JPH055604A - Position detection method and pattern forming method and apparatus using the same - Google Patents

Abstract

PURPOSE:To enable accurate alignment of a pattern formed on the surface of a sample by incorporating a deviation portion due to an angle of inclination of a sample substrate into a mark position detection value on the rear of a sample substrate to cancel an error portion of a position measured value by the deviation portion. CONSTITUTION:When an angle theta of inclination is given between a sample 1 and a set reference surface, a deviation between an actual position (distance delta from a reference value) of a mark 2 on the rear of the sample and a true position (distance W from reference position) on the surface of the sample corresponding thereto is dsintheta when the thickness (d) of the sample is given and hence, a detection value error epsilonof the actual position of the mark 2=delta-W)=dsintheta is obtained. An optical system 19 for detecting position is provided as containing a deviation portion E canceling the error in the position detection value and a parameter of the optical system 19 is selected beforehand so that the detection results (delta+ E) becomes equal to the true position. Thus, the error epsilon by the angle theta of inclination is always canceled whatever the angle theta of inclination may be thereby enabling the determining of the true position on the surface of the sample accurately.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of manufacturing a semiconductor device and an improvement of an apparatus for carrying out the method, and more particularly to alignment in an X-ray exposure apparatus, a reduction projection exposure apparatus, or an electron beam drawing apparatus. The present invention relates to a mark position detecting method suitable for improving accuracy, a fine pattern forming method using the method, and an apparatus thereof.

[0002]

2. Description of the Related Art When a fine pattern such as a semiconductor integrated circuit is formed by using a reduction projection exposure technique, an X-ray exposure technique, an electron beam drawing technique, or another lithography technique, the position of a sample placed on a stage. Must be accurately detected and then, for example, precise alignment with a reticle or the like must be performed. Further, as semiconductor devices have been highly integrated in recent years, even higher alignment accuracy is required, and most recently, 0.0
It is required to perform precise alignment with an error of 5 μm or less. When the alignment mark provided on the surface of the sample (the surface on which the pattern is formed) is used for this alignment, position detection error due to uneven coating of resist or damage to the mark occurs. Because
There is a limit to the high precision. Therefore, for example, Japanese Patent Publication Sho 55
As disclosed in Japanese Patent Publication No. 46053/1994, a method of providing an alignment mark on the back surface of the sample has been proposed. Further, JP-A-62-160722 and JP-A-63-2243.
The publication No. 27 also describes an alignment method using a mark provided on the back surface of the sample. These methods are
The position of the mark provided on the back surface of the sample is detected, and a pattern is formed on the front surface of the sample in accordance with the position of the mark.

[0003]

As described above, when the method of detecting the position of the mark provided on the back surface of the sample is adopted, for example, the alignment accuracy of 0.2 .mu.m or later required for the device is 0.2. In order to achieve 03 μm, it is necessary to solve the following new problems.
It is a problem that error factors that were almost negligible until now can no longer be ignored.
That is, when the sample surface deviates from the set reference surface due to the inclination of the sample, a positional deviation occurs between the back surface on which the mark is provided and the surface on which the pattern is to be formed due to the thickness of the sample,
A position detection error occurs due to this position shift. This error is no longer negligible for the alignment accuracy of 0.03 μm. In order to correct the position detection error caused by the tilt of the sample described above, it is possible to attach a tilt detector for detecting the tilt angle of the sample,
The attachment of such a tilt detector has a drawback that the device is complicated and large. Therefore, there is a demand for development of a position detection method using a back surface mark that can easily correct a position detection error caused by the tilt of a sample without providing such a tilt detector. further,
Since the detector for detecting the position of this back mark is provided on the back side of the sample, it is necessary to devise such as embedding the detector in the sample stage on which the sample is placed.
For that reason, the appearance of a sufficiently miniaturized detector is expected. However, in the above-mentioned prior art, no consideration has been given to such a point.

Further, when the position of the back mark is detected as described above, a sample stage for placing and moving the sample is provided on the back side of the sample. Then, in order to place the sample on the sample stage, at least a part of the area on the back surface of the sample is brought into close contact with the upper surface of the sample stage. Since the mark position cannot be detected in the sample back surface region which is in close contact with the sample stage upper surface, it is necessary to detect the mark position in a limited region other than this part. Therefore, it is necessary to arrange the mark position detection device in a limited space facing the back surface area of the sample that is not in close contact with the upper surface of the sample stage, and at the same time, a partial area on the back surface of the sample (in close contact with the sample stage There is a serious restriction that the position of the mark provided in the area) cannot be detected. For example, when a method of vacuum-sucking a sample on a sample stage is adopted, it is not possible to detect the position of the mark in the back surface region of the sample that is vacuum-sucked. In addition, when using an optical mark position detector, a method has been proposed in which the sample stage is made of a material that is transparent to the light used for mark position detection. Regarding the mark provided at the position corresponding to the portion where the vacuum suction groove is provided, the mark position cannot be detected by the optical detection means.

On the other hand, since it is necessary to keep the surface of the sample on which a desired pattern is to be formed as flat as possible, it is desirable that the vacuum suction positions on the rear surface of the sample be evenly distributed over as wide a range as possible. Therefore, an opening (a portion that does not contact the back surface of the sample) or a transparent portion (a portion that is transparent to the light used for optical position detection) is provided at a predetermined interval on the sample stage. Only marks at positions corresponding to are detected by optical means. Now, when detecting the position of a conventional alignment mark provided on the surface of a sample, this mark is formed at the same time as the pattern of a semiconductor integrated circuit, etc. It was equivalent to detecting the typical pattern position. When detecting the position of the mark provided on the back surface of the sample, it is common to start by obtaining the pattern position on the surface of the sample corresponding to the mark position. At this time, when the distance between the openings or the transparent portions provided on the sample stage matches the chip size or pattern pitch of the semiconductor integrated circuit or the like formed on the sample surface or is an integral multiple, The mark position can be detected without any problem. But,
If this condition is not satisfied, mark position detection becomes impossible. Therefore, in the latter case, the sample stage must be replaced with another sample stage having a vacuum suction port (or suction groove) array in which openings for optical position detection are arranged at intervals according to the chip size. There was a troublesome problem. As described above, in the conventionally proposed method in which the positioning mark is provided on the back surface of the sample and the position of the mark is detected by detecting the position of the mark, the sample stage A new position is formed on the surface of the sample by detecting the accurate position on the surface of the sample, such as a semiconductor integrated circuit having an arbitrary chip size or pattern pitch, without the need for replacement No consideration was given to facilitating the alignment of the pattern to be performed.

