JPH03249513A - Method and device for detecting inclination or height, projection exposure device, and method and device for interference fringe detection - Google Patents

Abstract

PURPOSE:To realize this projection exposure device which exposes a high-resolution pattern by irradiating a body to be measured with light from a coherence light source at a >=85 deg. angle of incidence and finding the inclination of the measured body surface and variation in height from the pitch and phase variation of interference fringes formed with reflected light and reference light. CONSTITUTION:The light from the coherent light source 1 is made into parallel illumination light 16, which irradiates an exposed area on the surface of photoresist on a wafer 4 slantingly at a >=85 deg. angle of incidence. The reflected light of this illumination light 16 and the reference light separate from the light from the light source 1 are reflected by a mirror 14 and superposed on an array sensor 3 to generate the interference fringes. Interference fringe information on them is detected to sample information based upon only the reflected light from the photoresist, the sampled information is used to correct the detected interference fringe information, and information on the pitch or the pitch and phase of the interference fringes is calculated from the corrected interference fringe information. The inclination or height of the surface is detected according to the information on the calculated pitch or the pitch and phase of the interference fringes.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a projection exposure apparatus for fine patterns such as semiconductor circuit patterns, display device patterns such as liquid crystals, etc., and in particular, it is capable of exposing the entire exposure area with high resolution. The present invention relates to a projection exposure apparatus equipped with means for detecting the inclination and height of an object to be exposed.

[Conventional technology]

Exposure of fine patterns for semiconductor integrated circuits or TFT
(Thin Film Transist.

r) When exposing a drive circuit pattern in a large field of view pattern of a display device such as a liquid crystal television, it is necessary to expose a pattern that is faithful to the original image and has little variation in line width over the entire exposure area. Especially in the field of semiconductor integrated circuits, line width patterns of 0.5 μm or less will be reduced to 15 mm in the future.
It is necessary to expose the entire near area, but as the pattern becomes finer, the imaging range (depth of focus) becomes 11 μm or less. For this reason, it is essential to accurately align the photoresist surface on the wafer with the pattern image formation surface. To achieve this, it is necessary to accurately detect the inclination and height of the exposed area of the wafer surface (photoresist surface).

In the first known example disclosed in Japanese Patent Laid-Open No. 63-7626, a semiconductor laser is focused on the wafer surface from an oblique direction, and the height is detected by detecting the focused position. . In addition, in this known example, multiple reflections due to the multilayer structure of the wafer are dealt with by using a three-wavelength semiconductor laser, and the light focusing position is changed to the oblique incident direction and the right angle direction, and the heights of different places on the wafer are determined. ing. This known example mainly detects the height, and it is also possible to detect the inclination by changing the measurement location in the oblique incident direction and the perpendicular direction, but it is also possible to detect the inclination.
It is difficult to obtain an accurate value of the inclination even by measuring two places in a narrow area of about 0 mm. In order to achieve highly accurate height detection in this known example, it is necessary to sufficiently condense the light on the wafer, that is, to make the condensed diameter as small as possible. It is necessary to increase the condensing angle (the angle of the outermost ray of the condensed beam with respect to the principal ray), and as a result, the angle of incidence of the principal ray must be made small. When this angle is reduced by /h (the angle from a line perpendicular to the wafer surface becomes smaller), the influence of multiple interference due to the multilayer structure of the wafer increases. In this known example, three wavelengths are used for this problem, but
Each wavelength is affected by interference, so this does not fundamentally solve the problem.

In addition, as a conventional method of detecting inclination,
In the second known example shown in Publication No. 20, tilt detection light different from the exposure wavelength is irradiated through a projection lens, the reflected light is focused, and the tilt is detected from the focused position. Since the light is incident perpendicularly or at a shallow angle, the influence of interference with light reflected from the substrate during transport, which will be described later, cannot be ignored, making accurate detection difficult.

Furthermore, in a third known example of a conventional height detection method for a multilayered object, disclosed in Japanese Patent Application Laid-Open No. 63-247741, the reflected light from the base film is separated separately. It is difficult to perform this method on thin films that appear in the process of creating semiconductor circuits.

