JPH0828319B2 - Projection exposure device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a projection exposure apparatus for semiconductor circuit patterns, display device patterns such as liquid crystal, and fine patterns such as
The present invention relates to a projection exposure apparatus equipped with means for detecting the inclination and height of an object to be exposed so that the entire exposure area can be exposed with high resolution.

[Conventional technology]

Exposure of fine pattern of semiconductor integrated circuit or TFT (T
hin Film Transistor) In the exposure of the drive circuit pattern in the large field pattern of the display device typified by a liquid crystal television, there is little variation in the line width over the entire exposure area.
It is necessary to expose a pattern that is true to the original picture. Particularly in the field of semiconductor integrated circuits, it is necessary to expose a line width pattern of 0.5 μm pattern or less over the entire area of about 15 mm in the future.
Image formation range (depth of focus) as patterns become finer
Is less than ± 1 μm. Therefore, it is essential that the photoresist surface on the wafer be exactly aligned with the pattern imaging surface. In order to realize this, it is necessary to accurately detect the inclination and height in the exposure area of the wafer surface (photoresist surface).

In the first known example disclosed in Japanese Patent Laid-Open No. 63-7626, the height is detected by focusing a semiconductor laser on the wafer surface from an oblique direction and detecting the focusing position. . Further, in this known example, multiple reflections due to the multi-layer structure of the wafer are dealt with by using a semiconductor laser of three wavelengths, the converging position is changed to the oblique incident direction and the direction at right angles, and the heights of different places on the wafer are obtained. ing. In this known example, the height is mainly detected, and it is possible to measure the tilt by changing the place in the direction perpendicular to the oblique incident direction, and the tilt can be detected, but two places in a narrow area of about 20 mm in diameter are measured. However, it is difficult to obtain an accurate value for the slope. In this known example, in order to realize height detection with high accuracy, it is necessary to sufficiently collect light on the wafer, that is, to make the light collecting diameter as small as possible. It is necessary to increase the converging angle (the angle of the outermost ray of the concentrated light flux with respect to the principal ray), and as a result, the incident angle of the principal ray must be reduced. If this angle is made smaller (the angle from the line perpendicular to the wafer surface becomes smaller), the influence of multiple interference due to the multilayer structure of the wafer becomes large for the reason described later. In this known example, three wavelengths are used for this problem, but interference is exerted on each wavelength, and this is not a fundamental excessive solution.

In the second known example disclosed in Japanese Patent Laid-Open No. 63-199420 as a conventional tilt detection method, tilt detection light different from the exposure wavelength is irradiated through a projection lens to collect reflected light and collect it. Although the tilt is detected from the position, since it is incident on the wafer at a substantially vertical or shallow angle, the influence of interference with the reflected light from the base cannot be ignored because of the reason described below, and accurate detection becomes difficult.

Further, as a conventional height detecting method for a multi-layered structure object, in the third known example disclosed in Japanese Patent Laid-Open No. 63-247741, the reflected light from the undercoat film is separated separately. It is difficult to perform for thin films that are registered in the process of semiconductor circuit fabrication.

[Problems to be Solved by the Invention]

The above-mentioned prior art does not consider the fact that the information on the inclination and height in the exposure area is accurately obtained with respect to a multilayer structure such as a wafer having a semiconductor circuit pattern.
There is a problem with the highly accurate tilt and height control required for circuit pattern exposure of 5 μm or less.

The object of the present invention is to solve the above conventional problems, accurately detect the inclination and height of the photoresist surface in the exposure region for any wafer in the semiconductor process, and always detect the resist surface or its vicinity on the image plane. It is an object of the present invention to provide a projection exposure apparatus that exposes a high-resolution pattern that is aligned with an optimum position and has a small line width variation.

[Means for solving the problem]

In order to achieve the above object, in the present invention, the light emitted from the coherent light source is converted into parallel illumination light, and the exposure area of the projection optical system on the photoresist surface on the wafer is obliquely irradiated with an incident angle θ. The reflected light and the reference light generated by separating the light emitted from the light source are incident on the pattern detector at a desired angle to each other to detect interference fringes. From this change in the interference fringe pitch and phase, it is possible to determine the change in the inclination and height of the photoresist on the wafer and the surface. Further, it is easy to set the incident angle to 85 ° or more in the present invention using a parallel light flux, and since the incident angle is large, the reflection on the surface of the photoresist becomes large, and the reflection on each layer of the underlying layer structure is reduced. As a result, the influence of the generated interference can be almost ignored. When the photoresist incident light is S-polarized, the reflection on the surface is further increased and the accuracy is improved.

Further, the light reflected on the photoresist surface is perpendicularly incident on the plane mirror, the reflected light is again incident on the photoresist surface, and if the reflected light is used as object light to obtain the information of the interference pattern, the tilt and height of the wafer can be obtained. Can be detected with twice the sensitivity, and detection can be performed with higher accuracy.