Therefore, an object of the present invention is to provide a position detecting method for detecting the position on the sample surface by detecting the position of the position detecting mark provided on the back surface of the sample. A new position detection method that eliminates the detection error of the sample surface position due to the sample inclination without adding an inclination detector for detecting the inclination angle of the sample and enables highly accurate sample surface position detection. Is to provide. Another object of the present invention is to provide a simple and compact mark position detecting device suitable for carrying out the position detecting method according to the present invention. Still another object of the present invention is to provide a novel pattern forming method using the position detecting method according to the present invention, which is particularly suitable for use in manufacturing a highly integrated semiconductor device. Is. Still another object of the present invention is to provide a novel pattern forming apparatus suitable for carrying out the above-described pattern forming method of the present invention. Still another object of the present invention is to change the sample size for any chip size without changing the sample stage that holds the sample by vacuum suction even when the chip size of the semiconductor device to be manufactured changes. It is an object of the present invention to provide a novel pattern forming method capable of effectively utilizing an alignment method using a mark provided on the back surface and a novel pattern forming apparatus suitable for carrying out the method.

[0007]

The above-mentioned object of the present invention is achieved as follows. That is, in the position detecting method according to the present invention, when the position on the sample front surface is detected by detecting the position of the position detection mark formed on the back surface of the sample using the position detector, the position detector described above is used. As a detector, a detector whose mark position detection value includes a deviation depending on the tilt angle of the sample is used, and the back surface of the sample represented by the product of the deviation, the tilt angle of the sample, and the thickness of the sample. The various parameters in the position detector are selected in advance so that the positional deviation amounts between the sample surface and the sample surface are exactly equal.
By doing so, the mark position detection value including the above deviation always represents an accurate position on the sample surface, regardless of the sample tilt angle, whereby the sample tilts. Even in this case, by detecting the position of the position detection mark on the back surface of the sample, the accurate position on the front surface of the sample can be known. In other words, even if the sample is tilted, the position on the sample surface can be accurately known as if it had no effect, and accurate alignment of a pattern to be newly formed on the sample surface is possible. Becomes Further, in the pattern forming method according to the present invention, the sample is formed into a desired pattern by finely moving the sample while accurately grasping the position on the sample surface by using the position detecting method and apparatus according to the present invention described above. After being accurately aligned with the position, a desired pattern is formed by exposure or drawing. Further, in the pattern forming method according to the present invention, the positions of the marks arranged on the back surface of the sample at a predetermined pitch are detected, and the position of the pattern to be formed on the sample surface is calculated from the detection result. Then, a desired pattern is formed by exposure or drawing at the position obtained by this calculation.

[0008]

As shown in FIG. 1, when the sample 1 is tilted from the set reference plane by the angle θ, the actual position of the mark 2 provided on the back surface of the sample (distance δ from the reference position) and the mark When the thickness of the sample 1 is d, a deviation of d sin θ from the true position (distance ω from the reference position) on the sample surface corresponding to 2 occurs. Therefore, the actual position (distance δ) of the mark 2 is detected, and the detected value is directly used as the true position (distance ω) on the sample surface.
In such a case, as a matter of course, a position detection error ε such that ε = (δ−ω) = d · sin θ (1) occurs. For example, when the thickness d of the sample is 600 μm and the tilt angle θ of the sample is 5 seconds, the error amount ε is about 0.015 μm. This value is
In the case where a strict alignment accuracy of 0.03 μm is required, it cannot be ignored anymore.
Therefore, in the present invention, the mark 2 provided on the back surface of the sample is apparently used by using the position detecting optical system 19 that includes the deviation ΔE that exactly cancels the error amount ε in the position detection value. Position (δ + ΔE) is detected so that this apparent position just represents the true position (ω) on the sample surface. That is, the apparent mark position detection result (δ + ΔE) by the position detecting optical system 19 is just equal to the true position (ω) on the sample surface corresponding to the mark 2, that is, ω = δ + ΔE. (2) The position detecting optical system 1 described above is formed so that the following relationship is established.
The various parameters of 9 are selected in advance. formula
ΔE of the second term on the right side of (2) is a deviation included in the position detection value by the position detecting optical system 19, and this deviation changes in accordance with the change of the tilt angle θ of the sample (that is, ΔE
= −ε = −d · sin θ), the error ε caused by the inclination of the sample is always canceled regardless of the inclination angle of the sample. By detecting the mark position provided, the true position on the sample surface corresponding to the mark position can be accurately obtained.

FIG. 2 shows an example of the structure of a position detecting optical system in which the detected value includes a deviation ΔE that cancels out the position detecting error ε caused by the inclination of the sample. The detection principle will be described in further detail with reference to FIG. 2 and 3, the phase of the diffracted light obtained when the mark 2 on the back surface of the sample is illuminated by the two light beams 11a and 11b having the wavelength λ is the position δ of the mark 2 and the illumination position A of both illumination lights. , B depending on the distance t measured in the direction perpendicular to the reference plane. Assuming that the distance between the two illumination lights is L and the thickness of the sample is d, between the above-mentioned distance t between the illumination positions and the tilt angle θ of the sample, from the figure, θ = tan ~ ¹ (t / L) ≈ t / L (3) The following relational expression holds. Further, the phase difference equivalent φ (ε) depending on the amount ε of displacement between the position δ of the mark 2 and the true position ω on the sample surface due to the inclination of the sample is P (ε) at the pitch of the mark 2.
Then, φ (ε) = 4πε / P (4) On the other hand, the phase difference equivalent φ (ΔE) due to the interval t between the illumination positions, which is a function of the tilt angle θ of the sample, is determined from the geometrical relationship shown in FIG. It is expressed by the following equation for the interval t. φ (ΔE) = 4πt / λ (5) In other words, the phase change φ (ε) represented by equation (4) and equation (5)
If the pitch P of the mark 2 and the interval L between the two detection lights are selected in advance so that the phase change φ (ΔE) represented by is exactly equal, the phase shift at an arbitrary sample tilt angle can be canceled. be able to. Therefore, when the detection light interval L that satisfies the above relationship is solved by simultaneous equations (1) to (5), a solution of L = λ · d / P (6) is obtained. .. For example, the mark pitch P is 6μ
m, the sample thickness d is 600 μm, the detection light wavelength λ is 633
It is understood that the beam interval L should be 63.3 μm, where nm is set.