Furthermore, in the later stages of the LSI exposure process, for example in the wiring pattern exposure process, the unevenness of the wafer surface becomes large, and the photoresist coated on it also has significant unevenness, although not as uneven as the original wafer surface. It becomes like this.

If the above-described conventional method is applied to such a structure, it becomes impossible to know which part of the uneven photoresist is measuring the slope or height, and therefore the accuracy deteriorates.

[Problem to be solved by the invention]

The above-mentioned conventional technology does not take into consideration the point of accurately obtaining information on the inclination and height within the exposure area for a multilayer structure such as a wafer having a semiconductor circuit pattern.
There was a problem with the highly accurate tilt and height control required for circuit pattern exposure of 5 μm or less.

An object of the present invention is to solve the above-mentioned conventional problems, to accurately detect the inclination and height of the photoresist surface in the exposure area for any wafer in a semiconductor process, and to always keep the image on the image plane at or near the photoresist surface. An object of the present invention is to provide a projection exposure apparatus that exposes a high-resolution pattern with little variation in line width in an optimal position.

[Means to solve the problem]

In order to achieve the above object, the present invention uses parallel illumination light emitted from a coherent light source and irradiates the exposure area of the projection optical system on the photoresist surface on the wafer obliquely at an incident angle θ. , the reflected light and the reference light created by separating the light emitted from the light source are incident on a pattern detector at a desired angle to detect interference fringes.

From changes in the interference fringe pitch and phase, it is possible to determine changes in the inclination and height of the photoresist surface on the wafer. In addition, it is easy to set the incident angle to 85° or more in the present invention, which uses a parallel light beam, and because the incident angle is large, most of the reflection occurs on the photoresist surface, and the reflection on each layer of the underlying layer structure. As a result, the influence of interference that occurs can be almost ignored. Furthermore, if the light incident on the photoresist is made into S-polarized light, the reflection on the surface will be further increased and the accuracy will be improved.

In addition, the light reflected on the photoresist surface is incident perpendicularly on a plane mirror, the reflected light is made incident on the photoresist surface again, and the reflected light is used as object light to obtain information on the interference pattern. It becomes possible to perform detection with twice the sensitivity, and even more accurate detection becomes possible.

In addition, by configuring the reference light to travel in substantially the same direction as the photoresist irradiation light and the object light (reflected light) and to pass through substantially the same area, each optical path is free from disturbances such as air fluctuations. This makes it possible to detect inclination and height that are less susceptible to changes in the surrounding environment.

In addition, if the information on the obtained interference fringes is subjected to fast Fourier transform, and the slope/θ and height/Z are calculated from the resulting information near the spectrum of the fringes, then /θ can be calculated at a high speed that can be considered as real time.
, /z is determined.6 If the photoresist irradiation position is in an optically conjugate (imaging) relationship with the light-receiving surface of the array sensor, which is the pattern detection means, then information on only the desired area on the wafer is selected, It becomes possible to find the slope and height of that part.

In the present invention, by increasing the incident angle, the light is irradiated onto the high parts of the resist, and the concave parts are hidden from the shadows of the white parts by increasing the incident angle. By doing so, the contribution to interference fringe detection is reduced.

As a result, in the present invention, the upper surface of the convex portion is detected, and the detection surface, which was uncertain in the conventional method, is clarified. Therefore, the position of the resist surface can be detected correctly. Furthermore, when the object to be measured has large irregularities, the area of the upper surface of the convex surface affects the intensity of reflected light. As a result, depending on the location of the measured surface, the amount of reflected light increases macroscopically where the area ratio of the top surface of the convex surface is large, so the distribution of interference fringes that occur between the array sensor light receiving surface and the reference light is uniform. It disappears. In other words, as mentioned above, when the object to be measured and the array sensor are in a nearly conjugate positional relationship, the amplitude of the interference fringes is large in the area where the area ratio of the convex top surface of the measured surface is large, and conversely, the amplitude of the interference fringes is large in the area where the area ratio is small. The amplitude becomes smaller. As a result, if such interference fringes whose amplitudes differ depending on the location are subjected to Fourier transform and the slope and height are determined from spectrum information corresponding to the fringe pitch, the accuracy will decrease. For such objects, in the present invention, the above-mentioned array sensor detects the intensity distribution of the pattern of only the reflected light from the measured object without superimposing the reference light, and uses this information to detect the above-mentioned interference. The stripe pattern information is corrected and then Fourier transform is performed. In this way, the corrected interference fringe pattern will have the same amplitude almost everywhere, regardless of the location.
It becomes possible to accurately determine slope and height from spectral information.