Further, by constructing the reference light so that it effectively travels in substantially the same direction as the irradiation light of the photoresist and the object light (reflected light) and passes through the same region, each optical path causes disturbances such as air fluctuations. In the same manner, it becomes possible to detect tilt and height that are less susceptible to changes in the surrounding environment.

Further, if the information of the obtained interference fringes is subjected to the fast Fourier transform and the inclination Δθ and the height θh are obtained from the information in the vicinity of the spectrum of the fringes as a result, the Δ is as fast as the real time.
θ and Δh are obtained. At this time, if the photoresist irradiation position is in an optically conjugate (image-forming) relationship with the array sensor light receiving surface that is the pattern detecting means, information on only a desired area on the wafer is selected, and the inclination and height of that portion are selected. It becomes possible to ask for it.

Also, when the above-mentioned interference measurement is performed with light of one wavelength, the height is obtained from the phase of the obtained interference fringes, but this phase change cannot be identified even if it changes from α to 2nπ + α (n: integer). , Using the second coherent light of different wavelength,
The same optical system (optical path) as the light of wavelength
If one wavelength is separated and the height is determined by using two pieces of interference fringe information, the height information can be accurately obtained over a wide range of height change. Also, by using other wafer height detection means such as an air micro, it is possible to eliminate the uncertainty factor in the height direction due to the detection of one wavelength and accurately obtain height information over a wide range of height change. It will be possible.

[Action]

Since the information of the interference fringes obtained by the above pattern detector has the information of the pitch and the phase, the information of the inclination and the height can be obtained at the same time. Moreover, when the incident angle is 85 ° or more, the inclination and height of the photoresist surface can be accurately obtained at the same time as described below. The incident angle, the polarization of the incident light, and the detection accuracy will be described below.

If the amplitude of the light incident on the measured object is S-polarized light and As-Ap for the P-polarized light, the amplitudes Rs, Rp and Ds, Dp of the light reflected and refracted on the surface of the object having the refractive index n are the incident angles θ, Refraction angle ψ
For (sin ψ = sin θ / n), it is given by the following formula.

For S-polarized light, the incident angle is 0 ° to 60 °, and for P-polarized light, the transmitted light is larger than the surface reflected light from 0 ° to 75 °.
The reflected light from the boundary of the underlying multi-layer structure causes large-amplitude interference with the surface reflected light. The amplitude of the surface-reflected light increases from the above value to about 85 °, which is an insufficient condition for realizing accurate measurement. The reasons are as follows. As shown in FIG. 17, when the light having the amplitude A incident at the incident angle θ is refracted at the refraction angle ψ amplitude D and is reflected by the amplitude reflectance Rb on the base, the amplitude of the reflected light becomes DRb. Here, assuming that the amplitude of the incident light A is 1, D is the amplitude transmittance. Therefore, when the light reflected on the ground passes through the surface, its amplitude becomes RbD ² . On the other hand, light incident with an amplitude A (= 1) is reflected by the surface and its amplitude becomes R. Here, R and D are expressed by Rs, Ds and Rp, Dp depending on whether the polarization of incident light is S or P, and the above equations (1) to (4) are established. The light R ₀ reflected on the surface and the light R ₁ reflected on the base layer overlap when the layer thickness d is small, and as a result, light having a complex amplitude A _R shown by the following equation is obtained.

Here, λ is the wavelength of light used for measurement. From the equation (5), it can be seen that the phase of A _R changes even with a slight change in the film thickness d shown in FIG. 3 (change in length one digit below the wavelength). Therefore, the relationship between the incident angle θ and R, D is as shown in FIGS. 5 and 6 for S and P polarized light, respectively. To make it easier to understand from this graph, the first component of the equation (5) which becomes a noise component is shown. If the amplitude ratio RbD ² / R of the second term to the term is obtained, the degree of error on the measurement can be evaluated. So consider the case of Rb = 1 as the worst case, D ²
FIG. 7 shows / R calculated with respect to the incident angle θ and two polarized lights. Since D ² / R is a noise (error) component in various detection methods, keep this value below 5%.
It turns out that the incident angle must be 85 ° or more. Also S
Noise can be further reduced if it is incident in the polarized state.
You can see from the figure.

As shown in FIG. 3, the method of reflecting twice on the surface of the object to be measured is to make the detection sensitivity of the inclination and height more than double as compared with the case where the light of 4α is inclined with respect to the inclination α of the surface. And enables highly accurate detection.

Also, the reflected light from the base surface is superimposed on the interference pattern and disturbs the pitch and phase of the interference fringes, but it is possible to almost eliminate this effect by using S-polarized light when incident at 85 ° or more and further using S-polarized light. Highly accurate detection is possible. Furthermore, by making the optical path of the reference light used for this interferometric measurement almost the same as the measuring light, stable and highly accurate measurement that is hardly affected by the measurement environment such as air fluctuation is realized.

