JPH04115518A - Exposure device - Google Patents

Abstract

PURPOSE:To enable an error at an exposure position and an off-axis microscope position to be detected and compensated without using other media and alignment accuracy to be improved by forming a latent image at a blank part of a substrate to be exposed. CONSTITUTION:Marks for correction 105L and 105R are placed marginally at a transferable region 101 on a reticle 1. The marks for correction 105L and 105R are projected and exposed at a blank part where no actual element pattern is created on a wafer 2 for forming a latent image and an AGA (Advanced Global Alignment) is executed to the formed mark for correction. The corrected array coordinates when exposing a wafer are determined from an alignment detection data and a state position data by an off-axis microscope 7 which is obtained at each shot of AGA, thus enabling aging error constituents of off-axis alignment system to be corrected without using other media and enabling alignment accuracy to be improved.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention relates to an exposure apparatus used in the manufacture of semiconductor devices such as ICs and LSIs, and specifically relates to an exposure apparatus used in the manufacture of semiconductor devices such as ICs and LSIs, and specifically, to an exposure apparatus used in the production of semiconductor devices such as ICs and LSIs. This relates to improvement of the superimposition function (commonly known as alignment).

[Background Art] There is no end to the miniaturization and higher integration of semiconductor devices (elements), and progress continues at a rapid pace. DRAM (dynamic random access memory), which plays a role as a driving force in miniaturization, has already entered the submicron range at the product level, and at the research level, it is possible to achieve accuracy of half a micron (0.5 μm) or less. Patterning is discussed.

An exposure device called a stepper, which appeared in the 256-bit DRAM era, is the main model in the production of 1M to 4M bit DRAMs, and it is predicted that it will continue to occupy this position in future ultra-fine devices.

When it comes to miniaturization, resolution is always a topic of discussion, but on the other hand, overlay accuracy is equally or more important than resolution. The required accuracy for overlay is 1/3 to 115 degrees of resolution.

Overlay accuracy can be broadly divided into the following two elements.

■Magnification, distortion component ■Alignment component The subject matter of the present invention is the alignment component (1).

The stepper, which first appeared with an off-axis alignment system, has since been proposed for many improvements, and is currently known as T.
TL alignment systems have become mainstream. The stepper alignment systems proposed so far can be broadly classified into the following three types.

■TTLONAXIS system: The alignment light is the same as the exposure light, and the feature is that the reticle and wafer can be observed at the same time.

■TTL N0NAxIS system: The alignment light is different from the exposure light, but the alignment light passes through the projection lens. Simultaneous observation of the reticle and wafer is difficult.

■Ofaxis system, an alignment microscope is placed completely separate from the projection lens.

Among these, off-axis steppers have many indirect error factors that intervene in the relative positioning of the reticle and wafer, and because the time and travel distance from alignment to exposure are long, the error components change greatly over time, resulting in High overlay accuracy cannot be obtained. Therefore, it is a method that is not used very often at present.

On the other hand, for resolution, the exposure wavelength is set to g-line (wavelength 4
The numerical aperture (NA) of the projection lens remains fixed at 36 nm).
It has been attempted to improve the resolving power Re by increasing .

However, as is clear from Rayleigh's equation %, the depth of focus DOF decreases as the NA increases, and on the other hand, the design and manufacture of the projection lens have reached their limits. Therefore, in order to support the future submicron generation, it is becoming necessary to shorten the exposure wavelength.

Currently, the i-line (wavelength 365 nm) stepper has entered the stage of practical use, and the next generation will include KrF excimer laser (
An excimer stepper using a light source with a wavelength of 248 nm is considered promising.

However, in order to put the excimer stepper into practical use, it has become necessary to review the aforementioned alignment system. This is because KrF excimer laser (wavelength 248n)
This is because the glass materials that allow the light (m) to pass through are limited to quartz and fluorite, and it is extremely difficult to correct chromatic aberration for light other than the exposure wavelength in terms of design.