[Operation] Since the interference fringe information obtained by the above pattern detector includes pitch and phase information, inclination and height information can be obtained at the same time. Furthermore, when the incident angle is set to 85 degrees or more, the reflection on the photoresist surface becomes large, and the inclination and height of the photoresist surface can be determined accurately and at the same time.

Furthermore, as mentioned above, in the latter stage of the exposure process, for wafers with large photoresist irregularities, only the reflected light from the object to be measured is detected, and data on this detected intensity distribution is used as a correction value to detect the convexities on the resist surface. Mainly the height and inclination of the top surface can be accurately determined. This effect will be explained in detail below. Now, if the cross-sectional structure of the wafer 4 is an uneven 42 with relatively large steps on the Si substrate 43 as shown in FIG.
When the layers 41 and 42 overlap, the photoresist 41 applied thereon is small compared to the uneven layer 42, but
Uneven steps remain. If the incident angle is 85° or more (for example, 88
When irradiated with a parallel laser beam (°), only the shaded part in Fig. 10 is specularly reflected, and the other parts of the beam are reflected as shown in Fig. 11 as an enlarged view of Fig. 10.
A, B1. As shown by the ray C□, the light is scattered and reflected in a direction different from the direction of the specularly reflected light. As a result, as will be described later, the detection system that extracts only specularly reflected light has A1. B engineering, C
Light that hits anything other than the top surface of the convex portion, such as the light ray 1, does not reach the top surface of the convex portion. In the case of such a cross-sectional structure consisting of concave and convex portions, light with an intensity approximately proportional to the area of the upper surface of the convex portion reaches the detector.

Moreover, since the wafer surface and the light-receiving surface of the array sensor have an almost conjugate relationship, the cross-sectional structure is mostly flat, or even if there is unevenness, it corresponds to a part where the area ratio of the convex top surface is large. tJs is large where it is strong, and vice versa. As a result, the intensity distribution Ox of the reflected light from the object to be measured on the array sensor varies in level depending on the location, as shown in FIG. 5, for example. When the light having such a distribution interferes with the reference light Rx at a constant level as shown in FIG. 6, the amplitude of the interference fringe intensity Ix will vary depending on the location, as shown in FIG. The phenomenon explained above will be further explained theoretically and quantitatively. The intensity of the reflected light from the measured object that enters the array sensor is Ox, the incident angle is C2, and the intensity of the other reference light is R.
x, the incident angle is 1α (minus indicates that the light is tilted to the opposite side of the reflected light from the object to be measured with respect to the normal to the array sensor surface), and both of these lights are tilted in the X direction. When you are there,
Detect stripes that change in the X direction. The obtained interference fringe intensity factor x (X) is where λ is the wavelength of the detection light, 80 is the inclination of the object to be measured, n is the number of reflections on the object to be measured, which will be described later, and m is This is the imaging magnification of the detection optical system.

φ(Z) is a phase change due to a change in height.

Equation (1) becomes as follows when n=1 since △θ〈×α1〈〈1.

