JPH03166712A - Mark detector and evaluating method of superposition and selecting method of wavelength - Google Patents

Abstract

PURPOSE:To detect the image of a mark optically transferred onto a substrate to be projected and exposed by a single device even at the time of numbers of kinds of selected wavelengths by providing a light source generating detecting illuminating light having a plurality of wavelengths, the wavelength selective means of the detecting illuminating light and the focusing means of the image of the mark. CONSTITUTION:A light source 11 generating detecting illuminating light having a plurality of wavelengths, the wavelength selective means 21, 22 of the detecting illuminating light and the focusing means 3, 9, 71-73 of the image of a mark are provided. A wavelength proper to process conditions from a plurality of wavelengths of the detecting illuminating light is selected by the wavelength selective means 21, 22. The focus is adjusted to the image of the mark in order to correct chromatic aberration in the direction of an optical axis generated by the wavelength selected of the detecting illuminating light by the focusing means 3, 9, 71-73, and the mark is detected. Accordingly, the image of the mark optically transferred onto a substrate to be projected and exposed can be detected by a single device even at the time of numbers of kinds of selected wavelengths.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides a mark detection device that optically detects a mark on a projection exposure substrate onto which a pattern formed on a mask is transferred by a projection optical system; The present invention relates to a transfer overlay evaluation method using this mark detection device and a wavelength selection method using the mark detection device and overlay evaluation method. [Prior Art] Conventionally, in the manufacture of semiconductor devices, a stepper is used that performs reduction projection exposure while sequentially moving a wafer in steps. Semiconductor devices are manufactured by sequentially overlapping the circuit patterns on the mask and exposing them to light. Therefore, the circuit patterns on the wafer and the circuit patterns on the mask are aligned with high precision (
(hereinafter referred to as alignment).
Usually, this alignment is performed by optically detecting the signal waveform of an alignment mark on the wafer or mask, and moving the wafer stage or mask stage based on the position information obtained from this. Therefore, the overlay accuracy depends first on the detection accuracy of the alignment marks. On the other hand, the detection signal waveform of the alignment mark on the wafer changes depending on the resist film thickness, coating unevenness, mark step shape and depth, etc. These conditions differ for each semiconductor manufacturing process, so depending on the process, the contrast and symmetry of the waveform may deteriorate, leading to a decrease in the accuracy of the detected position. However, even in such cases, changing the wavelength of the detection illumination light can improve the waveform shape and improve accuracy. For this reason, JP-A-63-7
As described in Japanese Patent No. 0521, a method has been proposed in which a plurality of light sources with different wavelengths are prepared and these are used depending on the process. By the way, overlay accuracy also deteriorates due to expansion and contraction of the wafer during the process, asymmetrical alignment marks, and changes in the magnification of the reduction lens. For this reason, before exposing the product, a trial exposure is usually performed first, and after development, the overlay accuracy is evaluated using a microscope, etc. Therefore, among the factors that deteriorate the overlay accuracy, a device for measuring the change in magnification of a reduction lens was proposed in Japanese Patent Laid-Open No. 60-23.
It is described in Publication No. 8836. This device detects the change in optical properties before and after exposure as a latent image, and measures the magnification change from the pitch of the two latent image patterns without development. [Problem to be solved by the invention] The technique described in Japanese Patent Application Laid-Open No. 1983-70521,
Since the detection optical system was fixed, a chromatic aberration correction optical system for the alignment illumination light of the reduction lens was required only for the type of wavelength to be selected. Furthermore, when applying the technique described in JP-A-60-238836 to evaluation of overlay accuracy, it is necessary to detect both the latent image and the alignment mark. At this time,
When the detection optical system for alignment was also used as the overlay accuracy evaluation optical system, in order to align the imaging positions of both the latent image and the 7 alignment mark, it was necessary to use the same illumination light for both. The first object of the present invention is to form a single image of a mark optically transferred onto a projection exposure substrate when using detection illumination light of a plurality of wavelengths, no matter how many different wavelengths are selected. The purpose of this invention is to provide a mark detection device that can be detected by the device. A second object of the present invention is to provide a mark detection device capable of detecting latent images and evaluating overlay accuracy. Furthermore, a third object of the present invention is to provide an overlay method that allows overlay errors to be evaluated accurately on-machine without development. A fourth object of the present invention is to automatically select an appropriate wavelength for each process condition from among various wavelengths of the detection illumination light, the overlay error of which is always smaller than the allowable value. An object of the present invention is to provide a wavelength selection method.