As another example, by using a non-imaging type position detecting optical system as shown in FIG.
A deviation ΔE depending on the tilt angle θ of the sample can be generated, and the deviation ΔE can cancel the error amount ε generated due to the tilt of the sample. That is, as shown in FIG. 13, the photodetector 51 is arranged at a position apart from the conjugate position C with respect to the sample surface in the optical system including the lenses 50a and 50b by a distance k upward. With this configuration, when the sample 1 is tilted, a deviation ΔE occurs in the detection result of the photodetector 51. This deviation ΔE
Is also a function of the tilt angle θ of the sample 1.
If the distance k is set in advance so that the deviation ΔE becomes equal to the positional deviation amount ε on the sample surface, the detection error due to the tilt of the sample (that is, the positional deviation amount).
ε can be offset. Besides this, as shown in FIG. 17, there is also a method of utilizing the optical path difference of the reflected light due to the inclination of the sample 1. That is, this is a method that utilizes that the optical path difference between the reflected lights 12 and 13 from the sample 1 changes depending on the tilt angle θ of the sample 1. It is to be noted that FIG. 18 shows how an optical path difference is generated between the two reflected lights when the sample 1 is tilted. As described above, in the method of recognizing the position on the sample surface by detecting the position of the mark provided on the back surface of the sample, make sure that there is no error in the position recognizing result on the sample surface even when the sample is tilted. The three types of position detection methods that can be performed and the device configuration for implementing the method have been shown. However, other than these methods, a deviation occurs in the position detection result depending on the tilt angle of the sample, and this deviation Any position detection method can be used as long as it has a function of canceling the amount of positional deviation generated between the sample back surface and the sample front surface due to the inclination of the sample.

Next, a method of arranging the position detecting marks on the back surface of the sample according to the present invention will be described. In the step of detecting the positions of the marks provided on the back surface of the sample at a predetermined interval, by detecting the positions of at least two marks, the translation error of the sample mounting position, the rotation error in the sample plane, and Obtain a uniform amount of expansion and contraction. The local in-plane deformation amount of the sample can also be obtained by detecting a large number of mark positions. Further, in the step of calculating the position of the circuit pattern to be newly formed by using these detection results, each position where the pattern is newly formed on the sample surface is calculated by the interpolation method. In this case, since the calculation is performed by the interpolation method, the pattern formation position on the sample front surface and the mark position provided on the sample back surface do not necessarily have to be in one-to-one correspondence. Since the position of the sample stage on which the sample is placed is precisely measured by the laser length measuring device, the sample can be accurately positioned at a position where a new pattern is to be formed.

[0012]

Embodiments of the present invention will now be described in detail with reference to the drawings. [Embodiment 1] FIG. 2 shows the basic construction of an optical system for position detection according to an embodiment of the present invention. According to this optical system, in the so-called back surface mark detection method in which the position of the mark provided on the back surface of the sample is detected and the position on the front surface of the sample corresponding to the mark position is obtained, the back surface of the sample can be It is possible to cancel the position detection error that occurs due to the positional deviation from the sample surface. First, the method of position detection will be described. FIG. 4 shows a schematic configuration of an optical system of a known roughness measuring device. This is described, for example, in APPLIED OPTICS Vol.20, No.4 / p610. This measuring device basically uses the principle of the Nomarski interferometer. Here, a laser light source 10 that emits light of two frequencies having slightly different wavelengths in the form of linearly polarized light is used. A light beam 11 from the light source is expanded via a lens system 17 and then split into two by a beam splitter 15. One of the split light beams is incident on the analyzer 8a, where heterodyne interference occurs. This interference light is detected by the photodetector 7a.
The reference signal S serving as a reference is detected by
You get r. The other split beam is incident on the Wollaston prism 5, where it is separated into a P-polarized beam (frequency ν ₁ ) beam 11a and an S-polarized beam (frequency ν ₂ ) beam 11b. The separated beams 11a and 11b are focused by the objective lens 3 and irradiate two points A and B on the back surface of the sample, respectively. The light reflected at both irradiation points A and B is
It passes through the objective lens 3 and enters the Wollaston prism 5. Here, the reflected lights from the irradiation points A and B overlap each other to form a single beam, which is guided to the analyzer 8b. The analyzer 8b is arranged at 45 degrees with respect to the plane of polarization of the light beam 11.
Since they are arranged so as to have an inclination of °, heterodyne interference occurs here. A detection signal Sd is obtained by detecting this interference light with the photodetector 8b. The cycle of the detection signal Sd is the same as the cycle of the reference signal Sr. Therefore, when the sample 1 tilts, the optical path difference between the detection lights 11a and 11b changes according to the tilt angle of the sample, and the phase difference of the detection signal Sd with respect to the reference signal Sr changes.

Basically using the detection principle shown in FIG.
FIG. 2 shows an example in which the position of the back mark 2 can be detected. As in the case of the detection method of FIG. 4, here also a two-frequency laser light source 10 that emits laser light of two frequencies having slightly different wavelengths in the form of linearly polarized light is used.
The beam 11 from the laser light source 10 is a P-polarized beam 11a and an S-polarized beam 1 by the Wollaston prism 5a.
1b, and the lattice-shaped mark 2 provided on the back surface of the sample is illuminated via the objective lens 3. The distance between the irradiation points A and B of both beams is L. At this time, irradiation point A,
Focusing only on the + 1st order light 12a, 12b and the −1st order light 13a, 13b among the diffracted light generated from B, the phases of these 1st order diffracted light are, as described above, It changes depending on the angle. The phase change φ ₁ can be obtained as a phase difference between the reference signal Sr from the detector 7a and the detection signal Sd from the detector 7b, as shown in the lower part of the figure. + 1st order diffracted light 12a, 12
b and the −1st-order diffracted lights 13a and 13b are made into parallel beams by the objective lens 3, and by using the polarization beam splitters 4a and 4b on the Fourier transform surface, the + 1st-order diffracted light from one irradiation point (for example, the irradiation point The + 1st-order diffracted light 12b from B and the -1st-order diffracted light from the other irradiation point (for example, the -1st-order diffracted light 13a from the irradiation point A) are selected. The selected beams 12b and 13a are condensed by the condensing lens 6, and the Wollaston prism 5b is placed at the intersection of the two beams to cause heterodyne interference, and the interference light is transmitted through the analyzer 8b. It is detected by the detector 7b. The phase difference φ ₁ between the obtained detection signal Sd and the reference signal Sr is, as shown in FIG. 3, the position of the mark 2 is δ, the pitch of the mark 2 is P, the wavelength of the detection light is λ, both irradiation points. If the amount of deviation between A and B in the direction perpendicular to the reference plane is t, then φ ₁ = (4πδ / P) + (4πt / λ) (7) Here, the direction of the shift amount t shown in FIG. 3 is assumed to be a negative direction. At this time, if the thickness of the sample is d, the tilt angle of the sample is θ, and the interval between the irradiation beams 11a and 11b is L, the amount of deviation between the sample back surface and the sample front surface due to the tilt angle θ of the sample (that is, , Position detection error) ε becomes ε = d · sin θ (8) Here, θ = tan to ¹ (t / L) ≈t / L (9) Then, the phase change amount caused by the shift amount ε and the phase change amount caused by the shift amount t become equal to each other.
The mark pitch P and the beam interval L may be selected. That is, (4πt / λ) = (4πε / P) ... (10) should be established. That is, it is understood from the formulas (7) to (10) that the following formulas should be satisfied. d / P = L / λ (11) From this equation, for example, the wavelength λ = 633 as the light source 10.
When using a He-Ne laser of nm, the thickness d of the sample 1
= 600 μm and mark pitch P = 6 μm, it is understood that the beam interval L should be 63.3 μm. The beam interval L is determined by the beam separation angle ξ of the Wollaston prism 5a and the focal length f of the objective lens 3. That is, when the required beam spacing L is known, the beam separation angle ξ is determined by the following equation. ξ = 1 / (2f) (12) That is, it is sufficient to use a Wollaston prism having such a separation angle ξ.