Ix (X) = Ox + Rx + 25 cost: (2 sin a, -
to j'cosa□)X+φ(Z)) −= (1)
'If Ox and Rx take constant values regardless of X, then (
1) Using the spectral peak position corresponding to the fringe period of the Fourier spectrum obtained by Fourier transforming the inspection signal given by the formula ' and the data in its vicinity, /
It becomes possible to obtain the values of θ and φ(Z). However, in general, Ox and Rx are not constant, and especially when the unevenness of the resist surface of the wafer differs depending on the location as described above, Ox changes as shown in FIG. 5. Ox in Figure 5
On the other hand, even if the reference light distribution Rx is a constant value Rc as shown in Fig. 6, the interference fringe intensity given by equation (1)' varies in amplitude depending on the location as shown in Fig. 7. . When such an amplitude change occurs, Tx has a spread around the peak value, and information about the original fringe period (inclination) and phase (height) is hidden, resulting in a decrease in accuracy.

Therefore, before detecting interference fringes, the reference light is blocked.
Only the reflected light from the object to be measured is detected by the same detection system. This value is naturally ○X. Furthermore, when Rx=Rc is not the case as in the example of FIG. 6, for example when Rx is as shown in FIG. 14, the intensity distribution Rx of only the reference light is measured. Using these two measured values (Ox, Rx) or one measured value (Ox) as a correction value when Rx=Rc, the following correction calculation is performed to derive a correction signal Icx.

Since Ix is given by formula (1) λ is found.

When this correction signal Icx is Fourier-transformed, the spread around the curve disappears, a sharp spectrum of a pure trigonometric function is obtained, and the slope and height can be determined accurately.

〔Example〕

Hereinafter, the present invention will be specifically explained with reference to Examples.

FIG. 1 shows an embodiment of the present invention. In this embodiment, the inclination and height detection of the present invention is applied to an exposure chip of a semiconductor exposure apparatus. Reference numeral 9 denotes a reticle on which a pattern original plate used for exposing a circuit pattern onto the wafer 4 is drawn. This reticle 9 is irradiated by an exposure illumination system 81 with g-rays emitted from a mercury lamp (not shown), i-rays, or far ultraviolet light emitted from an excimer laser. The light that has passed through the reticle passes through the reduction lens 8 and is projected as an image of the reticle onto the surface of the resist coated on the wafer 4 fixed on the surface of the wafer. As the pattern drawn on the wafer 4 becomes thinner to submicron, half micron, or even around 0.3 μm, the depth of focus becomes shallower, and when the line width becomes 0.5 μm or less, the slope and height are adjusted to ± If detection and control are not performed with an accuracy of within 10' rad, ±Q, 1 μm, product yield will decrease.

100X in FIG. 1 is a detection system that determines the horizontality of the resist surface on the wafer in the X direction, that is, the inclination and height associated with rotation about the y-axis. - In Figure 1, there is an X as well as 100X.
100Y, which detects the levelness of the direction, is omitted. Reference numeral 1 designates a highly directional light source such as a semiconductor laser or a gas laser, and the light emitted from this light source is split into two by a beam splitter 10 in a substantially parallel state with a desired spread.

One of the two divided beams 16 passes through a beam splitter 12° mirror 13 and irradiates the resist surface of the wafer 4 at an incident angle of 85° or more with respect to the vertical line and as S-polarized light. The resist is transparent to this irradiation light, but since the incident angle is large and the light is S-polarized, nearly 90% of the light does not enter the resist and is directly reflected, and enters the folding mirror 14 almost perpendicularly. The reflected light 26' follows the same optical path as the outgoing path but in the opposite direction, is reflected by the beam splitter, and becomes light 2 heading toward the detection optical path.
It becomes 6'. The other beam 17 separated by the beam splinter 10 is used as a reference beam. The beam 17 is turned off at a desired timing by a shutter (light shielding means) 330.