In addition, the obtained interference fringe information is subjected to fast Fourier transform (FTT).
When the spectrum is detected by the program, the information of the spectrum corresponding to the frequency of the fringe represents the pitch and the phase, and therefore the inclination and the height can be obtained at the same time from this value. Also FFT
Since it is a matrix operation, parallel operation processing is possible, and if such a parallel operation circuit is used, processing can be performed in 1 ms or less, and inclination and height can be detected and control can be easily performed in real time.

Further, in the case of interference detection, if a change of one pitch of the interference fringes occurs, the detected interference fringes become exactly the same. Therefore, a value obtained by adjusting the moving amount of the integer pitch remains as the uncertain value of the detected value. In the present invention, the second wavelength is used as the detection light to perform interference detection in the same manner as the first wavelength, and it is possible to accurately detect the height over a wide range from the phase relationship between the first wave and the second wavelength. The tilt and height can be controlled. Further, other wafer height detecting means such as an air micro is used to detect an indefinite range when detecting one wavelength, thereby enabling accurate height detection over a wide range.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG. The exposure light emitted from the exposure illumination system 81 illuminates the reticle 9, and the transmitted light is projected by the reduction projection lens 8 on the surface of the wafer 4 on the stage 7 as a reduced image of the pattern on the reticle. The relative position of the reticle 9 and the wafer 4 is detected by the alignment system 800, and fine exposure control of either the reticle 9 or the wafer 4 allows the pattern to be overexposed. 100X is a system for detecting the tilt and height in the X direction, and there is a similar optical system for the tilt in the Y direction, although not shown. The detection system will be described below.

The light emitted from the coherent light source 1 such as a semiconductor laser is converted into parallel light 15 by the lens 11. The parallel light 15 is split into parallel lights 16 and 17 by the beam splitter prism 10. Parallel light
Reference numeral 16 denotes a stage 7 having a beam splitter 12 and a mirror 13 which are irradiation means, and a vertical and biaxial suppression mechanism mounted thereon.
Is incident on the photoresist on the upper surface of the wafer 4, which is the object to be exposed, at an incident angle θ, (88 °). As described above, almost all the light is reflected by the photoresist surface, and this reflected object light 16 'is incident perpendicularly on the plane boundary 14 which is the folding detection optical system, traces the original optical path in the opposite direction, and is exposed. Reflected by the mirror 13, beam splitter 12, as object light 26 ″,
The pattern detection means 3 is reached via the lens 21, the minute aperture plate 23, and the lens 22. On the other hand, the reference beam 17 separated by the beam splitter 10 travels in the same direction as the irradiation beam 16 in the same direction (although strictly speaking, in the direction of an angle of 92 ° with respect to the wafer normal) and is reflected vertically by the plane mirror 14. Then, the light beam travels on the same path as the object light beam 26 ″ as the reference light beam 27 ″ and reaches the pattern detecting means 3 via the wedge glass 24. The reference optical path differs from the object optical path in that it is not reflected by the object 4 to be exposed and that it passes through the wedge glass 24. The lenses 21 and 22 emit the light incident as parallel light as parallel light, and the irradiation position of the irradiation light on the wafer,
That is, the exposure area 0 is imaged substantially on the pattern detector.
If the wedge glass is not present, the intersection point A of the object light and the reference light reflected and returned on the wafer surface forms an image behind the light receiving surface of the pattern detecting means. This means that the two lights are deviated on this light receiving surface. Therefore, a wedge glass is put in the reference light (or the object light) so that both the lights intersect on the light receiving surface and the exposure area 0 forms an image. The minute aperture plate 23 arranged in the folding detection optical system is located at the condensing position of the object light and the reference light, which are parallel lights incident on the lens 21, and has a minute aperture at the condensing point. This minute aperture plate removes back surface reflected light generated by a lens, a beam splitter, etc., which is a problem when using a laser light having high coherence, and prevents noise light from being superimposed on the light receiving surface of the pattern detecting means. The interference fringes detected by the pattern detecting means 3 have an intensity distribution Ix as shown in FIG. The pattern detection means 3 is a one-dimensional array sensor,
The intensity value is obtained at the position marked on the X axis in FIG. 2, and this data is transmitted to the processing circuit 5. When the surface of the exposure area of the wafer is horizontal and coincides with the image plane of the reticle 9 formed by the exposure image forming system (4 in FIG. 3), interference fringes having a solid line pitch P in FIG. 2 are obtained. If the surface of the exposed area is as shown in FIG.
When it is inclined by α as shown by the dotted line LL ′ in the figure, as is apparent from FIG. 3, the first reflected light is inclined by 2α and the returning second reflected light is inclined by 4α. As a result, the interference fringes obtained by the pattern detecting means have a pitch P'as shown by the dotted line in FIG. The interference signal obtained by the detection means is transmitted through the transmission line 31.
Is input to the processing circuit 5 shown in FIG. The input signal is A / D at the timing corresponding to each point indicated by the horizontal axis in Fig. 2.
It is converted and input to the FFT circuit. This FFT input signal is the 4th
It looks like Figure (b), and the result of FFT is the complex number C
As shown in FIG. 4 (C) obtained by (k) (however, the vertical axis of this graph is | C (k) |), there are generally spectrum peaks at k = 0 and K = m. k = 0 corresponds to the sine wave bias component, and k = m corresponds to the sine wave period. m corresponds to the pitch P, but the output can be obtained only discretely, so C (m) is used to obtain the true spectral peak position.
If the true peak position is obtained by performing an interpolation process from the data in and near the point, the slope Δψx can be obtained. Also, the complex number C (m)
Phase (information [Delta] Z of ^{tan -1 (Im (C (m} )) / Re (C (m))) from a height (Z) is obtained. Thus [Delta] [phi] _X obtained, [Delta] Z to the first The vertical and biaxial suppression mechanism of the stage 7 is based on Δψ _Y similarly obtained by the processing circuit 5 from the y-direction interference fringe information obtained by the y-direction detection detection system (not shown). Is controlled to align the reticle image forming surface and the photoresist surface with a desired positional relationship.