FIG. 5(a) shows the longitudinal chromatic aberration characteristics of a conventional g-line lens. FIG. 6(b) shows the longitudinal chromatic aberration characteristics of a conventional single quartz lens for excimer laser. In both graphs, the horizontal axis shows wavelength and the vertical axis shows longitudinal chromatic aberration.

In the case of the g-line lens shown in FIG. 5(a), it is usually possible to design a combination of glass materials so that the characteristic curves touch at the zero point at the target wavelength. On the other hand, Figure 5 (b
In the excimer lens of ), since there is no degree of freedom in the glass material, the line becomes an almost straight line that intersects at one point of the target wavelength.

For example, HeNe is used as alignment light for the g-line lens.
When a laser (wavelength: 633 nm) is selected, the axial chromatic aberration is approximately 10-odd μm. On the other hand, if, for example, an Ar laser (wavelength 500 nm) is selected as the non-exposure alignment light for the excimer lens, the axial chromatic aberration will reach the order of mm. From this reality,
Unless the axial chromatic aberration characteristic curve can be improved, a non-exposure alignment system in excimer lenses (steppers) is difficult to implement.

Based on this knowledge, as improvements to the above-mentioned drawbacks of off-axis alignment systems, we have developed a method using a dummy wafer (Japanese Patent Application Laid-Open No. 1-120820) and a method using a chuck loaded with recording material (Japanese Patent Application Laid-Open No. 1-120820).
No. 286309).

[Problems to be Solved by the Invention] However, the method using a dummy wafer requires a dummy wafer in addition to the normal wafer on which the actual device pattern is printed. Therefore, there are the following problems.

1) It takes time and effort to create dummy wafers.

2) The number of normal wafers is reduced by the number of dummy wafers, or a place other than the normal wafer location is required to collect the dummy wafers.

3) After measuring with a dummy wafer, it is necessary to collect the dummy wafer and carry in a normal wafer, which causes a decrease in throughput and changes over time in the baseline length, etc.

In addition, in the method using a chuck loaded with recording material,
Since the recording material is mounted inside the apparatus and placed outside the wafer outer diameter, there are the following problems.

1) A procedure and means are required to erase the image formed on the recording material.

2) Normally, since the recording material is placed at a position away from the position on the wafer where the actual element pattern is printed, differences in measured values such as baseline length between the wafer and the recording material occur due to stage yawing, pitching, etc. .

In view of the above-mentioned problems in the conventional example, an object of the present invention is to significantly improve alignment accuracy in an exposure apparatus of an off-axis alignment system without using a chuck on which a dummy wafer or recording material is mounted.

[Means and effects for solving the problems] In order to achieve the above object, the present invention provides an exposure method in which a pattern on an original plate (reticle) is projected onto a photosensitive layer on a substrate (wafer) to be exposed through a projection optical system. In the apparatus, in order to form an image of a specific mark on the original plate as a latent image on a photosensitive layer at a predetermined position of the exposed substrate through the stage means for moving the exposed substrate and the projection optical system, illumination means for illuminating only a specific mark; detection means for detecting an image of the specific mark formed as a latent image on the photosensitive layer;
The apparatus is characterized by comprising a control means for controlling movement of the stage means based on the detection output of the detection means.

Currently, photosensitive materials containing photochromic materials have been developed. According to this, a latent image (an image that detects changes in the refractive index and transmittance of a photosensitive layer due to light irradiation under a microscope using white light (for example, JP-A-61-114529, etc.))
can be easily formed. Furthermore, there are some blank areas on the substrate to be exposed, where actual device patterns are not printed.

With the above configuration, the present invention forms a latent image in the blank area of the substrate to be exposed, and can detect and correct errors between the exposure position and the off-axis microscope position without using other media such as a dummy wafer or recording material.

This improves the alignment accuracy of the ophaxis system.

[Example] Hereinafter, the present invention will be explained in detail using the drawings.

FIG. 1 shows the main components and their overall arrangement of a stepper, which is an exposure apparatus according to an embodiment of the present invention.

