JPH068926B2 - Surface position detection method - Google Patents

Description

The present invention relates to a surface position detecting method for detecting an image plane position of a projection optical system with respect to an object, and particularly to an image plane position of the projection optical system with respect to a wafer of a semiconductor printing apparatus. The present invention relates to a surface position detection method for detecting a.

[Prior Art] In recent years, semiconductor integrated circuits tend to be miniaturized more and more. And as the most important step of the manufacturing process,
There is a photo process of transferring a mask pattern onto a wafer. The apparatus used in this photo process is called a semiconductor printing apparatus, and the most influential method at present is the optical projection exposure method. For this purpose, a contact method is used in which a mask and a wafer are brought into contact with each other, a mask and a wafer are separated by a gap of several tens of μm, a proximity exposure method is used in which a mask pattern is shadow-printed on the wafer, and a high-resolution lens is used to move the wafer once There are a step-and-repeat method in which a mask pattern is transferred onto a wafer while feeding each by one by one, and a mirror projection method in which a mask pattern is imaged on a wafer using a high-resolution mirror optical system and printed.

Of these methods, the mirror projection method, in particular, guarantees the focus on the entire surface of the wafer because of the characteristic that a fine pattern in a line of about 1 μm is transferred at one time over the entire surface of a large diameter wafer having a diameter of 5 inches or more. However, there is a problem that unevenness of the order of several microns exists on the wafer surface. Therefore, there is a demand for an apparatus capable of measuring the state of unevenness at several points on the wafer and performing an average focus adjustment on the entire surface of the wafer.

2. Description of the Related Art Conventionally, as a semiconductor printing apparatus, a distance from a reference to a wafer is measured at a place different from the wafer printing position, and the wafer is positioned at the wafer printing position based on the measured distance. That is, since a projection optical system is not provided for focus detection, a method called an off-axis method is widely used. The measurement method at that time is to inject an air flow onto the wafer, measure the back pressure of the air to detect the distance to the wafer, measure the electric capacity between the wafer and the detection device, and use ultrasonic waves. There are a method of measuring a distance by using the method, a method of projecting a light beam obliquely on the wafer surface and detecting a positional deviation of reflected light.

However, in the case of the off-axis method, the position of the wafer in the optical axis direction can be managed in principle, but it is detected whether the wafer surface is located at the position where the mask image is truly in focus. It is not possible. The reason is that if the focus position of the projection optical system moves due to heat accumulation of illumination light during wafer printing or fluctuations in the ambient environment temperature, the defocused image will actually appear even if the wafer is set at an appropriate position. Is formed.

Therefore, in order to handle such changes, at the printing position,
Detection that looks into the wafer surface through the projection optical system is required.

[Problems to be Solved by the Invention] An object of the present invention is to perform focus detection through an objective optical system, and to enable detection with a relatively simple configuration. A subordinate object is to provide a surface position detecting method for focusing an image of a semiconductor manufacturing mask on a wafer by utilizing an aberration characteristic of a reflective projection optical system.

[Means for Solving the Problems] In order to achieve the above object, the surface position detecting method according to the present invention comprises:
An image plane position of a detected image height in the projection optical system, which is a method of detecting an image plane position of the projection optical system of a projection device that projects a pattern on the first object plane onto a second object plane via the projection optical system. Based on the detection step of optically detecting a signal corresponding to the distance between the object surface and the second object surface via the projection optical system, and the signal obtained by the detection step and the image plane curvature characteristic of the projection optical system. , Determining an image plane position of an image height different from the detected image height with respect to the second object plane.

In particular, the projection apparatus is a projection exposure apparatus that moves a first object and a second object in a one-dimensional direction to perform scanning exposure.

Further, the detected image height and an image height different from the detected image height are located side by side in the scanning direction.

[Embodiment] Before describing an embodiment of the present invention, a basic structure of a reflection type projection printing apparatus to which the embodiment is applied will be described first. In FIG. 1, PS is a reflection type projection system, and the basic type is composed of concavo-convex spherical surfaces that are arranged to face each other.