Next, the allowable value ΔL of the setting error when setting the required beam interval L will be considered. As shown in FIG. 5, when the allowable value ΔL has a finite value, a position detection error Δε due to the tilt occurs. Here, for the target value of overlay accuracy of 0.03 μm (3σ), Δε <0.01
It is necessary to suppress to about μm. Further, the tilt angle θ due to the warp of the sample 1 is usually about ± 0.15 m rad. At this time, Δε is expressed by the following equation. Δε = (ΔL / L) θ · d (13) When the above formula is arranged, the allowable value Δ of the setting error of the beam interval L is Δ
It is understood that L needs to be suppressed within about 15% of the required beam interval L. Therefore, considering some margin, Δ
It is sufficient to keep L within 10% of L. That is, the beam interval L is 0.9 (d · λ / P) <L <1.1.
It is desirable to select a value within the range of (d · λ / P).
Further, as shown in FIG.
The irradiation spot diameter γ of a and 11b is preferably γ = n · P (14), where n is a natural number. This is because the influence of the step of the mark 2 can be suppressed to a low level. Also,
When the thickness of the sample changes, the pitch P of the mark 2 may be determined based on the following equation obtained by modifying the equation (11). P = d · λ / L (15) Next, the allowable value ΔP of the setting error of the mark pitch P will be considered. From the equation (10), when the mark pitch P has an error of ΔP, the position detection error Δε caused by the error
Is Δε = (t / λ) ΔP ... (16) Here, the setting error allowable value Δ of the beam interval L is first
If the same conditions as when L is calculated are applied, the allowable value ΔP of the setting error of the mark pitch P should be ± 0.7 μm. Since this allowable value ΔP is about 12% with respect to the required mark pitch P, even if some margin is taken into consideration,
It can be said that ΔP is sufficient if it is within 10% of P. That is, the mark pitch P is 0.9 (dλ / L) <P <1.1.
It is desirable to select a value within the range of (dλ / L). By doing so, the position of the mark provided on the back surface of the sample can be detected without providing the tilt detector for detecting the tilt angle of the sample, and the desired pattern formation position on the surface of the sample can be detected. Can be obtained with high accuracy.

As a method of forming the mark 2 described above, it is convenient to use a laser marker. Further, the formed mark 2 is, as shown in FIG. 7, a protective film 1 such as an oxide film.
It is preferably protected by 8. Further, as shown in FIG. 8, by forming the mark 2 in the protective film 18b provided so as to overlap the protective film 18a, it is not necessary to directly process the sample (for example, Si substrate) 1. By doing so, it is possible to prevent the occurrence of defects due to transfer or the like inside the Si substrate. In addition, for example, electronic materials January 1991 issue 55,
Scanning tunneling microscope as described on page 60
It is also possible to form a mark having an atomic order size by using a fine processing technique using (STM: Scanning Tunneling Microscope) and use the mark for position detection.
It is even more preferable to use a mark having such an atomic order size because it is possible to improve the position detection sensitivity and reduce the detection error due to the deformation and distortion of the sample. Also, sample 1
Is not limited to the Si substrate, but may be a GaAs substrate, a glass substrate, or a plastic substrate, or may be a dummy substrate in which a photosensitive thin film is attached to these substrates.

FIG. 9 shows a modification of the above embodiment. The configuration of FIG. 9 corresponds to the configuration shown in FIG. 2 in which the positions of the incident light on the sample and the reflected light from the sample are replaced. The laser beam from the dual-frequency laser light source 10 is collimated by the lens 9 and then split into two beams by the polarization beam splitter 4. One of the reflected beams is incident on the sample through the objective lens 3 and the other is transmitted. The reflected beam is reflected by the reflection mirror 22 and then is incident on the sample through the objective lens 3. At this time, the reflection angles of the beam splitter 4 and the reflection mirror 22 are set so that the irradiation points of both beams on the back surface of the sample are separated from each other by the distance L. Of the light diffracted by the mark 2 provided on the back surface of the sample, the diffracted light obtained in the direction perpendicular to the back surface of the sample (± 1st order diffracted light)
Only the object is taken out and condensed by the objective lens 3. By using the Wollaston prism 5 having a separation angle that matches the intersecting angle of both the collected beams, the two beams are superposed on one another, and then the polarizing plate 8 causes heterodyne interference to generate the interference light. It is detected by the detector 7.
With this configuration, the position detecting device as a whole can be further downsized. Further, the thickness d of the sample 1
If you want to change the
It suffices to finely adjust the reflection angles of the reflection mirror 22 and the polarization beam splitter 4 and set again so that the beam interval L satisfies the expression (11). For fine adjustment of the reflection angle, reflection angle adjusting means 42a and 42b are provided.

[Embodiment 2] FIG. 13 shows the configuration of a position detecting device according to another embodiment of the present invention. This example
The non-imaging type position detection optical system is used to perform highly accurate position detection. In Example 1 described above, the mark provided on the back surface of the sample was irradiated with two light beams, and the position on the sample surface was detected from the phase of light reflected and diffracted from the sample. ,
By using a non-imaging type position detection optical system as shown in FIG. 13, it is possible to cause a deviation in the position detection value depending on the tilt angle of the sample, and the deviation causes the tilt of the sample. It is possible to cancel the detection error ε. First, the principle of the position detection method using the non-imaging type position detection optical system will be described with reference to FIG. The laser beam emitted from the laser light source 101 passes through the polarization plate 8 and the collimator lens 9, is guided to the λ / 4 plate 44 by the polarization beam splitter 4, and then is condensed by the objective lens 3 to be on the sample 1. Irradiate the mark 2. The + 1st-order diffracted light 12, the -1st-order diffracted light 13, and the 0th-order diffracted light 45 reflected and diffracted from this irradiation position are again converted into the objective lens 3 and
After passing through the / 4 plate 44 and the polarization beam splitter 4, the light is focused by the condenser lens 6 and is incident on the two-divided light receiving element 43 which is installed at a position separated by a distance k upward from the conjugate position C of the optical system. The incident state of the light beams 12, 13, 45 on the light receiving surface of the light receiving element 43 will be described with reference to FIG. The 0th-order diffracted light 45 is incident on the center of the light-receiving surface of the two-divided light-receiving element 43, and the + 1st-order diffracted light 12 and the -1st-order diffracted light 13 are incident on both sides thereof. With such an arrangement, when the mark 2 is displaced in the horizontal direction (X direction), the light intensities of the regions C and D change due to interference. Therefore,
When the difference value of the outputs from the light receiving units 43a and 43b is obtained by the subtractor 46, the difference value changes linearly with respect to the detection position (X) as shown in the lower part of the figure.
This makes it possible to detect the mark position. The above-mentioned light receiving element 43 is not limited to two divisions, and the same result can be obtained by using any of two or more divisions.