In the figure, reference numeral 1 indicates a photo original plate (hereinafter referred to as a "reticle"), and 2 indicates a semiconductor substrate (hereinafter referred to as a "wafer") as a substrate to be exposed.

A light beam 31 emitted from the excimer laser 30 passes through the illumination optical system 3 and illuminates the reticle 1. This light beam reaches the wafer 2 via the projection lens 4, whereby the pattern on the reticle 1 can be transferred to the photosensitive layer on the wafer 2. 31 is the CPU 9 that drives the excimer laser 30.
This is an interface for controlling from 1.

The excimer illumination system will be briefly explained.

The beam directed upward by the mirror 41 passes through the incoherent optical system 32, the fly's eye lens 33, the condenser lenses 34a and 34b, and the mirror 35 before reaching the masking image plane. 36 is a masking blade;
7a is a first masking imaging lens; 38 is a mirror 3
Fb is a second masking imaging lens.

The reticle 1 is supported by a reticle holder 11. The reticle holding base 11 is supported by a moving stage (not shown).

The wafer 2 is exposed to light while being vacuum-adsorbed by the wafer chuck 21. FIG. 1 shows a state in which the wafer 2 is supported by the second support member 22. As shown in FIG. The wafer chuck 21 is movable in each axis direction by the stage 5.

The stage 5 is supported by a stage base 50. The stage 5 is a Y stage 51, an X stage above the Y stage 51,
It has a stage 52, a fine movement stage 54 for leveling and Z direction movement above the X stage 52, a rotational (θ) fine movement stage 55 above the fine movement stage 54, and a vertical (Z) fine movement stage 56. The fine movement stage 54 for leveling and Z direction movement includes three piezoelectric elements (piezo elements) 53.
It is designed to move up and down.

The wafer chuck 21 is placed on the Z fine movement stage 56. Above the leveling and Z fine movement stage 54,
The mirror 85 that serves as the reference for the position coordinates of the stage system is
They are placed in each direction. The mirror 85 reflects the laser beam emitted from the laser interferometer 86. This allows you to know the position of the stage and the distance traveled. 87 is a receiver that converts an optical signal into an electrical signal.

On the other hand, on the leveling and Z fine movement stage 54, θ
A coarse movement mechanism 57 and a Z coarse movement mechanism 58 are configured. Furthermore, the second support member 22 is placed thereon.

A reticle optical system 6 is arranged above the reticle 1. The reticle optical system 6 is an optical system that handles light of the same wavelength as the exposure light. Therefore, by releasing the switching mirror 41, the light beam 31 emitted from the excimer laser 30 is transferred to the reticle optical system 6 via the mirrors 42 and 43.
It's like supplying it to someone.

The reticle optical system 6 is a binocular optical system having two objective lens systems. A target mark formed in a small area on the reticle 1 is irradiated by the reticle optical system 6. This makes it possible to transfer the target mark to the wafer 2 as a latent image. An incoherent optical system 40 is disposed between mirrors 42 and 43 so that the illumination by the reticle optical system 6 has the same illumination σ as the illumination system 3 and to reduce uneven illuminance. Note that illumination σ refers to the ratio of the numerical aperture NA of the illumination system to the numerical aperture NA of the projection lens (σ=NA, /NA),
By changing 200, the resolution and depth of focus of the exposure change.

The laser beam entering the reticle optical system 6 passes through a beam splitter 44, a relay lens 45, and an objective lens 46, and is bent downward by a total reflection prism 47 to illuminate a target mark placed at a predetermined position on the reticle 1. . Thereby, the image of the target mark is transferred to the wafer 2 via the projection lens 4. 48 is a CCD for observing the target mark on the reticle 1. C
The output of the CD48 is sent to the CP via the interface 61.
Input to U91.

An off-axis microscope 7 is arranged above the stage 5 and adjacent to the projection lens 4. The off-axis microscope 7 is a monocular microscope that handles non-exposure light (white light).

Its main role is to detect the relative position between the internal reference mark 70 and the alignment mark on the wafer. The off-axis microscope 7 has the following configuration.