The catoptric projection optical system is a system that is axisymmetric with respect to the optical axis OO ', and
The aberration is corrected at y ₀ , and the radius of this good image area
An annular zone of y ₀ is used. As one method for facilitating scanning for transfer, the optical paths are bent by plane mirrors FM ₁ and FM ₂ , respectively, and the light flux emitted from one point of the mask Ma is reflected by the plane mirror FM ₁ and then the concave mirror M ₁ It is reflected by the convex mirror M ₂ and the concave mirror M ₁ in action again, the optical path is bent by the plane mirror FM ₂ , and converges to a point on the wafer W.

Since the ring-shaped region satisfies the required high performance, the arc-shaped portion included in this region is used for projection. FIG. 2 shows the aperture of width SW projected on the wafer W.
AP is drawn, and a partial image of the mask defined by the aperture AP is formed on the wafer in this way. Therefore, in actuality, the mask Ma and the wafer W are integrally scanned to scan the mask in the direction of the white arrow. The entire image is transferred onto the wafer. The mask Ma is illuminated by an illumination system IS corresponding to the aperture AP, and this illumination system is composed of an ultraviolet / far-ultraviolet light source, a condenser, a collimator, and the like. AS is an alignment optical system, which has a function of detecting the degree of alignment between the mask Ma and the wafer W, but details are omitted.

In general, the sagittal image plane S and the meridional image plane of the optical system have large deviations away from the optical axis. However, the catoptric projection optical system adopted here is different from the sagittal image plane at the off-axis specific image height (y = y ₀ ). The meridional image plane is designed to match. This aberration correction method is described in JP-A-52-5544.

In the vicinity of the image height (y = y ₀ ), the sagittal and meridional image planes are formed by intersecting inclined straight lines as shown in FIG. Z is the optical axis direction. In fact, when printing is performed, a band SW including the best image height, about 1 mm above and below, or at most a few times as wide as the band SW is used. However, since the optical system is rotationally symmetrical as described above, it becomes an arc.

Based on the above, the embodiments of the present invention will be described below.
It is assumed that a meridional image plane is used as the image plane curve of the optical system.

Now, assuming that the wafer W is arranged so as to coincide with the surface on which the mask image is formed in the apparatus of FIG. 6, the image height is vertically swung by an equal amount Δy with respect to the best image height y ₀ . Patterns P _1W and P _2W are provided on the wafer W that hits the heights of y ₁ and y ₂ . Since this optical system PS is a unity magnification system, the images of these patterns are mask Ma
Also on the same scale, images are formed at points P ₁ and P ₂ at image heights y ₁ and y ₂ . This is shown in FIG.

FIG. 4 shows an example of the projected pattern, but the width a is increased in the y direction so that the blurring of the image due to the meridional image plane is likely to appear.
A rectangular pattern having a length b (b> a) in the x direction (about 1 μm) is used.

On the other hand, on the light receiving side, slits of the same shape are provided at the respective image forming positions P ₁ and P ₂ of the mask Ma, and the light beam passing through these slits is received by the light receiving element. In practice, an aperture A is provided at a position optically conjugate with the mask Ma with respect to the microscope objective L ₀ , the optical path bending mirror M ₀ , and the relay lens Lr as in the embodiment shown in FIG. A light receiving element D capable of independently detecting the output of the region is installed and the output is guided to the comparator C. In the figure, the system below the microscope objective L ₀ is shown enlarged, but
Actually, the size is set so as not to block the illumination light from the illumination system (not shown). Of course, independent light receiving elements may be arranged. The degree of focusing and the direction of defocus can be found by taking the difference in the output from the light receiving area of the light receiving element D. This will be described with reference to FIG.

Pattern when using the meridional image plane (Fig. 4)
Are placed at image heights y ₁ and y ₂ , and the best focus positions at these image heights are indicated by A and B. The y-coordinate coincides with the ideal image plane, and the image is blurred regardless of whether the Z-coordinate moves in the (+) direction or the (-) direction.