Next, the state of the detection signal when the sample 1 is tilted will be described with reference to FIG. Since the light receiving element 43 in this mark position detecting optical system is separated from the conjugate position C of the optical system by the distance k as described above, the respective light beams 12, 13, 45 are laterally offset by the distance j, and the light receiving element 43. Will be incident on. In such a case, the 0th-order diffracted light 45 is largely incident on the light receiving portion 43a side and only slightly incident on the light receiving portion 43b side, so that the output on the light receiving portion 43a side and the output on the light receiving portion 43b side are generated. An imbalance occurs between the two, and as a result, the output difference value from the subtractor 46 includes an offset (deviation). Here, by providing the distance i between the region C ′ and the region D ′, the offset amount included in the above output difference value can be made to have a linear relationship with the tilt angle θ of the sample 1. it can. Further, when the thickness d of the sample 1 is different, the distance k may be adjusted according to the equation (8) so that the shift amount ε shown in FIG. 1 becomes equal to the offset amount. With this configuration, it is possible to obtain a highly accurate position detection device that does not cause an error in the position detection value due to the inclination of the sample 1.

[Third Embodiment] FIG. 17 shows the arrangement of a position detecting apparatus according to a third embodiment of the present invention. In this embodiment, the principle of an interferometer is applied so that highly accurate position detection can be performed. In FIG. 17, the beam emitted from the laser light source 101 is irradiated onto the mark 2 formed on the back surface of the sample 1 via the mirror 47c. + 1st-order diffracted light 1 reflected and diffracted from this irradiation position
The second and −1st order diffracted light 13 are reflected by mirrors 47a and 47b, respectively.
After being reflected by, the light is caused to interfere on the slit 48. Light and dark of interference fringes caused by this interference are detected by the photodetector 7 through the slit opening. If the opening width of this slit opening is set to just 1/2 of the pitch of the interference fringes, a detection signal with a high S / N ratio can be obtained. In this state, when the mark 2 moves in the X direction, the output signal of the photodetector 7 changes as shown in the lower part of FIG.
The position of the mark 2 can be detected from the output signal.

Next, how the diffracted light path changes when the sample 1 is inclined by θ from the reference plane will be described with reference to FIG. When the sample 1 tilts, the optical paths of the first-order diffracted lights 12 and 13 shift from the positions shown by the broken lines to the positions shown by the solid lines. Therefore, an optical path difference occurs between the first-order diffracted lights 12 and 13, and the position of the interference fringes generated on the slit 48 also shifts in the lateral direction. That is, the tilt of the sample 1 causes the mark position detection result to include an offset. The distance between the mirrors 47a and 47b and the pitch P of the detection mark 2 are selected so that the offset amount is exactly equal to the error amount ε shown in the equation (8). The error can be canceled out, and the sample surface position can be detected with high accuracy.

[Embodiment 4] FIG. 10 shows an embodiment in which the rear surface mark detection method according to the present invention is applied to a reduction projection exposure apparatus. The backside mark detection device configured to cancel the position detection error caused by the inclination of the sample 1 as shown in Examples 1 to 3 is currently the mainstream of pattern formation devices for manufacturing semiconductor devices. It is particularly suitable for use in position detection in a reduction projection exposure apparatus. Figure 1
At 0, the original image pattern formed on the reticle 31 is illuminated by the monochromatic light source 30 and is reduced and projected onto the surface of the sample 1 via the reduction projection lens 24.
As described above, when the reduction projection lens 24 is present between the reticle 31 and the sample 1, it is difficult to dispose the mark detector on the upper side (front side) of the sample. It is particularly effective to use the mark position detecting device 19 to detect the mark on the back surface of the sample from the lower side (back surface side) of the sample. While the mark position detection device 19 detects an accurate position on the sample surface, the sample fine movement device 25 finely moves the sample in the XY directions to reduce and project the original image pattern at a desired pattern formation position on the sample surface. As described above, alignment is performed on the sample side. In this alignment, the position of the mirror 38 attached to the sample table 14 is precisely measured by the laser length measuring instrument 37, and the sample fine movement device 25 is controlled via the control device 39 based on the measured value. Done by.

The position of the reticle 31 is detected by detecting the mark 33 provided on the reticle via the reticle mark detecting optical systems 26a and 26b.
This is done by detecting in b. Based on this reticle position detection result, the reticle fine movement device 40 finely moves the reticle to align the reticle 31. As a result, the reticle 31 is positioned with respect to the absolute reference. A focusing device 32 is provided for focusing the reduction projection lens 24 on the pattern formation position on the sample surface. This focusing device may be configured to maintain a constant relative distance between the sample surface and the reduction projection lens by using a gap holding means utilizing the principle of an air micrometer, for example. Also,
The thickness of the sample is measured by the sample thickness measuring device 41, and it is determined whether or not the deviation included in the position detection value of the mark position detecting device 19 needs to be corrected according to the measured sample thickness. Done.

[Embodiment 5] FIG. 11 shows an embodiment in which the back surface mark detection method according to the present invention is applied to the alignment of an electron beam drawing apparatus. As shown in FIG. 11, in the electron beam drawing apparatus, the electron gun 34 and the electron lenses 35a, 35b, 35c are controlled by the pattern data stored in the drawing data storage unit 36, so that the desired surface on the sample 1 is controlled. A pattern is formed. On the lower side (back side) of the sample 1, the mark position detecting device 19 according to the present invention is installed. The position detecting method in the mark position detecting device 19 is the same as that described in the first, second, or third embodiment, and therefore detailed description thereof is omitted here. As described above, according to the present invention, the alignment of the electron beam drawing apparatus can be performed with the highest accuracy ever.

[Embodiment 6] FIG. 12 shows an embodiment in which the back mark detection method according to the present invention is applied to alignment in a 1: 1 proximity exposure apparatus. As shown in this figure, the rear surface mark detection method according to the present invention is 1: 1 in the same manner as in the case of the above-described reduction projection apparatus or electron beam drawing apparatus.
It is also applicable to the proximity exposure apparatus. In this 1: 1 proximity exposure apparatus, a mask 2 on which a desired pattern is drawn
In the state where 0 and the sample 1 are in close proximity to each other, the pattern is exposed on the surface of the sample 1 by exposure with X-rays or the like from the X-ray light source 49. The pattern mask 20 is held on the mask fine movement device 21, the position of the mirror 38 'is precisely measured by the laser length measuring device 37, and the mask fine movement device 21 is controlled via the control device 39 based on the measured value. The precise positioning of the mask 20 is performed by controlling the. The method for detecting the mark position in this embodiment is the same as that in the first, second, or third embodiment, and the pattern formation method is the same as in the fourth, fifth embodiment. Therefore, detailed description is omitted here.