[Means for Solving the Problems] The object of the first step is to provide a light source that generates detection illumination light of a plurality of wavelengths, a wavelength selection means for the detection illumination light, and a focusing means for the image of the mark. By what you do, you will be achieved.

Further, the first object is to provide a stage for moving a part or the entire detection optical system of the focusing point means with respect to an image of the mark formed by the projection optical system, and a stage for driving the stage. By equipping himself with the means,
It is achieved. Furthermore, the second object is to make the mark a latent image formed by transferring the mark to a photosensitive layer coated on the projection exposure substrate by the projection optical system, and furthermore, to detect the mark. The illumination light has a wavelength at which the difference in transmittance of the photosensitive layer before and after exposure is maximum,
This can be achieved by using light close to the wavelength.

Furthermore, the third purpose is to form an overlay evaluation mark on the projection exposure substrate in advance, and after aligning the overlay evaluation mark formed on the mask, expose the photosensitive layer on the projection exposure substrate. This can be achieved by forming a latent image by doing this, detecting the overlay evaluation mark and the latent image on the projection exposure substrate by the mark detection device, and calculating the overlay error.

The fourth purpose is to use the wavelength selection means to
This is achieved by selecting the wavelength of the detection illumination light, calculating the overlay error using the above-described overlay evaluation method, and selecting a wavelength at which the overlay error is at least smaller than the allowable value. [Operation] In the mark detection device of the present invention, detection illumination light having a plurality of wavelengths is generated from a light source.

Then, the wavelength selection means selects a wavelength appropriate for the process conditions from among the plurality of wavelengths of the detection illumination light.

Then, in order to correct the chromatic aberration in the optical axis direction caused by the selected wavelength of the detection illumination light, the focusing means focuses on the image of the mark and detects the mark.

Further, in the mark detection device of the present invention, the focusing point means
Since the structure includes a stage for moving part or all of the detection optical system with respect to the mark and a drive means for the stage, no matter how many wavelengths of detection illumination light are selected, By moving the stage, the image of the mark can be detected more accurately with a single detection optical system.

Further, in the mark detection device of the present invention, the mark is transferred by the projection optical system to a photosensitive layer coated on the projection exposure substrate to form a latent image, and this latent image is detected. The overlay accuracy can be evaluated by

Furthermore, in the mark detection device of the present invention, the detection illumination light is set to either a wavelength at which the difference in transmittance of the photosensitive layer before and after exposure is maximum, or light close to the wavelength. Detect the latent image by improving the contrast of the latent image,
Overlay accuracy can be evaluated.

Furthermore, in the overlay evaluation method of the present invention, marks for overlay evaluation are first formed on the projection exposure substrate in advance. Furthermore, after aligning the overlay evaluation mark formed on the mask, a latent image is formed by exposing the photosensitive layer on the projection exposure substrate.

Next, the overlay evaluation mark and the latent image on the projection exposure substrate are detected by the mark detection device, and the overlay error is calculated. As a result, without development,
Overlay errors can be evaluated with high accuracy while still on-machine.

In the wavelength selection method of the present invention, the wavelength of the detection illumination light is selected by the wavelength selection means of the mark detection device.

Next, an overlay error is calculated using the overlay evaluation method described above.

In addition, at least a wavelength for which the overlay error is smaller than an allowable value is selected.

This allows the overlay error evaluation function to automatically select an appropriate wavelength of the detection illumination light for each process condition.