The objective lens 71 and the relay lens 72 enlarge and project the wafer pattern onto an imaging plane 74 .

The erector lenses 77 and 78 function as low-magnification erector lenses when both are inserted on the optical axis, and as high-magnification erector lenses when the erector lens 78 is withdrawn, so that the aerial image on the imaging plane 74 is received by the CCD 79. Project onto a surface.

25 is an optical fiber that guides light from a light source (not shown). The light emitted from the optical fiber 25 becomes illumination light that illuminates the wafer 2 via the illumination lens 26 and beam splitter 73. Similarly, 27 is an optical fiber that guides light from a light source (not shown), and the light emitted from this optical fiber 27 illuminates the reference mark 70 via the illumination lens 28.

The beam splitter 75 is arranged so that the pattern surface of the reference mark 70 and the imaging surface 74 have the same optical path length with respect to the light receiving surface of the CCD 79. Therefore, the reference mark 70 is also projected and imaged onto the light receiving surface of the CCD 79 by the erectors 77 and 78.

Note that it is desirable that the off-axis microscope 7 is equipped with a focus detection function. This is possible, for example, by performing signal processing for detecting the amount of blur on the output signal of the C0D79.

A control circuit 9 is used to control each of the above-mentioned components. The CPU 91 issues commands to each element according to predetermined sequence software, and also judges data from each element to determine the next procedure. The arithmetic circuit 92 is
It is mainly used for arithmetic processing that requires high speed and high precision, such as calculating the relative position between the reticle 1 and the wafer 2 from stage coordinates and detection results of an off-axis microscope. The storage circuit 93 is used to store the measurement data and calculation data. A detailed partial structure will be described later.

FIG. 2 is a plan view of the reticle 1 used in the exposure apparatus of this embodiment. In the figure, 101 indicates a transferable area of the projection lens 4, and 102 indicates an actual element pattern area. Precision wafer alignment marks 103 are placed on the left and right sides of the scribe line area in contact with the actual device pattern area 102.
L, 103R are arranged. Further, a rough alignment mark 104 is arranged above the scribe line area. Further, correction marks 105L and 105R are arranged at the very edges of the transferable areas on the left and right sides of the X-axis. Precision wafer alignment marks 103L and 103R are for use in the next process.

FIG. 3(a) is a plan view of a wafer and a wafer chuck in the exposure apparatus of this embodiment.

In the figure, patterns created in the previous process are arranged on a wafer 2. Marks corresponding to the alignment marks 103L, 103R, and 104 described in FIG. 2 are formed in portions corresponding to the scribe lines between the respective shot areas. Among these, the pattern pre-alignment marks are appropriately selected, such as using the coarse alignment mark 110 belonging to the 34th shot and the coarse alignment mark 111 belonging to the 42nd shot. For precision alignment marks, please contact AGA (Advanced Global Al
ignment) is the basic, so for example, the AGA shot shown by hatching [7, 13, 19, 35° 38.41
, 57, 63.69].

As shown in the figure, there is a blank area around the wafer 2 in which no actual device pattern is formed. For example, figure numbering 113°114.115,116
This is the part indicated by . The correction mark 105L described in FIG. 2 is provided in this portion.

By projecting and exposing 105R, for example, mark 12
2 is formed as a latent image.

FIG. 3(b) shows a state in which the precision alignment mark 117 and the correction mark 122 are observed in the field of view of the off-axis microscope 7. In this example, the above 2
The reticle is designed so that each type of mark has the same shape. 13o is a frame screen of the CCD 79, and 131 is a reference mark 7o built into the off-axis microscope 7.

Assuming that FIG. 3(b) shows the state in which the precision alignment mark 117 is observed, the mark 132 is the mark 103L on the reticle, and the mark 133 is the mark 103R on the reticle, which were created by projection exposure. Become something. In addition, FIG. 3(b) shows the correction mark 12.
2 represents the observed state, mark 132
is a mark 105L on the reticle, and mark 133 is a mark 105R on the reticle, which are created by projection exposure.