When the shutter 330 is turned on, the reference light passes through the beam splitter 12 and the mirror 13 and directly enters the folding mirror 14 perpendicularly. The reflected beam 27' follows the same optical path as the forward path but in the opposite direction, is reflected by the beam splinter 12, and becomes light 27' heading toward the detection optical path. The two said beams 26' and 271 towards the detection optical path are connected to the lens 21.22.
The light passes through a light shielding plate 23 having a pinhole or a minute rectangular opening for removing noise, and is superimposed on the array sensor 3 in a substantially parallel state. On the array sensor 3 there is a second
Interference fringes as shown by the solid lines in Figures and Figure 3 occur,
This intensity distribution Ix is detected by the array sensor 3. If the parallelism of the wafer 4 is poor, or if the wafer warps as it goes through various processes, if the wafer 4 is exposed chip or multiple chips in step-and-repeat steps, the imaging surface of the reduction lens and the resist may Since the surfaces do not match in terms of inclination and height, the processing circuit 5 detects the inclination and height using the method described below based on the detection information of the array sensor 3 described above, and based on this detection information, The wafer stage 7 was equipped with the following. A stage control mechanism (for example, using a piezo or mechanical fine movement mechanism) is controlled to align the formed image with the resist surface and expose the resist. In FIG. 1, 800 is an alignment system used for overlapping exposure.

FIGS. 2 and 3 show signals detected by the array sensor 3 in the embodiment shown in FIG. 1, and the signal Ix shown by a solid line in each drawing is the one when the slope and height are in the optimal exposure state. It is. When the wafer 4 deviates from the optimum exposure state and tilts, the pitch of the stripes changes from P to P' as shown by the dotted line in FIG. Further, when the wafer deviates from the optimum exposure position in the height direction, the phase φ2 changes as indicated by the dotted line in FIG. The pitch P and phase φ2 of the interference fringe due to this change in inclination and height can be calculated by entering n=2 in equation (1), and from cos α and zl, λ P;
...(3) 2 sin a 1+(4Δθ/m) λ holds true. However, θ□ is the angle of incidence on the wafer, and φS is the initial constant of the phase.

FIG. 4 shows a portion 51 of the processing circuit 5 of FIG. 1 according to the present invention, and shows one embodiment of the present invention. It is detected by the array sensor 3 in FIG.
The converted interference fringe digital information Ix is temporarily stored in a memory 501 shown in FIG. Prior to detection of the interference fringes, the wafer 4 is loaded onto the stage 7, and before exposure begins, the shutter (light shielding means) 330 shown in FIG. Only 26 is detected by the array sensor 3. Similar to the detection of interference fringes, this signal Ox is A/D converted by an A/D conversion circuit 31 and stored in the memory 2 of 502 shown in FIG. 4 in the form of digital information.
is stored in. As mentioned above, if the enlarged view of the surface of the resist on the wafer 4 has an uneven shape as shown in FIG. 10, it will contribute to the specularly reflected light among the light incident at the incident angle θ1 (85° or more, for example, 88°). This is the top surface of the white part.

The hatched beam in the enlarged view of FIG. 10 is specularly reflected light. When the area of the upper surface of the convex portion becomes smaller, the image on the array sensor 3 due to the reflected light from the uneven structure portion becomes darker. On the other hand, the image on the array sensor 3 corresponding to a portion with small unevenness or a large area on the upper surface of the convex portion of the unevenness becomes brighter, and as a result, the reflected light image ○χ from the wafer 4 as shown in FIG. formed on the sensor 3. As mentioned above, the shutter 33 shown in FIG.
Since this Ox can be detected by closing 0, this information is stored in 502 of the memory 2. When moving the wafer 4 and repeating exposure by step-and-repeat, if the array sensor 3 detects interference fringes prior to exposure, if the distribution of the reference beam 17 is uniform as shown in FIG. 6, then as shown in FIG. An interference pattern Ix such as
is stored in Therefore, these two pieces of information Ix and Ox are taken out from each memory 501 and 502 and 503 of calculation means 1.
The following calculation is performed to derive the corrected interference waveform 1 (X.

However, the intensity of the reference light is approximately a constant value Rc. As shown in FIG. 8, Icx obtained in this manner has significantly fewer components other than the fundamental frequency. As a result, this signal is subjected to fast Fourier transform (
A pure spectrum of wavenumbers obtained by FFT) is obtained, and from this spectrum information, highly accurate slope and height information (Δθ , ΔZ) are obtained.