[Examples] Examples of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing the light flux passing through the mark detection device of the present invention. The device of the embodiment shown in FIG. ) and a wafer stage 8l.
, a pulse motor 82 and a laser length measuring device 83 for detecting the position of the wafer 6; a wavelength selection means for selecting an arbitrary wavelength from the alignment illumination light 14 which is detection illumination light of a plurality of wavelengths; It is equipped with a focusing point means. The mercury lamp 11 generates and irradiates light used as exposure light 13 and alignment illumination light 14. The circuit pattern projection optical system includes a reticle 4 and a reduction lens 5. The wavelength selection means for the alignment illumination light 14 includes an achromatic lens 21, an interference filter 22, a rotating holder 231 for the interference filter 22, and a pulse motor 232 for the same. The focusing means for the image of the mark includes the detection optical system 3, a detection optical system stage 71 that moves the detection optical system 3, a motor 72 which is a driving means for this, and the detection optical system 3 via the detection optical system stage 71. It is equipped with an encoder 73 for detecting the position of , and a control circuit 9 . The detection optical system 3 includes:
an achromatic lens 31, a beam splitter 34,
An objective lens 32, another achromatic lens 35, a reflecting mirror 39, another achromatic lens 36, and so on. It includes a cylindrical lens 37 and a one-dimensional light receiving element 38. The control circuit 9 includes a detection optical system stage 71
motor 72 and rotation holder 2 of the interference filter 22.
31, the pulse motor 82 of the wafer stage 81, etc. A folding mirror 42 is arranged between the objective lens 32 and the reduction lens 5 of the detection optical system 3. Further, a reflecting mirror 25 is arranged between the interference filter 22 and the achromatic lens 31 of the detection optical system 3. The light emitted from the mercury lamp l1 is divided into exposure light 13 and seven-line illumination light 14 by a beam splitter I2.
It is divided into The exposure light l3 is applied to the reticle 4 by an optical system (not shown).
The circuit pattern on the reticle 4 is guided by the reduction lens 5.
is transferred to a resist (not shown), which is a photosensitive layer coated on the wafer 6. On the other hand, the alignment illumination light l4 is turned into a parallel light beam by the achromat lens 21, and enters the interference filter 22, where the wavelength of the transmitted light 24 is selected. The interference filter 22 is mounted on a rotating holder 231
By rotating the pulse motor 232, the interference filter can be replaced with another interference filter, and the wavelength of the transmitted light 24 can be changed. In addition, the pulse motor 232
is controlled by control circuit 9. The transmitted light 24 is passed through the reduction lens 5 by the bright lens 3l.
The light is focused on plane A, which is at a conjugate position with the entrance pupil 51 of . The objective lens 32 realizes Koehler illumination by forming an image of the A plane onto the entrance pupil 51, thereby making the illuminance distribution on the wafer 6 uniform. Note that the folding mirror 42 is the reticle 4.
It contains the information of the circuit pattern above and is placed in a position that does not block the exposure light 13. The return light 4l from the wafer 6 is reflected by the folding mirror 42 to the detection optical system 3, focused again on the A plane by the objective lens 32, reflected by the beam splitter 34, and transmitted to the achromatic lens 35. By 36, A
The light is focused on point B, which is the conjugate position with point B. The cylindrical lens 37 images the point B onto the one-dimensional light receiving element 38, and
The dimensionally compressed image signal of the wafer 6 is sent to the control circuit 9. Next, FIG. 2 shows the imaging relationship between the reduction lens 5 and the detection optical system 3. The upper surface of the wafer 6 is imaged at an imaging position C1 by the reduction lens 5. The image forming position C is imaged at point D by the objective lens 32 and the achromatic lens 35, and the image at point D is formed onto the one-dimensional light receiving element 38 by the achromatic lens 36. In addition, in FIG. 2, the cylindrical force lens 37
and one-dimensional light-receiving element 38 are side views of the cylindrical drill lens 37 and one-dimensional light-receiving element 38 in FIG.

By the way, since the reduction lens 5 is designed for the exposure light 13, it has longitudinal chromatic aberration for the alignment illumination light 14 having a different wavelength, and the imaging position C1 of the wafer 6 is designed for the alignment illumination light 14. Move by wavelength. Therefore, when the wavelength is changed by the interference filter 22, it is necessary to move the imaging position of the detection optical system 3 in order to image the imaging position C1 on the one-dimensional light receiving element 38.

For this purpose, a detection optical system stage 71 that supports the entire detection optical system 3 is driven by a motor 72. Further, it is necessary to position the detection optical system stage 71 so that the focal position of the objective lens 32 and the imaging position C coincide with each other. For this reason, an encoder 73 is attached to the motor 72 to detect the rotation angle. The signal from the encoder 73 is sent to the control circuit 9, and the motor 72 is controlled. Note that the signal for positioning may be obtained by separately attaching a length measuring device that detects the moving distance of the detection optical system stage 71. Alternatively, the detected waveform from the one-dimensional light receiving element 38 may be monitored and the detection optical system stage 71 may be positioned at a position where the contrast of the detected waveform is maximized.