Since the light receiving element is measured at the positions of P ₁ and P ₂ , a pattern image blurred by the defocus amount (ΔZ) in the optical axis direction is measured. Since a slit having the same shape as the pattern is arranged in front of the light receiving element, the light amount indicated by the hatched portion in FIG. 5 (A) is received. In this case, the difference between the outputs of both light receiving regions is 0 because the defocus amounts are equal (P ₁ A = P ₂ B).

Next, consider the case where the wafer is out of focus. FIG. 6 also depicts a state where defocus occurs. In this case, since the meridional image plane shifts from M to M ′ in the focus planes of the points P _1W and P _2W provided on the wafer at the image heights y ₁ and y ₂ , (the sagittal image plane is S → S ′), P ₁ and the defocus amount of the output in this case of the light receiving element placed in the P ₂ (respectively △ Z _1, △ Z ₂₎ changes in accordance with the. Therefore, the output at the point P ₁ decreases because the defocus amount increases (ΔZ ₁ > ΔZ), as shown in FIG. 5 (B), and conversely, the output at the point P ₂ decreases the defocus amount. Decrease (△ Z ₂
<ΔZ) Increase. Therefore, if the difference between the two outputs is calculated, the degree and direction of defocus can be detected.

FIG. 7 is a diagram showing the relationship between the defocus amount and the amount of light received by the light receiving element that receives the light flux that has passed through the slit, and the output when the defocus amount is 0 is normalized to 1.0. However, the numerical values in the figure are values when the optical system Fe = 3.5, λ = 633 nm, and the width of the projection pattern and the light-receiving slit is 2.2 μm. The reference numerals in FIG. 7 correspond to the reference numerals in FIG. 3, and in order to increase the sensitivity, two light receiving elements may be arranged at the positions A and B of the inflection points of the bell-shaped curve. At that position (image height), the tilt of the image plane (tan θ) is known, so Easily requested from.

FIG. 8 is a diagram showing the relationship between the output difference between a pair of light receiving elements and the defocus of the wafer, which is the defocus detection sensitivity curve of this system.

The above example is an apparatus in which a detection pattern is provided in advance on a wafer, but a pattern projection system may be provided on the apparatus side, and FIG. 9 shows an embodiment in that case.

That is, inside the detection optical system DS, the luminous flux emitted from the lamp LA is condensed at the position of the mask AFM by the condenser lens Lc. The mask AFM is provided with a detection-like mark, which is illuminated and the light flux advances. The light beam is focused on the mask lower surfaces P ₁ and P ₂ by the beam splitter BS ₀ , the relay lens LR, the mirror M ₀ , and the microscope objective L ₀ , and enters the projection system PS.

When this method is adopted, the sensitivity of detecting the out-of-focus is doubled. Because the detection pattern AFM provided in the detection optical system
Image on the mask, and then forms an image on the wafer through the projection system. At this time, if the wafer is defocused, the wafer acts as a mirror surface and the actual blur amount of the wafer A virtual image of the pattern is formed at a position 2d that is twice as blurred as d. Therefore, the image forming point of the return light is also shifted by twice at the detection position D.

In the above-described embodiment, the defocus amount is obtained from the output imbalance according to the sensitivity curve of FIG. 8, but the zero method can be performed. The method will be described with reference to FIG. 10, and the reference numerals in the figure are the same as those described above, for example, in FIG.

First, the light receiving elements are placed on P ₁ and P ₂ . When the wafer is defocused from the best focus, the focus planes of the sagittal and meridional lines move from S to S'and M to M ', respectively, so that an output difference occurs between both light receiving elements (ΔZ ₁ and ΔZ ₂₎ . Is supported). Now, the pair of light receiving elements are moved from the positions of P ₁ and P _{2 to} the positions of P ₁ ′ and P ₂ ′, respectively. However, the image height difference between the light receiving elements is kept constant at 2Δy. In P ₁ ′ and P ₂ ′, since the defocus amounts of both light receiving elements and the best focus surface of the meridional image plane M ′ are equal (= ΔZ), P ₁ ′ = A ″,
P ₂ ′ = B ″ and the differential output is 0. The wafer defocus amount (δ = ΔZ ₁ −ΔZ = ΔZ−ΔZ ₂ ) is the amount of movement of the light receiving element in the image height direction (Δ). If S) is known, δ = ΔS × tan θ (1) (where tan θ is the tilt angle of the image plane). It can be easily calculated according to this equation (1).