FIG. 12 shows an embodiment of the present invention. The same part numbers as in FIG. 1 represent the same parts. In FIG. 12, the laser beam irradiated on the transmission grating 18 with a nearly parallel beam is applied to the wafer 4.
The beam is separated into light 16' and reference beam 17'. Each light is blocked by a light shutter 330'. That is, when the shutter 330' is in the state shown in FIG. 12, interference fringes are detected by the array sensor 3;
30' moves to the right, opening A. When overlaps with the position of AR in FIG. 12, the wafer illumination light 16' is blocked and the reference light 17
' is detected by the array sensor 3.

When this reference light 17' is detected, a signal Rx whose distribution is slightly uneven as shown in FIG. 14 is detected. Also shutter 3
30' moves to the left, opening A1 becomes A in FIG. When the reference light 17' overlaps with the position , the reference light 17' is blocked and only the wafer reflected light is detected by the array sensor 3. This wafer reflected light has a distribution ○X as shown in FIG. 13, similar to the embodiment shown in FIG. These distributions Rx and Ox are stored as outputs of the A/D conversion circuit 31 in the memory 2' at 502' in the slope and height-to-lateral ratio circuit 51' in FIG. 17, which is a part of the processing circuit 5' in FIG. and are stored in 502 of memory 2, respectively.

502' of memory 2' and 50 of memory 2 prior to exposure.
Calculating means 1 using the above Rx and ○X stored in 2.
5o3', the following calculation is performed before exposure by step-and-repeat.

Ix is stored in 501 of memory 1 as in the case shown in FIG.

Ix-Ox-Rx 1°"" 2Ei Ding ""6) This Icx is compared with the interference detection waveform Ix (stored in 501 of memory 1) shown in FIG. 15, and as shown in FIG. , since the fundamental frequency component is the main component, it is possible to accurately determine the slope and height (Δθ, ΔZ) from this signal Icx by the fast Fourier transform means 504 and the calculating means 2' 505, as described above. In this embodiment, as compared to the embodiment shown in FIG. 1, accurate detection is possible even if there is some unevenness in the distribution of the reference light 17'.

Not only is it possible to omit an optical system such as a pinhole for making the distribution of the reference light 17' uniform, but the efficiency of light utilization is also increased, and detection can be performed with a light source with a small output. .

FIG. 18 is a diagram showing a specific configuration of the highly accurate calculation means 2' 505 for calculating the pitch (inclination) and phase (height) of interference fringes by the spectral information processing shown in FIGS. 4 and 17. Since the detected and corrected signal Ixc contains almost no signal other than the fundamental frequency component, the pitch and phase can be accurately determined by performing the calculations described below. I
xc is a real number, which is converted into a complex Fourier transform means 504
When the complex Fourier transform is performed, the following formula (FBI) is obtained.

It has a sharp peak value at the spectral position j=,jo corresponding to the wavenumber component. jO, j, I, Js of the discretely obtained spectrum Ajo calculated in calculation step 5051
The value A 3 + I HA J at the -1 position. - From □, the pitch P and phase φ2 can be accurately determined by the method shown below. A jo HA 3゜+1. Rewrite the complex value of Ajo-1 into a real part and an imaginary part as follows.

A jo= Ro+ i I o
...(8) Ajo+, =R++iI
...(9) A,, -1=R-+i I-...
(10) From these values obtained by FFT, the next value (inner product value of the complex vector) is calculated in the inner product calculation step 5052.

So = Ro2+ I o2(= (Ajo−Ajo
)) −(11) S + = RoR+ + X o
I +(=(Ajo・A J-1)) −(12)S
=RoR-+IoI-(=(AjoAxa-t)l・
・(13) Use this value to find the next Δ.

S+-S Using Δ thus obtained, the true spectral peak j is determined by the following equation at the true peak position step 5055.

jR=JO+△...(14) Next, the phase value φ2 is given by the following equation using the above △ and the phase step 5056 at the true peak position, where φjo(
6) is obtained from the equation and satisfies the following equation.

exp jφjo= (Ro + i I o )/
1”・(16) Using △ given by equation (14), (1
The initial phase value φ2 is found from equation 5).