Furthermore, FIG. 3 is a diagram showing another embodiment in which a portion of the detection optical system 3 is moved. The detection optical system 3
Instead of moving the entire achromat lens 31 and objective lens 32 on a small stage 74, as shown in FIG.
The six small stages 74 may be moved by a motor 75.
The moving distance is monitored by an encoder 76. The signal from the encoder 76 is sent to the control circuit 9, and the motor 75 is controlled. As described above, positioning may be performed using a length measuring device or the contrast of the detected waveform.

Next, FIG. 4 is a diagram showing the relationship between the center of the detected waveform and the reference position of the one-dimensional light receiving element. Next, the above-mentioned FIGS. 2 and 4
The alignment method will be explained with reference to figures. First, the wafer stage 81 is moved by the pulse motor 82 until the designed center position of the detected waveform matches the reference position X0 of the one-dimensional light receiving element 38. The coordinates are read into the control circuit 9 by the laser length measuring device 83. Actual detected waveform 381
The center of the reference position from ΔX
It shifts by just that. Alignment is performed by moving the wafer stage 81 by an amount corresponding to this ΔX. Next, a method of using the mark detection device as an overlay error evaluation device will be described.

Figure 5 is a diagram showing the principle of detection of latent images, Figure 6 (a) is a cross-sectional view showing the formation of overlay evaluation marks on the base, and Figure 6 (b) is a diagram showing the formation of overlay evaluation marks in the resist. FIG. 3 is a diagram showing a detected waveform of a latent image.

First, the principle of latent image detection will be explained using FIG.
．． ate Technology/Japanese version/S ep
tember 1988, Pp26-33 (Solid State Technology/Japanese version 79th issue 198g
, pP26-33), the transmittance characteristics with respect to wavelength change before and after exposure. That is, at 300 n ra to 450 nm, a portion that was opaque before exposure becomes transparent after exposure. This change is greatest around a certain wavelength, for example the h-line (408 nm). Therefore, if this phenomenon is utilized, it becomes possible to detect the pattern after exposure without development, and for example, if the pattern is illuminated with H-line, the contrast of the detected image will be maximized.

Next, the latent image, the overlay evaluation marks, and their detected waveforms will be explained with reference to FIGS. 6(a) and (b). The overlay evaluation mark 62 is formed by, for example, two recesses in a previous process. A resist 61 is applied thereon, and a pattern for forming a latent image on the reticle 4 shown in FIGS. 1 to 3 is exposed. As a result, as shown in FIG. 6(a), the resist 61 has an exposed area 611 and an unexposed area 611.
Divided into 2. When these are detected with the h line, the detection signal 6
3, a valley is formed at a location corresponding to the edge portion of the overlay evaluation mark 62 and the unexposed portion 612. From these, the position XA*Xi of the overlay evaluation mark 62 and the position IY^yYB of the unexposed area 612 are determined, for example, by the
It is determined by the method disclosed in Publication No. 2-2284. Using these values, the center position 1 of the overlay evaluation mark 62 is determined! Xc, center position 1 of unexposed area 612
! Yc is determined by the following formula. XC= (XA+XIl)/2 Yc= (YA+YB)/2 From these, the overlay error e is given by the following formula. a = YC - XC The overlay error e may be evaluated by the average value of the overlay errors e of several shots or 3σ (σ is the standard deviation). However, since the h-line is the sensitive wavelength of this resist,
The overlay evaluation mark 62 and the unexposed portion 612 should be formed in an area that does not affect the circuit pattern, such as a scribe area. Further, the resist 6l often contains a light absorbing agent that absorbs exposure light in order to prevent the cross section of the pattern from undulating due to the generation of standing waves. In such a case, as shown in FIG. 2, first select the h-line using the interference filter 22, and move the detection optical system stage 71 so that the focal position of the objective lens 32 and the h-line imaging position of the reduction lens 5 coincide. After that, the unexposed portion 612 is detected. After that, interference filter 2
2, select a wavelength that is not absorbed by the light absorbing agent, move the detection optical system stage 71 so that the image formation position of the reduction lens 5 and the focal position of the objective lens 32 at this wavelength match, and mark the overlay evaluation mark 62. Detect. However, if the detection optical system stage 7l has yawing or the like, an offset will occur in the detection position due to the movement, so it is necessary to calculate this in advance and correct the detection position. FIG. 7 is an explanatory diagram of correction of the detection position of the detection optical system shown in FIG. 2, FIG. ) is a diagram showing the offset of the detected waveform center value due to changes in the wavelength of illumination light. If an offset occurs in the detection position of the detection optical system 3, as shown in FIG. The fiducial mark 64 is first detected by the h-line, and the
The center position YCo shown in FIG.
Calculate the center position XCo shown in c). At this time, the offset Δ is Δ=Y Co Incidentally, the S/N of the detection waveform for alignment may deteriorate depending on process conditions such as the thickness of the resist, the depth of the target step, and the refractive index of the underlying film. However, even in such a case, the S/N can be improved from the detection waveform 382 to the detection waveform 383 by changing the wavelength for alignment illumination, as shown in FIG. Therefore, in order to improve alignment accuracy, it is necessary to optimize the wavelength of alignment illumination light for the process conditions. Hereinafter, a method will be described in which this detection optical system 3 is used as an overlay error evaluation optical system and a wavelength is automatically selected from the evaluation results. FIG. 10 is a diagram showing the steps of a wavelength selection method for selecting a desired wavelength from N types of wavelengths using the detection optical system shown in FIG. 2, and FIG. 11 (
a) to (C) are diagrams showing the exposure method in the steps shown in FIG. 10. These figures 2, 10 and 11 (a) to (c)
First, in step 901, one wavelength is selected by the interference filter 22, in step 902 the detection optical system stage 71 is moved, in step 903 alignment is performed, and in step 904, first, as shown in FIG. Perform exposure to light.