FIG. 8 shows an embodiment of the present invention. The same numbers as in FIG. 1 are the same. Also, as in FIG. 1, the diagram of the y-direction tilt detection system is omitted. The semiconductor laser 1 has a wavelength of λ ₁ , and the semiconductor laser 1 ′ has a wavelength of λ _2. For example, λ ₁ = 810n.
The light emitted from the semiconductor laser 1,1 ′ with m, λ ₂ = 750 nm is collimated by 11 and 11 ′, respectively, and the diffraction grating 18,1
8'separates into 0th and 1st parallel rays. The four separated parallel beams are transmitted by the wavelength separation mirror 19 for the light of λ ₁ and reflected by the light of λ ₂ , and the prism 110 makes the four parallel beams parallel to each other. The beams 16 and 16 'of wavelengths λ ₁ and λ ₂ travel through exactly the same optical path, are reflected by the mirror 13, are incident on the wafer at θ ₁ , and the reflected light becomes object light, which is detected by the mirror 23, lenses 21 and 22. The light enters the pattern detection means 3 through the optical system. On the other hand, the beams 17 and 17 'of the wavelengths λ ₁ and λ ₂ pass through exactly the same reference optical path and enter the pattern detecting means 3 at a constant angle with the object light. The object light path and the reference light path pass through exactly the same optical component except for reflection on the wafer surface. The processing circuit 5'blinks the semiconductor lasers 1 and 1'alternately and alternately receives the interference fringe information of the wavelengths λ ₁ and λ ₂ from the pattern detecting means 3. FIG. 9 shows the interference fringe information of the wavelength λ ₁ received by the processing circuit. The solid line is at the best height position and the dotted line is when the height changes by ΔZ. When there is no change in the slope of both detection signals, the phase difference Δφz represented by the following value is generated.

However, when obtaining ΔZ from the detected phase difference Δφz, there is an uncertainty shown by the following equation.

Here, n is an integer. For example, λ ₁ = 0.81 μm, θ ₁ =
If it is twice, it has an uncertain value of true value + 11.6nμm. In this embodiment, this problem is solved by the measurement with the second wavelength λ ₂ . Figure 10 (a) is a detected intensity Iz with respect to the height variation of the wafer surface at the reference position when detected at a wavelength of _{λ 1 (X = X 0)} , FIG. 10 (b) are also of lambda ₂ It is at the wavelength. The intensity of the detected pattern is given by the following equation.

Here, X is the coordinates of the light receiving surface of the detecting means, and M is the image forming magnification. Therefore, Iz = I (X ₀ , ΔZ; λi). λ ₁
It is assumed that the phase value detected in step ₁ is Δψ _1, and the height corresponding to this value corresponds to ... P _-2 , P _-1 , P ₁ , P ₂ , P ₃ ... as shown in FIG. Then, which of these points corresponds to Δ
I don't know if Z is a true value. Assuming that the phase value detected at λ ₂ is Δψ ₂ , the corresponding ΔZ is shown in FIG.
_-1 , ₀ , ₁ , _2, ... The interval S ₁ S ₀ between ΔZ = S ₀ having the same phase and ΔZ = S _{1 having} the same phase next in FIGS. 10A and 10B is given by the following equation: During this period, the phase of λ ₁ is Δψ ₁ , and the phase of λ ₁ becomes Δ ₂ at only one point of ΔZ ₀ , and the next value of ΔZ satisfying this condition is the height given by the following equation.

ΔZ = ΔZ ₁ + mS ₁ S ₀ (10) Here, m is an integer.

At λ ₁ = 0.81 μm, λ ₂ = 0.75 μm, θ ₁ = 2 °, S ₁ S ₀ is
It becomes 145 μm. The height of the wafer surface does not fluctuate in such a wide range, and if wafers of different types are used, the value will be known in advance and no problem will occur. In the embodiment shown in FIG. 1, since a folded plane mirror is used to reflect the light twice on the wafer surface, the value corresponding to S ₁ S ₀ is half that in the embodiment shown in FIG.
It becomes 72.5 μm.