In this case, the precision alignment mark 117 and the correction mark 122 are both marks that are located at separate positions on the reticle 1 (mark 103L and mark 103 that is apart from this).
R, mark 105L and a separate mark 105R)
Adjacent marks are formed by moving the stage and exposing twice. This method has been proposed in Japanese Patent Laid-Open No. 13329/1983 by the present applicant.

Next, the operating procedure of the exposure apparatus of this embodiment will be explained. The procedure is shown in the flowchart of FIG. The operation of each part will be explained according to FIGS. 1 to 3.

A stepper continuously performs alignment exposure processing on a large number of wafers.

Referring to the flowchart in FIG. 4, when the exposure process for a predetermined reticle 1 starts in [Step 150], the wafer 2 is replaced in [Step 151], and the wafer 2 is placed on the second support member 22. is set. wafer 2
The replacement is performed with the second support member 22 shown in FIG. 1 protruding from the surface of the wafer chuck 210.

In [Step 152], left and right pattern pre-alignment marks 110 and 111 on the wafer 2 are detected. For this purpose, the stage is first moved so that, for example, the right pre-alignment mark 110 (FIG. 3) is included in the low magnification field of the off-axis microscope 7, focused, and the mark 1
10 positions are detected. Then, the stage position at that time is memorized. Note that focus detection is performed by the focus detection function of the off-axis microscope 7, and driving is performed by the two coarse movement mechanisms 59. Since the magnification is low, the accuracy is poor. mark 1
After position detection at 10, the stage moves to the pre-alignment mark 111 on the left, and focuses and performs the same steps as on the right.
Performs position detection and coordinate storage.

By this step 152, the optical axis of the off-axis microscope 7 (
In this case, data is obtained that roughly indicates the coordinate relationship of the pattern engraved on the wafer 2 with respect to the device coordinates with the reference mark 70) as the origin and the XY stage as the coordinate axis.

In [Step 153], the arithmetic circuit 92 calculates the coordinate error based on the data obtained above, and
The correction amounts ΔX1ΔY and Δθ are calculated.

In [Step 154], the XY stage adjusts the correction amounts ΔX and Δ
Taking Y into consideration, movement is started toward the next target position (the position where step 156 is performed). The θ coarse movement mechanism 58 is rotationally driven to make the θ correction amount Δθ zero.

In [Step 155], the second holding means 22 is lowered, and at the same time, two mechanisms including the wafer chuck 21 are raised to transfer the wafer 2.

In [Step 156], preliminary measurement of AGA is performed.

In this step, for example, the 13th. No. 19. No. 63,'s57
The precision alignment marks of the four shots are off-axis marked 1i! Read with 7. Mainly, the residual error in θ is read, and this is corrected in rotation using the θ fine movement stage.

In [Step 157], the correction marks 105L and 105R on the reticle 1 are exposed and transferred to a blank area where there is no actual element pattern. Whether or not to perform this process on all of the four blank areas 113, 114, 115, and 116 where there is no actual element pattern can be determined by checking only the baseline length, or by checking the stage magnification and orthogonality. It can be set arbitrarily based on the judgment whether to check or not. Also, whether to perform this process for each wafer or once for several wafers can be set arbitrarily depending on the stability of the baseline and stage magnification and orthogonality over time and the required accuracy. can.

In [Step 158], shot number 13, 19.35
, 38, 41, 57, 63. AGA is performed on the precision alignment marks of each shot of 69. At this time, it is possible to arbitrarily set whether to detect marks on only one side of the relevant shot or to detect marks on both sides depending on the required accuracy.

In [Step 159], AGA is performed on the correction mark formed in the blank area where there is no actual element pattern.

In [Step 160], the corrected array coordinates at the time of wafer sale are determined from the alignment detection data and stage position data by the off-axis microscope shot by the AGA in [Step 158] and [Step 159]. From the data obtained in [Step 159], the baseline length, the magnification component and orthogonality component of the stage, and the magnification component of the projection lens can be calculated. The magnification component of the lens is not the focus of this embodiment. [From the data obtained in step 158, the x of the shot array within the wafer is
@The Y-axis component, the magnification components in the X and Y directions, and the chip rotation direction can be calculated. From the above results, the array coordinates at the time of exposure are determined.