[Embodiment 7] FIG. 19 shows still another embodiment of the present invention. This embodiment relates to a case where the arrangement pitch of the pattern to be formed on the sample front surface and the arrangement pitch of the position detection marks provided on the back surface of the sample are different in the reduction projection exposure apparatus using the position detection method according to the present invention. is there. In the figure, light from a light source 30 illuminates a reticle 31 on which a pattern such as a semiconductor integrated circuit is drawn via a condenser lens 60. The light diffracted by the pattern on the reticle 31 is reduced by the reduction projection lens 24.
To reach the surface of the sample 1 via. As a result, the pattern on the reticle 31 is reduced and projected on the surface of the sample 1. The sample 1 is adsorbed and fixed on the sample table 14, and the sample table 14 is moved by a stage driving device 62 in the x, y, and z directions, respectively.
Since it is mounted on 64 and 63, the exposure is repeated while the sample 1 is positioned at a predetermined pitch by operating the driving device 62, so that a plurality of samples 1 are formed on the surface of the sample 1 at the above-described predetermined pitch. The above pattern can be formed. Mirror 3 by laser length measuring device 37
The position of the sample table 14 can be accurately obtained by measuring the position of 8. Then, the obtained sample stage position is compared with the target position, and the difference is fed back to the driving device 62 via the control device 39. Therefore, if the target position for forming the pattern is given, Positioning of the sample table 14 at the target position is accurately performed.

The alignment between the reticle 31 and the sample 1 is performed in the order shown in FIG. First, in step 70, the reticle 31 is positioned. That is, by mounting the reticle 31 on the fine movement stage 40 and finely positioning the fine movement stage 40 by the control device 39, the center position of the reticle 31 coincides with the optical axis of the exposure optical system. Next, in step 71, position detection of the plurality of position detection marks 2 already provided on the back surface of the sample 1 is performed. That is, the position of the position detection mark 2 provided on the back surface of the sample 1 is set to the opening 8 provided on the sample table 14.
It is detected by the mark position detecting device 19 fixedly installed on the base 66 via 0. Then, in step 72,
From the error in the mark position (error in both the x and y directions) for at least two marks, the translation error of the sample placed on the sample table 14, the rotational mounting error in the sample surface, and the error in the sample surface. The isotropic expansion / contraction amount is calculated by the calculation unit 61. Furthermore, if mark position errors for a large number of position detection marks 2 are detected, the average mounting error such as translational error, in-plane rotational mounting error and expansion / contraction amount of the sample 1 is increased by statistical calculation processing. It can be calculated with accuracy. This average mounting error calculation is performed at step 72. Furthermore, when it is determined in the determination step 73 that the local deformation amount of the sample 1 also needs to be obtained,
In the calculation step 74, the local deformation amount of the sample 1 is calculated and obtained. Based on the above detection results, in the exposure position calculation step 75, individual positions where a pattern is to be formed on the sample 1 are calculated. Finally, in the exposure step 76, the stage drive unit 62 exposes the pattern while sequentially positioning the sample at the plurality of pattern formation positions obtained by the above calculation.

As shown in FIG. 21, the surface of the sample table 14 is provided with a plurality of openings 80 having pitches Sx and Sy in the x and y directions, respectively. The sample 1 is attached to the surface part of the sample table left between the openings.
2 is provided with a suction groove for vacuum sucking
In FIG. 1, the suction groove is not shown. On the other hand, on the back surface of the sample 1, the mark 2 for position detection is provided with the opening 8
They are provided at the same pitch as the 0 pitch Sx, Sy.
As a result, the above-mentioned opening portion 80 is provided from the lower side of the sample table 14.
Through it, all the position detection marks 2 provided on the back surface of the sample 1 can be directly observed. Here, the arrangement pitch (or chip size) Px, Py of the plurality of patterns 81 to be formed on the front surface of the sample 1 and the arrangement pitch Sx, S of the position detection marks 2 previously provided on the back surface of the sample 1.
It is not necessary to match y. In this embodiment, the arrangement pitches Px, Py of the patterns 81 to be formed on the surface of the sample 1 are arranged.
By detecting the position detection mark 2 provided on the back surface of the sample from the back side of the sample with pitches Sx and Sy that are completely unrelated to, the mounting error of the sample 1 and the local deformation amount on the surface of the sample are obtained. Then, the individual positions of the pattern 81 to be formed on the surface of the sample 1 are calculated. Therefore, when manufacturing a semiconductor integrated circuit or the like, the position of the position detection mark 2 previously provided on the back surface of the sample 1 is detected even when the first layer pattern is exposed.
In the present embodiment, the arrangement pitch Sx and Sy of the openings 80 provided in the sample table 14 are both set to 25 mm. A material (for example, glass) that is transparent to the light used for detecting the mark position may be embedded in the opening 80. Further, the entire sample supporting portion of the sample table 14 may be made of a material transparent to the mark position detecting light. Also, omit the groove for vacuum suction, instead,
A method of electromagnetically fixing the sample may be adopted.

The calculation of the individual positions of the pattern 81 to be formed on the surface of the sample 1 is performed by the arithmetic unit 61 shown in FIG. An example of this calculation will be described with reference to FIG. The four marks 91, 92, 93, 94 shown in the figure
Is a part of a large number of position detection marks 2 provided on the back surface of the sample 1 in advance, and shows a regular mark position when the sample 1 is accurately placed on the sample table 14. ..
The arrangement pitch of these marks is Sx in the x direction and Sy in the y direction. Now, due to the mounting error of the sample 1, the four marks described above are displaced from their respective predetermined positions, and
It is assumed that detection is performed at the positions indicated by ', 92', 93 ', and 94'. Then, the positional displacement amounts of the respective marks at this time are (X ₁ , Y ₁ ), (X ₂ , Y ₂ ), (X ₃ ,
Y _3), and was detected with (X _4, Y _4). On the other hand, it is assumed that one of the plurality of patterns 81 to be formed on the surface of the sample 1 is the pattern 90 indicated by the broken line. The regular exposure position of the pattern 90 is the four marks 9 described above.
1, 92, 93, and 94, and the mark pitch Sx in the x direction is Δx: 1− as shown in the figure.
At the ratio of Δx, the mark pitch Sy in the y direction is Δy: 1−
It is assumed that the positions are internally divided by the ratio of Δy. Since the array pitch of the patterns 81 to be formed on the front surface of the sample 1 and the array pitch of the position detection marks 2 previously provided on the back surface of the sample 1 are both known, if the array number of the pattern 90 is specified, the above Δx , Δy values are uniquely determined. The actual exposure position 90 'of the pattern 90 is the predetermined position 9
It is a position displaced from 0 by the displacement amount (X, Y), and the displacement amount (X, Y) of the pattern position is the displacement amount (X ₁ , Y ₁ ), (X ₂ , Y ₂ ) of the mark position described above. ), (X ₃ , Y ₃ ),
It is expressed by the following equation using (X ₄ , Y ₄ ). _{X = {x 1 (1-} Δx) (1-Δy) + x 2 Δx (1-Δy) + x 3 ΔxΔy + x 4 (1-Δx) Δy} / 4 ......... (17) Y = {y 1 (1- Δx) (1-Δy) + y ₂ Δx (1-Δy) + y ₃ ΔxΔy + y ₄ (1-Δx) Δy} / 4 (18) The above equations (17) and (18) are linear approximation arithmetics. It is an expression.
The content of the calculation performed by the calculation unit 61 is not limited to this, and it is conceivable to perform calculation using a spline function or the like using the position detection results for a larger number of marks. vice versa,
Only by detecting the position of a small number (minimum of 2) of typical marks among the large number of marks, the average translation error of the sample 1 placed on the sample table 14, the in-plane rotational mounting error, etc. It is also conceivable to calculate each exposure position and calculate the exposure position.