FIG. 11 shows the illumination light 16 in the x direction and the y direction and ▲ ▼ for the region 41 which is exposed on the wafer in one exposure. The position of the illumination light 16 on the wafer is the pattern detection means 3
This corresponds to the address of the array element on the light-receiving surface 301 of. Among the addresses js to je corresponding to the entire illumination area, only a desired area is extracted, for example, Is to Ie in FIG. 11 or Is _{1 to} Ie ₁ and Is _{2 to} Ie ₂ in FIG. 12, and this data is extracted. It is easy to perform FFT using only Since it is possible to specify an arbitrary part in this way, for example, by specifying the part containing the fine pattern as the detection area and removing the part of the rough pattern, the inclination and height of the fine pattern part can be accurately obtained. , It is possible to perform exposure at a position closer to the focus.

FIG. 13 shows an embodiment of the present invention. The same numbers as in FIGS. 1 and 8 represent the same items. Light emitted from the semiconductor laser ₁ having wavelengths λ ₁ and λ ₂ and the collimator lens
11 and ▲ ▼ form parallel light, and the wavelength separation mirror 19 forms the same optical path. The cylindrical lenses 110 and 120 are used to widen the beam diameter in the y direction. FIG. 14 shows the irradiation light 16 that has been expanded to the exposure area 41 on the wafer 4.
The range of (dotted line) is shown. The illuminated portion is the light receiving surface 302 of the pattern detecting means 3 "and 3'comprising a two-dimensional array element
Imaged above. Information of only the desired regions 42 and 43 shown in FIG. 14 among the interference fringes of the irradiation portion obtained two-dimensionally is processed. The inclination and height in the x direction are obtained at the respective positions of 42 and 43, and the inclination and height in the x and y directions of the entire surface 41 are obtained.

FIG. 15 shows an embodiment of the present invention. This drawing is a plan view, and the exposure optical system of the projection exposure apparatus is omitted. Rotation axis 71 in the x-axis direction and rotation axis in the y-direction of the stage restraining mechanism
It is possible to rotate wx, wy minutely around 72. The exposure area 41 on the wafer 4 is irradiated with irradiation light for tilt and height detection from a direction (x direction) tilted by 45 ° from the x-axis, and reflected light is a plane mirror.
It is returned vertically at 14, and the tilt and height detection optics 10
Detected at 0. In this embodiment, the detection light has an inclination of 45 ° with respect to the biaxial elevation mechanism, and an interference pattern as shown in FIG. 16 is generated on the two-dimensional image pickup surface of the pattern detecting means. The interference pattern is a solid line when the wafer surface is kept horizontal, and a dotted line when the wafer surface is inclined in the y direction. Therefore, if the pitch and the phase in the x direction and the y direction on the imaging surface are obtained, the inclination and the height in the x and y directions can be obtained from one detection optical axis system. Further, in the embodiment of FIG. 15, the light beam may be divided into two just before the pattern detecting means, and the x direction and the y direction may be detected by different pattern detecting means.

FIG. 18 shows an embodiment of the present invention. The same numbers as in FIG. 1 represent the same items. The difference from FIG. 1 is the following five points. A tube laser 101 typified by a He-Ne laser or the like is used as a laser source, and a pinhole plate 102 for selecting a part of laser light (center portion of Gaussian distribution) is folded back to form a flat mirror 1
A point provided at a position conjugate with 4 and the beam 15 pass through the beam splitter 12, are reflected by the mirror 14, and the number of reflections until reaching the pattern detecting means 3 is even or odd both in the reference optical path and the object optical path. The point pattern detecting means 3 and the folding plane mirror 14 are formed into a conjugate relationship (image forming relationship), and an intersection point A of the object light and the reference light reflected and returned on the wafer surface is imaged on the pattern detecting means 3. As parallel plane glass
A point where 201 is inserted in the reference optical path 27 ″ (or an object optical path 26 ″), and a point where a correction lens 204 is inserted immediately before the pattern detecting means so that the object light and the reference light incident on the pattern detecting means 3 become plane waves. The point is that the air micro 82 is also used to detect the height of the wafer 4.