Next, the process returns to step 901, another wavelength is selected using the interference filter 22, and the following steps 902 to 904 are followed to perform exposure as shown in FIG. 11(b). In this way, steps 901 to 904 are repeated for N types of wavelengths, and the wafer exposure position is changed for each wavelength as shown in FIG. By performing exposure in this manner, the number of wafers required for trial exposure can be reduced to one.

Next, in step 905, the wavelength of the latent image detection light is selected, and in step 906, the detection optical system stage 7l is moved,
In step 907, the latent image and overlay evaluation mark of the shot aligned with the alignment illumination light of wavelength k are detected, and in step 908, the overlay error e is calculated.
In step 909, e is compared with a preset tolerance value,
Select the alignment illumination light when the tolerance is met. Instead of step 909, a method may be adopted in which the overlay errors e at each alignment wavelength are relatively compared and the smallest one is selected. [Effects of the Invention] According to the invention described in claim 1 of the present invention described above, a light source that generates detection illumination light of a plurality of wavelengths, a wavelength selection means for the detection illumination light, and a combination of images of the mark are provided. According to the invention as claimed in claim 2, the focusing means is provided as a part of the detection optical system or a part of the detection optical system with respect to the image of the mark formed by the projection optical system. Since the structure includes a stage that moves the entire structure and a driving means for the stage, when using detection illumination light of multiple wavelengths, no matter how many different wavelengths are selected, it is possible to This has the effect of being able to detect images of optically transferred marks with a single device. Furthermore, claim 3 of the present invention
According to the invention described, the mark is a latent image formed by being transferred by the projection optical system onto the photosensitive layer coated on the projection exposure substrate, so the latent image is detected and superimposed. There is an effect that alignment accuracy can be evaluated, and furthermore, according to the invention according to claim 4, the detection illumination light is set to a wavelength at which the difference in transmittance of the photosensitive layer before and after exposure is maximum. , or light close to the wavelength, it is effective to improve the contrast of the latent image, detect the latent image, and evaluate the overlay accuracy. Furthermore, according to the fifth aspect of the present invention, an overlay evaluation mark is formed on the projection exposure substrate in advance, and after alignment of the overlay evaluation mark formed on the mask, the projection exposure substrate is A latent image is formed by exposing the upper photosensitive layer, and the overlay evaluation mark and the latent image on the projection exposure substrate are detected by the mark detection device to calculate the overlay error. This has the effect of allowing overlay errors to be evaluated accurately on-machine without development.