According to the present invention, the mark 2 for position detection is used.
Is provided on the back surface of the sample where the pattern of the semiconductor integrated circuit or the like is not formed. Therefore, when the pattern of the semiconductor integrated circuit or the like is formed on the surface of the sample as in the conventional case, at the same time, for alignment in the next step, There is no need for troublesome work such as forming a required mark pattern. That is, the original image pattern on the reticle 31 does not necessarily need to include the mark pattern for alignment. Further, in addition to the reduction projection exposure apparatus shown in FIG. 19, an exposure apparatus using X-rays, an electron beam drawing apparatus, or the like is the same as that shown in this embodiment for detecting the position on the sample surface. It goes without saying that the above method can be adopted. In the present embodiment, the position detection mark provided on the back surface of the sample is arranged along the rectangular coordinate system.
Although the marks are dimensionally arranged at equal intervals, the positions of the position detection marks are not limited to this. For example, they may be arranged at unequal intervals according to the arrangement of the opening portions 80 provided in the sample table 14, Alternatively, the same effect can be obtained by arranging along the polar coordinate system.

[0030]

According to the present invention, when the position on the surface of the sample is obtained by detecting the position of the position detection mark provided on the back surface of the sample, the position detection value of the position detection mark is used as the sample. Since the deviation that changes according to the tilt angle of the sample is included and the position detection error ε caused by the tilt of the sample is offset by this deviation, the tilt angle of the sample is intentionally detected. It is possible to reduce the position detection error ε caused by the inclination of the sample to such an extent that there is almost no problem, without adding any inclination detection means for this. Further, in the present invention,
Since the position error on the surface of the sample is first obtained from the position detection result of the alignment mark provided on the back surface of the sample, and then the individual positions where the pattern of the semiconductor integrated circuit etc. should be formed are calculated. A pattern having an arbitrary chip size, which is irrelevant to the arrangement pitch of the alignment marks, can be formed on the sample surface. Also,
Since an opening for observing the mark position is provided on the sample table just corresponding to the position of the alignment marks provided on the back surface of the sample, any position can be adjusted from the back surface side of the sample through the opening. The position of the mark can be observed directly.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of a mark position detecting method according to the present invention.

FIG. 2 is a schematic configuration diagram of a mark position detection device according to an embodiment of the present invention.

FIG. 3 is a schematic diagram for explaining the operation principle of the mark position detecting device according to the present invention.

FIG. 4 is a schematic configuration diagram of a roughness measuring device using a conventional Nomarski interferometer.

FIG. 5 is a schematic diagram for explaining a setting allowable value of a light beam interval in the mark position detecting device according to the present invention.

FIG. 6 is a schematic diagram for explaining a desirable setting example of the light beam diameter in the mark position detecting device according to the present invention.

FIG. 7 is a schematic view showing one configuration example of a position detection mark provided on the back surface of the sample in the mark position detection method according to the present invention.

FIG. 8 is a schematic view showing another configuration example of the position detection mark provided on the back surface of the sample in the mark position detection method according to the present invention.

FIG. 9 is a schematic configuration diagram of a mark position detection device according to another embodiment of the present invention.

FIG. 10 is a schematic configuration diagram of a reduction projection exposure apparatus using a mark position detection device according to the present invention.

FIG. 11 is a schematic configuration diagram of an electron beam drawing apparatus using the mark position detecting apparatus according to the present invention.

FIG. 12 is a schematic configuration diagram of a proximity exposure apparatus using the mark position detection device according to the present invention.

FIG. 13 is a schematic diagram showing a principle configuration of a position detecting device according to still another embodiment of the present invention.

14 is a schematic configuration diagram of a position detection device specifically configured according to the principle configuration shown in FIG.

15 is a detailed configuration diagram of a light receiving element in the position detecting device shown in FIG.

16 is a schematic diagram for explaining how the light beam spot incident on the light receiving element shown in FIG. 15 changes when the sample is tilted.

FIG. 17 is a schematic diagram showing a principle configuration of a position detecting device according to still another embodiment of the present invention.

FIG. 18 is a schematic diagram for explaining the mark position detection principle in the position detection device shown in FIG.

FIG. 19 is a schematic configuration diagram of a pattern forming apparatus according to an embodiment of the present invention.

20 is a process diagram showing an alignment procedure in the pattern forming apparatus shown in FIG.

21 is a schematic diagram showing an arrangement relationship between an opening 80 provided on the sample suction surface of the sample table 14 and the position detection mark 2 provided on the back surface of the sample in the pattern forming apparatus shown in FIG.

22 is a schematic diagram for explaining a method of calculating a pattern forming position in the pattern forming apparatus shown in FIG.

[Explanation of symbols]

1: sample, 8: polarizing plate, 2:
Position detection mark, 10: laser light source, 3: objective lens, 11: light beam, 4: beam splitter, 12: + 1st order diffracted light, 5: Wollaston prism, 13: −1st order diffracted light, 6: condensing lens, 15: Beam splitter, 7: Photodetector, 19: Optical system for position detection.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Masaaki Ito 1-280, Higashi Koikekubo, Kokubunji, Tokyo Inside the Central Research Laboratory, Hitachi, Ltd.