The light emitted from the laser source 101 enters the minute opening of the pinhole plate 102 provided at a position conjugate with the folding plane mirror 14 and selects the center of the Gaussian distribution. Next, the lenses 103 and 105 form parallel light 15 having a desired thickness, which is then incident on the beam splitter 106. The parallel light 15 is split into parallel lights 16 and 17 by the beam splitter 106. Here, the parallel light 16 is transmitted through the beam splitter 106 (the number of reflections is 0), and the parallel light 17 is reflected twice in the beam splitter 106. This is beam 15
Passes through the beam splitter 12 and is folded back by the mirror 14,
The number of reflections to reach the pattern detection means 3 is set to be even or odd in both the reference light path and the object light path.
When the direction of the incident light 15 on the light fluctuates, the fluctuation of the crossing angle between the reference light and the object light can be suppressed to a small degree, and as a result, the interference fringe pitch hardly changes, which enables highly accurate detection. . In the case of the present embodiment, the number of reflections of the reference light and the object light up to the pattern detection means after the beam splitting by the beam splitter 106 is 6 times, which is even. The parallel light 16 emitted from the beam splitter 106 is the beam splitter 1
2, almost all the light is reflected by the photoresist surface on the upper surface of the wafer 4 which is the object to be exposed on the stage 7 having the vertical and biaxial tilting mechanism through the mirror 13.
It is vertically incident on the folding plane mirror 14. The parallel light 16 reflected by the folding plane mirror 14 reversely twists the original optical path again, and the object light 26 ″
As a result, it reaches the pattern detection means 3 through the mirror 13, the beam splitter 12, the lens 202, the minute aperture plate 23, the lenses 203 and 204. On the other hand, the collimated light 17 separated by the beam splitter 106 travels in substantially the same optical path as the collimated light 16, and the beam splitter 12,
After passing through the mirror 13 and directly incident on the folding mirror 14 vertically, the original optical path is reversed again and the reference light 27 ″ is mirrored.
13, beam splitter 12, parallel plane glass 201, lens 202,
The pattern detection means 3 through the minute aperture plate 23 and the lenses 203 and 204
Leading to. The reference optical path differs from the object optical path in that it is not reflected by the object to be exposed 4 and passes through the plane-parallel glass 201. The lenses 202, 203 and 204 form an image on the pattern detecting means 3 on a plane perpendicular to the optical axis between the reflecting surface of the folding plane mirror 14 and the exposure center O. This is when detecting the tilt and height from only the information of the part corresponding to the desired location of the exposed object,
This is because the correspondence between the position in the interference fringes and the position on the wafer must be clear. By adopting the above configuration, the image of the wafer in the optical path of the going and returning can be almost even. An image can be formed on 3 and can be detected with high accuracy even if it is partially detected. However, when a plane perpendicular to the optical axis between the reflecting surface of the folding plane mirror 14 and the exposure center O is imaged on the pattern detecting means 3, the exposure area and the intersection point A do not coincide with each other. The reference light and the object light cannot be superposed on the detection means 3. Then, the plane parallel glass 201 is used as the reference optical path 27 ″ (or the object optical path 2).
6 "), and the reference light was moved in parallel so that the reference light and the object light were overlapped on the pattern detecting means 3.
The lens 204 is a lens for correcting the spherical waves of the reference light and the object light generated by the lenses 202 and 203 into a plane wave, and is arranged immediately before the pattern detecting means 3. By making both wavefronts plane waves, it is possible to eliminate variations in the pitch of interference fringes and obtain high detection accuracy.

This detection method has a problem of uncertainties in the height of the wafer, which is shown by the equation (7) in the case of detecting one wavelength. Although the solution using the two-wavelength illumination has been described above, the solution can also be solved by using other wafer height detecting means together. In this embodiment, this problem is solved by using the air micro 82 together. That is, the wafer height is positioned by the air micro to the range where the height can be reliably detected by the present detection method shown by the equation (6), and the present detection method is used within the range of the phase change shown by the equation (6). As another method, the height of the wafer may be detected by other detecting means such as an air micro, and the tilt may be detected by this detecting method. FIG. 21 shows the principle of the air micro, in which pressurized air is supplied from the air pressure source 821 to the air micro nozzle 822 and the reference air micro 823, and inside the air micro nozzle determined by the gap between the air micro nozzle 822 and the wafer 4. The back pressure 824 and the back pressure 825 of the reference air micro 823 are converted into an electric signal by the differential pressure converter 826, and the height of the stage 7 is controlled by the processing circuit 5 so that the differential pressure becomes zero. stop.

FIGS. 19 and 20 are diagrams for explaining the effect of the method of selecting and illuminating a part of the Gaussian distribution according to the present embodiment. FIG. 19 shows a case where a part of the Gaussian distribution is not selected. The figure shows the illuminance distribution on the pattern detection means 3 of the reference light (solid line) and the object light (broken line) when a part of the Gaussian distribution is selected. The vertical axis shows the illuminance Ix, and the horizontal axis shows the detection position x of the pattern detecting means 3. When the wafer is tilted, the object light moves on the pattern detecting means 3, and the overlapping state (hatched portion) of the reference light and the object light changes. At this time, in the embodiment of FIG. 19 in which a part of the Gaussian distribution is not selected, the illuminance at the overlapping portion of the reference light and the object light changes greatly, and the interference intensity also changes greatly. On the other hand, in the illumination in which a part of the Gaussian distribution in FIG. 20 is selected, the illuminance change at the overlapping portion of the reference light and the object light is small,
The fluctuation of the interference intensity is also small. The smaller the fluctuation of the interference intensity, the smaller the error in the process of signal processing and the more accurate detection becomes possible. Since the signal processing of this embodiment has been described above, the description thereof will be omitted.