In addition, if you want to correct up to the die rotation, insert a step to correct the rotation of θ for the die rotation after step 160, and then perform stepwise exposure in the next step 161. do it.

In step 161, step exposure of the wafer is performed according to the array coordinates calculated in step 160 above.

[In step 162, the exposed wafer is transferred from the chuck to the second holding means, and the wafer is recovered.

In [Step 163], if the next unexposed wafer is waiting, the process returns to Step 151; if there is no waiting wafer, the sequence is completed.

[Modification of Example] Kanbori” mouth.

In the flowchart of FIG. 4, if step 157 precedes step 159, these two steps can be placed anywhere under the following conditions.

■Step 157 and step 159 are the same as step 155 and step 1.
You can put it anywhere between 60 and 60.

■Step 159 may be divided into several steps.

For example, another procedure may be inserted between the first and second steps of step 159.

Modified Example 2 In this example, the peripheral area of the wafer was used as an example of the empty space other than the area where the actual device pattern is printed, but if there is empty space between shots as well as the peripheral area, A latent image may be formed at that location (for example, in the margin of a shot portion for a test chip or on a scribe line) and detected.

Although the embodiment has been described using an off-axis type alignment system, the present invention is applicable and effective to all systems except a die-by-die on-axis alignment system.

1 Beard 1 As explained in the embodiments, the magnification of the projection lens can be measured in the process of implementing the present invention. In this case, the magnification reading accuracy includes a detection error of the microscope and a feeding error of the stage. Therefore, it is somewhat impossible to correct the magnification with one detection.

However, the factors that change the magnification of the lens are atmospheric pressure and environmental temperature, and it is impossible for them to change suddenly. Therefore, the lens magnification is calculated each time in step 160 of FIG.
Accuracy can be improved by averaging this multiple data, and it is appropriate to correct the lens magnification using this average value. It is a known technique that the lens magnification can be corrected by moving some lenses of the projection lens in the optical axis direction. Alternatively, in the case of excimer monochromatic quartz lenses, it has already been proposed that magnification can be corrected by shifting the exposure wavelength.

[Effects of the Invention] As explained above, by forming a latent image in the empty space other than the part where the actual device pattern is printed on the wafer, it is possible to eliminate the disadvantage of off-axis alignment systems over time without using any other medium. We were able to provide a concrete means for correcting the changing error component, and were able to significantly improve alignment accuracy.

Furthermore, the present invention also makes it possible to simultaneously detect the magnification error of the projection lens and feed back the detected error to the correction means, thereby making it possible to improve the precision of the magnification distortion component.

[Brief explanation of drawings]

FIG. 1 is a sectional view of the entire system of an exposure apparatus according to an embodiment of the present invention. FIG. 2 is a plan view of a reticle used in the exposure apparatus of the above embodiment. A plan view of a wafer, a wafer chuck, and a latent image forming part used in the exposure apparatus, FIG. 4 is a flowchart for explaining the operation of the exposure apparatus of the above embodiment, and FIG. 5 shows a g-line lens and an excimer lens. This is the characteristic curve of 1 reticle, 2: wafer, 3: illumination system, 4: projection lens, 5: stage, 6: reticle microscope, 7: off-axis microscope, 9: control circuit.

Claims (1)

[Claims] (1) In an exposure apparatus that projects and exposes a pattern on an original plate onto a photosensitive layer on a substrate to be exposed through a projection optical system, a stage means for moving the substrate to be exposed, and a specific mark on the original plate through the projection optical system. illuminating means for illuminating only a specific mark on the original plate in order to form an image of the image as a latent image on a photosensitive layer at a predetermined position of the exposed substrate; and the specific mark formed as a latent image on the photosensitive layer. An exposure apparatus comprising: a detection means for detecting an image of the stage; and a control means for controlling movement of the stage means based on a detection output of the detection means.