Claims (1)

Claim: What is claimed is: 1. When indirectly measuring the position of a point corresponding to the characteristic point on the surface of the sample by detecting the position of a certain characteristic point on the back surface of the sample, The position detection value for the feature point is set to include a deviation amount that changes corresponding to the tilt angle of the sample, and the position on the sample back surface and the position on the sample front surface due to the tilt of the sample due to the deviation amount. By offsetting the amount of deviation generated between them, the position detection value for the feature point including the deviation amount always indicates the true position of the point corresponding to the feature point on the sample surface. A position detection method characterized by the above. 2. The position detecting method according to claim 1, wherein the feature point on the back surface of the sample is a mark provided on the back surface of the sample in advance. 3. The deviation included in the position detection value for the characteristic point on the back surface of the sample is an amount substantially equal to the product of the thickness of the sample and the tilt angle of the sample. The position detection method according to item 1. 4. A step of detecting the position of an alignment mark provided on the back surface of the sample substrate, and the positioning of a region on the surface of the sample substrate where a pattern is to be formed with reference to the detected mark position. In a pattern forming method including at least a step of performing and a step of forming a desired pattern on the positioned pattern forming area, the position detection value of the alignment mark in the mark position detecting step includes a sample substrate A deviation amount that varies depending on the tilt angle of the sample substrate is included, and the deviation amount caused between the position on the back surface of the sample substrate and the position on the front surface of the sample substrate due to the tilt of the sample substrate. Are canceled out, and as a result, the position detection value of the alignment mark including the deviation indicates the true position on the sample substrate surface. Turn-forming method. 5. The deviation included in the position detection value of the alignment mark is an amount substantially equal to the product of the thickness of the sample and the tilt angle of the sample. The position detection method described. 6. The alignment marks are two-dimensionally arranged at a predetermined pitch on the back surface of the sample substrate,
The pattern forming method according to claim 4, wherein the arrangement pitch of the marks is a constant pitch that is independent of the arrangement pitch of the pattern to be formed on the surface of the sample substrate. 7. Mark position detecting means for detecting the position of an alignment mark provided on the back surface of the sample substrate, and on the front surface of the sample substrate with reference to the mark position detected by the mark position detecting means. In a pattern forming apparatus having at least a positioning means for positioning an area where a pattern is to be formed and a pattern forming means for forming a desired pattern on the positioned pattern forming area, the positioning means has the mark position. Calculation means for calculating the position of the area to be patterned on the sample substrate surface from the mark position detected by the detection means, and the pattern formation area obtained by this calculation is the pattern formation position by the pattern formation means. And means for positioning the sample substrate so that they coincide with each other. Turn-forming apparatus. 8. The mark position detecting means comprises a light source for emitting light to be irradiated onto the sample substrate, a means for collecting and irradiating the light from the light source onto the sample substrate, and a mark on the sample substrate. Means for collecting the light reflected from the mark provided on the
It comprises at least a photo-detecting means having at least two light-receiving surfaces for detecting the condensed reflected light, and a computing means for obtaining a difference between detection signals from the respective light-receiving surfaces of the photo-detecting means. The pattern forming apparatus according to claim 7, wherein 9. The mark position detecting means is provided on a sample substrate, a light source for emitting light to be irradiated on the sample substrate, a means for irradiating the sample substrate with light from the light source. 8. The device according to claim 7, further comprising at least means for reflecting the reflected light from the mark, means for selecting a part of the reflected light, and means for detecting the selected reflected light. Pattern forming device. 10. The mark position detecting means comprises a light source means for emitting dual frequency light rays having slightly different wavelengths, and a first light beam splitting means for splitting a light beam from the light source means into two light beams. Means and means for detecting one of the light beams obtained by the first light beam splitting means to obtain a reference signal, and the other light beam similarly obtained by the first light beam splitting means, Second light beam splitting means for splitting the light beam into two light beams, and a grating provided on the back surface of the sample substrate for respectively collecting the two light beams obtained by the second light beam splitting means. The objective lens means for irradiating two different irradiation points of the circular mark, the positive phase component of the reflected light of one wavelength and the opposite phase component of the reflected light of the other wavelength from the reflected light from the two irradiation points. And the means to separate them It is characterized by comprising at least means for causing interference by re-superimposing the two component light beams separated and taken out, and means for detecting light obtained by this interference to obtain a detection signal. The pattern forming apparatus according to claim 7. 11. The two-frequency light beam having a slightly different wavelength is composed of two linearly polarized lights having different polarization planes, and the second light beam splitting means has an optical system including a Wollaston prism. The pattern forming apparatus according to claim 10, wherein the pattern forming apparatus is configured. 12. The beam separation angle ξ of the Wollaston prism is ξ = L / (2.multidot.L) where f is the focal length of the objective lens means and L is the distance between the two irradiation points.
The pattern forming apparatus according to claim 11, wherein the value is selected so as to substantially satisfy the relationship of (f). 13. The distance L between the two irradiation points on the back surface of the sample substrate is d, the thickness of the sample substrate, and λ is the central wavelength of the two-frequency light beam irradiated on the back surface of the sample substrate. When the pitch of the lattice marks is P, 0.9 (dλ / P) <
11. The pattern forming apparatus according to claim 10, wherein the value is selected within a range of L <1.1 (dλ / P). 14. The distance L between the two irradiation points is L = dλ /
14. The pattern forming apparatus according to claim 13, wherein the value is selected so as to substantially satisfy the relationship P. 15. The pattern forming apparatus according to claim 10, further comprising means for finely adjusting the distance L between the two irradiation points on the back surface of the sample substrate. 16. The spot diameter γ of the irradiation light beam at the two irradiation points on the back surface of the sample substrate is γ = nP (where n is a natural number), where P is the pitch of the lattice marks.
11. The pattern forming apparatus according to claim 10, wherein the value is selected so as to substantially satisfy the following relationship. 17. The pitch P of the grid-like marks is such that the thickness of the sample substrate is d, the central wavelength of the two-frequency light beam irradiated onto the back surface of the sample substrate is λ, and the pitch P is 2 on the back surface of the sample substrate.
When the distance between irradiation points is L, 0.9 (dλ / L) <P
The pattern forming apparatus according to claim 10, wherein the value is selected within a range of <1.1 (dλ / L). 18. The pitch P of the lattice marks is P = d
18. The pattern forming apparatus according to claim 17, wherein the value is selected to substantially satisfy the relationship of λ / L. 19. The pattern forming apparatus according to claim 10, further comprising means for finely adjusting the distance L between the two irradiation points on the back surface of the sample substrate. 20. The pattern forming apparatus according to claim 10, further comprising means for measuring the thickness d of the sample substrate. 21. The mark position detection means detects at least four mark positions provided on the back surface of the sample substrate, and the sample substrate is detected based on the mark position detection data at the four positions. 8. The pattern forming apparatus according to claim 7, further comprising calculation means for calculating and calculating the local deformation amount of. 22. A sample table for holding the sample substrate is further provided, and a sample substrate holding surface of the sample table has an optical path of a light beam to be irradiated on the back surface of the sample substrate. The pattern forming apparatus according to claim 10, wherein a plurality of openings to be formed are arranged at a predetermined pitch. 23. The arrangement pitch of the plurality of openings and the arrangement pitch of the position detection marks provided on the back surface of the sample substrate are set to be substantially equal to each other. The pattern forming apparatus according to.