FIG. 22 shows an embodiment of the present invention. In this embodiment, laser light is irradiated in the diagonal direction of the exposure area 41 of the wafer 4.
100x is a system that detects inclination and height in the x direction, and 100Y is y.
Reference numeral 14 is a folded plane mirror, which detects the tilt and height of the direction. FIG. 23 is an enlarged view of the exposure area 41, which is an example having two circuit parts (memory etc.) in one exposure area. In the figure, 412 indicates a circuit portion and 413 indicates a boundary portion, and the laser beam 411 irradiates the exposure region in a diagonal direction. FIG. 24 shows a cross section taken along the line I--I of FIG. 23. Generally, the circuit portion 412 and the boundary portion 413 have different heights. For this reason, when the laser light is irradiated in a direction perpendicular or parallel to the side of the exposure area, the inclination and height of the boundary portion may be detected instead of the inclination and height of the circuit portion to be originally detected. . On the other hand, as shown in FIG. 23, it is possible to accurately obtain the inclination and height of the circuit section by irradiating a laser beam in the diagonal direction of the exposure area and selecting an arbitrary detection range as necessary. . Further, when the laser light is irradiated in the diagonal direction of the exposure area, the irradiation range can be set to be the longest, and there is an effect that the inclination and height of the exposure area can be detected with a higher degree of accuracy.

Although the semiconductor exposure apparatus has been described in the above-described embodiments, the present invention can be similarly applied to other exposure apparatuses for display devices such as liquid crystal displays and can exert a great effect.

〔The invention's effect〕

Since the present invention is configured as described above, it has the effects described below.

(1) The tilt and the height can be simultaneously obtained by the interferometric measurement.

(2) By arranging the reference light so that the reference light passes through substantially the same place as the irradiation light and the detection light, stable inclination and height detection that is not affected by disturbance factors such as air fluctuation can be performed.

(3) By setting the incident angle to the object to be exposed to 85 ° or more and by making the incident light S-polarized, the inclination and height of the photoresist surface are not affected by the underlying film structure. , Can be accurately detected.

(4) By returning the reflected light of the light obliquely irradiated to the exposed object vertically, and irradiating the exposed object again, the accuracy of detecting the inclination and the height can be doubled.

(5) Particularly precise by making the imaging surface of the pattern detecting means conjugate with the beam irradiation position on the surface of the exposed object and detecting the tilt and height only from the information of the portion corresponding to the desired location of the exposed object. You can focus on where you need to focus.

As explained above, it becomes possible to detect the inclination and height of the surface of the object to be exposed with high accuracy and stability without being affected by the base, and especially for a focal depth such as an LSI with a line width of 0.5 μm or less. This contributes to a significant improvement in the yield of the exposure process, as compared with the pattern exposure without the exposure.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of an embodiment including a folding detection optical system according to the present invention, FIG. 2 is a diagram showing a detection pattern signal waveform, and FIG. 3 is a diagram for explaining an effect of the folding detection optical system. FIG. 4, FIG. 4 is a diagram showing an embodiment of a processing circuit, FIGS. 5 to 7 are diagrams showing the relationship between incident angle and reflection / transmission complex amplitude, and characteristics of noise component ratio, and FIG. diagram illustrating the configuration of those using two wavelengths in one embodiment of the invention, Figure 9 represents the change in the accompanying detection pattern signal to the height variation diagram, when detecting Fig. 10 2 wavelengths lambda _1, at lambda ₂ Of the signal Iz at the point of interest associated with the height change .DELTA.Z of the light, FIG. 11 and FIG. 12 show the irradiation light for the exposure area and the calculation processing area, and FIG. 13 is an embodiment of the present invention. FIG. 14 is a diagram showing the configuration of a two-dimensional detection with two wavelengths, and FIG. 14 shows an arithmetic processing area for the exposure area of the embodiment shown in FIG. FIG. 15 and FIG. 15 are views showing a detection system for detecting tilts in two directions in an embodiment of the present invention, and FIG.
FIG. 17 is a diagram showing the detection pattern, FIG. 17 is a diagram for explaining the amplitude component of the irradiation detection light, and FIG. 18 is one embodiment of the present invention in which an air micro is used in combination to further improve the optical system. FIG. 19 is a diagram showing the configuration, FIG. 19 and FIG. 20 are diagrams for explaining the effects of the illumination method according to the present invention, FIG. 21 is a diagram showing the principle of the air micro according to the present invention, and FIGS. 3A and 3B are explanatory diagrams when laser light is irradiated in a diagonal direction of an exposure region according to the present invention. 1 …… Laser light source, 10,12 …… Beam splitter, 14 …… Folded plane mirror, 21,22 …… Lens, 23 …… Micro aperture plate, 24 …… Wedge glass, 3,3 ″, ″ …… Pattern detection Means, 4 ... Wafer, 5 ... Processing circuit, 7 ... Stage, 8 ... Exposure projection lens, 81 ... Illumination system, 9 ... Reticle, 82 ... Air micro.