According to the invention as set forth in claim 6 of the present invention, the wavelength selection means of the mark detection device selects the wavelength of the detection illumination light, and the overlay error is calculated by the overlay evaluation method, so that at least the overlay error is calculated. Since the wavelength is selected for which the error is smaller than the allowable value, for each process condition, the appropriate wavelength is automatically selected from among the various wavelengths of the detection illumination light, with the overlay error always being smaller than the allowable value. There are effects that can be selected.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing a light flux passing through a mark detection device which is an embodiment of the present invention, Fig. 2 is a diagram showing the imaging relationship of the detection optical system of Fig. 1, and Fig. 3 is a diagram showing the detection of Fig. 1. A diagram showing another example of the optical system, FIG. 4 is a diagram showing the relationship between the center of the detected waveform and the reference position of the one-dimensional light receiving element, and FIG. 5 is the relationship between the change in transmittance of the resist and the wavelength before and after exposure. Figure 6 (a)
), (b) is a diagram showing the overlay evaluation mark on the base, the latent image for overlay evaluation in the resist, and the detected waveform;
Fig. 7 is an explanatory diagram of correction of the detection position of the detection optical system shown in Fig. 2, Fig. 8 (a) is a diagram showing fiducial marks for detection position correction, Fig. 8 (b), (C) is a diagram showing the offset of the detected waveform center value due to a change in the wavelength of the illumination light, Figure 9 is a diagram showing the change in the detected waveform due to a change in the wavelength of the alignment illumination light, and Figure 10 is a diagram using the detection optical system shown in Figure 2. FIGS. 11(a) to 11(C) are diagrams showing the exposure method in the steps shown in FIG. 10. 1l...Mercury lamp as a light source, 12...Beam splitter, 13...Exposure light, 14...Detection illumination light,
4... A reticle of a pattern projection optical system, 5... Also a reduction lens, 6... A wafer as a projection exposure substrate, 8l... A wafer stage, 22... A wavelength selection means. interference filter, 231... rotary holder for interference filter, 232... pulse motor, 24
. . . transmitted light, 42 . . . folding mirror, 3 . . . detection optical system constituting the focusing means, 71 . A certain motor, 73... Encoder for position detection of the detection optical system, 74... Small stage of the detection optical system, 75
...Motor for driving the small stage, 76...Encoder for detecting the position of the small stage, 61...Resist which is a photosensitive layer on the wafer, 62...Mark for overlay evaluation, 611... Exposed part, 612... Unexposed part, e.
...Overlay error, 63...Detection signal, 64...
VIDUAL MARK, 901-909...Steps of a method for selecting a desired wavelength from N types of wavelengths. Figure 2 Figure 4 Figure xo 62 overlap evaluation 7-7 63 & 8ll. #+ 6] 1 Tribute part 612: Un☆ Maru decoration No. 7 Fig. 61' Entering Timer》 Rema 7 Fig. 8 Yco Yco: Middle IC position 9 Fig. 10 Fig. 11 To Figure 6-1 JP-A-3-166712 (H)

Claims (1)

[Scope of Claims] 1. In a mark detection device that detects marks on a projection exposure substrate onto which a pattern formed on a mask is transferred by a projection optical system, a light source that generates detection illumination light of a plurality of wavelengths; A mark detection device comprising: means for selecting a wavelength of the detection illumination light; and means for focusing an image of the mark. 2. The focusing means includes a stage for moving part or the entire detection optical system with respect to the image of the mark formed by the projection optical system, and a driving means for the stage. The mark detection device according to claim 1, characterized in that: 3. The mark detection according to claim 1 or 2, wherein the mark is a latent image formed by being transferred by the projection optical system onto a photosensitive layer coated on the projection exposure substrate. Device. 4. The mark according to claim 3, wherein the detection illumination light has a wavelength at which the difference in transmittance of the photosensitive layer before and after exposure is maximum, or light close to the wavelength. Detection device. 5. A method for evaluating the superposition of a pattern formed on a mask with the projection exposure substrate when forming an image of the pattern on the projection exposure substrate using exposure light via a projection optical system, the method comprising: An overlay evaluation mark is formed in advance on the projection exposure substrate, and after aligning the overlay evaluation mark formed on the mask, a latent image is formed by exposing the photosensitive layer on the projection exposure substrate, and 4. An overlay evaluation method, comprising: detecting overlay evaluation marks and latent images on a substrate to be projected and exposed using a mark detection device according to claim 3, and calculating an overlay error. 6. A method for selecting the wavelength of detection illumination light of a plurality of wavelengths generated from the light source according to claim 1, wherein the wavelength of the detection illumination light is selected by the wavelength selection means according to claim 1,
A wavelength selection method, comprising calculating an overlay error using the method described above, and selecting a wavelength for which at least an overlay error is smaller than a tolerance value.