a main part of a first embodiment of the present invention.

2 is an explanatory view of the alignment mark of FIG.

FIG. 3 is an explanatory diagram showing a relationship between a signal obtained from an alignment mark according to the present invention and the intensity of illumination light.

FIG. 4 is an explanatory diagram showing a relationship between a signal obtained from an alignment mark according to the present invention and intensity of illumination light.

FIG. 5 is an explanatory diagram showing a relationship between a signal obtained from an alignment mark according to the present invention and intensity of illumination light.

FIG. 6 is an explanatory view of a conventional alignment mark.

FIG. 7 is an explanatory diagram of alignment marks when the resist film is uneven.

FIG. 8 is an explanatory diagram of a cross-sectional shape of an alignment mark.

FIG. 9 is an explanatory diagram of a signal obtained from an alignment mark in a horizontal direction of a wafer.

[Explanation of symbols]

 1 XY stage 2 θ stage 3 wafer 4, M1, AA, A, B alignment mark 5 projection lens system 6 bar mirror 7 laser interferometer 8 reticle 9 illumination system 10 reticle stage 11 mirror 12 objective lens 14 light source 15 lens 16, 20 half Mirror 17 Reference mark 18 Lens 19 Light source 21 Elector lens 22 CCD camera

Claims (4)

[Claims] 1. Alignment marks in which N marks A having a width W are arranged in a one-dimensional direction at a pitch P (N is an integer of 2 or more) are formed on an object surface to be aligned, and the N alignment marks are formed. Of the marks A, the first measurement value obtained from M (M ≦ N) light fluxes passing through the marks A and M− of the marks B having a width P−W formed between the marks and the adjacent marks. A position detecting device characterized in that position information of the object is detected using a second measurement value obtained from a light beam passing through one mark B. 2. The position detecting device according to claim 1, wherein the mark A and the mark B have mutually opposite cross-sectional shapes. 3. When the pattern on the first object is projected and exposed on the second object by the projection optical system after the relative alignment between the first object and the second object is performed by the position detecting device, The position detection device forms N marks A having a width W at a pitch P (N is an integer of 2 or more) in a one-dimensional direction on a surface of an object to be aligned, and forms the alignment marks. Of the marks A, the first measurement value obtained from M (M ≦ N) light fluxes passing through the marks A and M− of the marks B having a width P−W formed between the marks and the adjacent marks. A projection exposure apparatus, wherein position information of the object is detected by using a second measurement value obtained from a light beam passing through one mark B. 4. A relative position between a reticle and a wafer is adjusted by a position detecting device, and then a pattern on the reticle surface is projected and exposed on the wafer surface through a projection lens to develop the wafer. When manufacturing a semiconductor device via the, the position detecting device should align the alignment marks in which the marks A having the width W are arranged in the one-dimensional direction at N (N is an integer of 2 or more) pitch P. It is formed on the object plane and is formed between a first measurement value obtained from a light beam passing through M (M ≦ N) marks A among the N marks A, and the mark and a mark adjacent thereto. Manufacture of a semiconductor device, characterized in that position information of the object is detected using a second measurement value obtained from light fluxes passing through M-1 marks B among the marks B having a width P-W. Method.