Claims (17)

[Claims] 1. A stage for holding an exposure illumination system, a mask or reticle, a projection optical system, an object to be exposed, and moving the object to be exposed in three directions orthogonal to each other, and a relative position between the mask or reticle and the object to be exposed. In a projection exposure apparatus including an alignment system that detects a target position and controls alignment, the light emitted from the coherent light source is converted into parallel irradiation light, which is oblique to the exposure area of the projection optical system on the surface of the object to be exposed. From at least one irradiation means for irradiating at an incident angle θ, at least one detection optical system for guiding the object light reflected by the exposed object to the pattern detection means, and the light emitted from the coherent light source are separated, A means for generating reference light and at least one for guiding the reference light to the pattern detection means, superimposing it on the pattern detection means at a desired angle with respect to the optical axis of the object light, and causing interference. A illuminating means and a processing circuit for obtaining at least one of the inclination and height of the exposed object from the information of the interference pattern obtained by the pattern detecting means, and at least 1 of the exposed object based on the information. The tilt of the direction,
Alternatively, the projection exposure apparatus is provided with a stage control system for controlling at least one of the heights. 2. The projection exposure apparatus according to claim 1, wherein the irradiation light, the object light and the reference light practically travel in substantially the same direction and pass through the same region. 3. The projection exposure apparatus according to claim 1, wherein the incident angle θ is 85 ° or more. 4. The projection exposure apparatus according to claim 1, wherein the irradiation light is S-polarized light. 5. A pattern is formed by arranging a folding plane mirror, wherein the object light reflected by the object to be exposed is reflected substantially vertically by the plane mirror, and the same optical path as the outward path is reversed and reflected by the object to be exposed again. 2. The projection exposure apparatus according to claim 1, further comprising a folding detection optical system for guiding to the detection means, and causing interference with the reference light on the pattern detection means. 6. The projection exposure apparatus according to claim 5, further comprising means configured so that the reference light passes through substantially the same direction and the same region as the object light. 7. The pattern detecting means is an array sensor, and a pitch P corresponding to the inclination Δθ and height Δh of the exposed object.
And transmitting a sine wave signal having a phase φ to the processing circuit,
7. The projection exposure apparatus according to claim 1, wherein a high-speed Fourier transform is executed by the processing circuit, and Δθ and Δh are obtained from information in the vicinity of the spectrum corresponding to P. 8. The irradiation position on the object to be exposed is in a conjugate relationship (imaging relationship) with the pattern detecting means by the detection optical system or the folding detection optical system. 7. The projection exposure apparatus according to 5 or 6. 9. The projection exposure apparatus according to claim 1, wherein the object light and the reference light incident on the pattern detecting means are effectively plane waves. 10. The pattern detecting means is an array sensor, and the irradiation position on the object to be exposed is in a conjugate relationship with the pattern detecting means by the detection optical system or the folding detection optical system. A processing circuit that selects only the desired region of the information obtained in step 1, performs a high-speed Fourier transform operation, and obtains at least one of the inclination and height of the desired region of the exposure region of the exposed object from the obtained spectral information. 7. The projection exposure apparatus according to claim 1, 5 or 6, further comprising: 11. A noisy light component reflected from the front and back surfaces of an optical component in both optical paths by providing a minute aperture in a light converging portion of both the optical paths in the optical path of the detection optical system or the reflection optical system and the reference light. The light is shielded from the light.
The projection exposure apparatus according to claim 1. 12. A coherent light source having a wavelength different from that of the coherent light source is provided, and light emitted from the second coherent light source is introduced into the irradiating means, and the first coherent light source is provided. The object light and the reference light having substantially the same optical path are formed, and the first light from the second light is converted into the second light in the detection optical system or the folding detection optical system.
Means for separating the light having the wavelength of 2 is arranged, the separated object light and reference light having the second wavelength are detected by the second pattern detecting means, and the interferences obtained by the first and second pattern detecting means are detected. 7. The projection exposure apparatus according to claim 1, wherein an uncertain factor in the height direction is removed from the stripe information so that accurate height information can be detected in a wide range. 13. A single beam passes through a beam splitter, and the number of reflections of the reference light and the object light before reaching the pattern detecting means are both even or odd. The projection exposure apparatus described. 14. By using another wafer height detecting means such as an air micro in combination, an uncertain factor in the height direction due to one wavelength detection is removed, and accurate height information can be detected in a wide range. The projection exposure apparatus according to claim 1, 5 or 6, characterized in that. 15. The projection exposure apparatus according to claim 1, wherein the laser light is applied in a substantially diagonal direction of the exposure area. 16. A pinhole plate for selecting a part of laser light is provided in the illumination optical path.
Or the projection exposure apparatus according to item 6. 17. The projection exposure apparatus according to claim 16, wherein a pinhole plate for selecting a part of the laser light is provided at a position substantially conjugate with the exposure area.