In order to realize this, in the embodiment shown in FIG. 6, the aperture A and the light receiving element D are connected to the optical axis X of the relay lens.
It may move in the vertical direction in a plane perpendicular to, or if only the aperture A moves if the light receiving area of the light receiving element D has a margin.
The drive mechanism AT for movement may be a combination of a screw feed mechanism and a reduction mechanism, or a voltage may be applied to the laminated body of piezoelectric elements to utilize its expansion and contraction to move the aperture coupled thereto. Also in the embodiment of FIG. 9, similarly, the aperture A and the light receiving element D may be moved until the output difference becomes zero.

The above zero method has some advantages, and the first one is the high precision which is a characteristic of the zero method itself. Secondly, if only the tilt angle of the image plane is known, the defocus amount of the wafer can be easily back-calculated from the moving amount of the light receiving element.
There is no need to remember the sensitivity curve. Therefore, it is possible to perform very stable detection with respect to fluctuations in the device and installation errors in the detection optical system. The third advantage is that the position setting accuracy of the aperture and the light receiving element may be loose. This is because the sagittal and meridional image plane characteristics are outstanding, and therefore the light receiving element must be largely moved in the image height direction to detect a minute defocus amount. By the way, in the case of a mirror projection optical system mainly composed of two concave and convex mirrors, it is necessary to move the light receiving element by 100 μm to detect a defocus of 1 μm. It is easy to achieve this accuracy.

In the above example, the example in which the light receiving element is mechanically moved to realize the zero method has been described above, but a method of substituting it with a devised output of a photoelectric position detection element such as a CCD may be considered. That is, as shown in FIG. 12, photoelectric position detecting elements Q ₁ and Q ₂ are provided which have their centers at image heights y ₁ and y ₂ , respectively. On the pair of photoelectric elements, an electric mask is applied to first fix the integration regions R ₁ and R ₂ (the rectangular meridional pattern) of the received light amount.

Next, at the time of defocusing the wafer, according to the above-described principle, the light amount integration regions on both the light receiving elements are moved while maintaining the shape and the image height difference so that the output difference between the both light receiving elements becomes zero. Go. As a result, R ₁ → R ₁ ′ and R ₂ → R ₂ ′, respectively. When the amount of movement (ΔS) is calculated electrically, the above (1)
From the formula, the defocus amount of the wafer can be known.

The following example concerns deformation of the sensing aperture. This is because the aperture A used in FIGS. 6 and 9 has an opening (slit) having the same shape as the pattern in FIG. 4, but the pattern image is shielded by using the mask having the same shape. Characterized by points. By doing so, the light flux in the peripheral portion shown by the diagonal lines in FIGS. 11 (A) and (B) is received, and the light amount of the portion that is not significantly changed is shielded, so that improvement in detection sensitivity can be expected. . However, in this example, the detection curve is inverted as compared with FIG. 7, and becomes as shown in FIG.

FIG. 13 (A) shows an example of arrangement when detecting at multiple positions.
In the example described above, the detection is performed at one location on the subject, whereas in this example, P ₁ P ₂ , P ₁ ′ P ₂ ′, P ₁ ″ P ₂ ″.
Are measured at three points. That is, the arc indicated by AP is a position where the meridional image plane intersects with the ideal image plane (because the optical system has rotational symmetry, it becomes an arc), and a pattern is provided or projected at this position by sandwiching this position. P ₁ output of P ₁ 'P _2' from P _2, the output of P ₁ "P _2" is the defocus amount is averaged is calculated, based on the wafer W in FIG. 6 and FIG. 9 to the calculated value An adjusting mechanism (not shown) moves vertically to achieve focus. The mask may be moved instead of the wafer, or the optical system may be moved if possible.

The following example is a method (FIG. 14) in which the detection position is set to a low image height (or a high image height) for both points P ₁ and P ₂ from the aberration-corrected best image height (y = y ₀ ). is there. Image plane tilt angle (ta
If nθ) is known, it is operationally the same as the case of arranging P ₁ P ₂ across the best image height. This is achieved by moving the pair of the aperture A and the light receiving element D upward or downward in correspondence with the meridional image plane and by arranging them forward or backward in the configuration of FIG. 6 or FIG. This configuration can also be used in the arrangement shown in FIG. 13 (B). P ₁ P ₂ , P ₁ ′ P ₂ ′ and P ₁ ″ P ₂ ″ are the positions where the pattern is projected.

According to this method, when the slit moves on the wafer, the focus shift can be detected immediately before each point on the wafer is burnt. If the wafer scanning speed and the difference between the image height to be detected and the image height to be detected (Δyd) are known, it is easy to correct the focus shift when each point on the wafer reaches the baking position. . Further, by providing a plurality of pairs of detection points, the average value of the focus deviation at each detection point can be obtained. In this way, there is an advantage that during the scanning of the wafer, the focus can always be continuously and continuously focused on the entire baked portion on the arc.

EFFECT OF THE INVENTION A signal according to the distance between the subject and the image plane of the projection optical system, and
From the image plane curvature characteristics of the projection optical system, it is possible to detect the image plane position with respect to the subject, especially with respect to the usage area, even if the means for directly detecting the image plane position cannot be structurally arranged. There is.

Further, as described in the embodiments, since the reflective system constituted by only the mirror has no chromatic aberration, there is no difficulty in restricting the design of the projection optical system due to the wavelength of the focus detection light.

[Brief description of drawings]

FIG. 1 is an optical sectional view showing a conventional example, FIG. 2 is a plan view of a wafer position, FIG. 3 is a view showing a main part of astigmatism, FIG. 4 is a plan view of a detection pattern, and FIG. FIG. 6 is a diagram for explaining the amount of light received by the light receiving element, FIG. 6 is an optical sectional view showing an embodiment of the present invention, FIG. 7 is a relational diagram of the amount of received light and defocus, and FIG. 8 is a differential output. FIG. 9 is a cross-sectional view showing another embodiment of the defocusing, FIG. 10 is a view showing the main part of astigmatism, and FIG. 11 is a view for explaining the amount of light received by the light receiving element. FIG. 12 is an explanatory view of an example using an optical position detecting element, FIG. 13 is a view showing an arrangement example of detection positions, FIG. 14 is a view showing a main part of astigmatism, and FIG. The related figure of the amount of defocus. In the figure, Ma …… Mask W …… Wafer PS …… Reflective projection system M ₁ …… Concave mirror M ₂ …… Convex mirror M …… Meridional image plane S …… Sagittal image plane L ₀ …… Microscope objective lens Lr …… Relay lens A ... Aperture D ... Light receiving element AFM ... Mask with detection pattern LA ... Illumination light source.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI technical display location 7352-4M H01L 21/30 311 N

Claims (3)

[Claims] 1. A method for detecting an image plane position of a shadow optical system with respect to a second object plane of a projection device, which projects a pattern on a first object plane onto a second object plane via a projection optical system, wherein: A detection step of optically detecting, via the projection optical system, a signal corresponding to the distance between the image plane position of the detected image height in the optical system and the second object surface; the signal obtained by the detection step and the projection optics Determining the image plane position of the image height different from the detected image height with respect to the second object plane, based on the image plane curvature characteristic of the system. 2. The surface position detecting method according to claim 1, wherein the projection device is a projection exposure device that moves a first object and a second object in a one-dimensional direction to perform scanning exposure. . 3. The surface position detecting method according to claim 2, wherein the detected image height and an image height different from the detected image height are positioned side by side in the scanning direction.