In the same way as when formula (1) was calculated, from formula (1), the pitch P of the interference fringes is determined by the number n of reflections on the object to be measured (wafer 4). The number of sample points per pitch of the interference fringe signal (real number) is given by P = N / J R... (17), so from equations (3)' and (14), Δθ = C5X F, 1J
s)/n...(18) However, C1=λ
m/ (2N) (constant) j, :2 N5inα
□/l (constant) to the stage tilt control amount is determined by stage tilt control amount calculation step 5057.

This 80 is the stage tilt control amount.

On the other hand, from the true phase value φZ, the constant initial phase φ, (Δ
2=0, that is, the phase at the initial focal point), and in the same way as formula (4) was obtained, Δ2=λ(φ2−φg)/(4nπcosθ1), so (in formula (4), n =2) ΔZ=Cz(φ2-φ8)/n However, C2=λ/(4πcosθ□) (Constant) Stage vertical control amount △Z is stage vertical control amount calculation step 50
58. This Δ2 becomes the stage height control amount. Similar to the above 100X X-axis detection, the Y-axis is also 1
The above stage tilt control amount Δθ and stage height control amount ΔZ are determined from 00Y. Δθ of this X axis and Y axis,
Let ΔZ be ΔθX, ΔZx, Δθy, and Δzy, respectively.

Two values are obtained for ΔZ, so either use only one, or use the average ΔZ = (ΔZx + ΔZy)/2 and calculate Δ
By controlling the two directions perpendicular to the optical axis of the wafer stage 7 and the optical axis direction using the three values θX, Δθy, and ΔZ (or ΔZx or ΔZy), the focal plane (tilt and height of the exposure optical system is constant) ), and the circuit pattern formed on the reticle 9 can be exposed onto the wafer 4 using the exposure reduction lens 8 as a high-resolution circuit pattern. Moreover, this detection calculation can be performed at high speed (at the bottom), so if it is performed every time a chip on the wafer 4 is moved step by step and exposed, the entire wafer can be exposed with high resolution and high throughput. becomes possible.

In the above embodiment, Fourier transform is used as a method for determining the pitch and inclination from the corrected interference fringes. It is also possible to obtain the slope, and the method is not limited to Fourier transform. Furthermore, the inclination and height detection method of the present invention is not limited to the semiconductor exposure apparatus of the above embodiment, but can be particularly effectively applied to objects where the distribution of reflected light from the detection object is uneven. It is.

〔Effect of the invention〕

According to the present invention, it is possible to accurately detect the inclination and height of the surface of an optical multilayer object such as a semiconductor wafer, which has an uneven surface shape. As a result, especially for the future miniaturization of LSIs, it is possible to expose LSI patterns with a high yield in all processes, regardless of the surface condition of the wafer, even if a reduction exposure device with a relatively shallow depth of focus is used. becomes.

Further, according to the present invention, the inclination and height of the surface of a wide range of objects can be determined with high accuracy, without being limited to the above-mentioned objects, and regardless of the layer structure or pattern state of the surface.

Further, according to the present invention, the period, pitch, and initial phase of a signal that generally has a periodic waveform can be detected very accurately, and the above-mentioned measurement with high precision can be performed over a wide range of applications.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram showing one embodiment of the present invention, FIGS. 2 and 3 are diagrams for explaining the invention in detail, and FIG.
A schematic configuration diagram showing a part of the processing circuit shown in the figure, FIGS. 5 to 9 are diagrams each showing the effects of the present invention, and FIG. 10 shows reflected light of irradiation light at a high angle of incidence on a concavo-convex pattern. Figure, 11th
The figure is a partially enlarged view of FIG. 10, FIG. 12 is a configuration diagram showing another embodiment of the present invention different from FIG. 1, and FIGS. 13 to 16 are diagrams each showing the effects of the present invention. Figure 17 is the 12th
FIG. 18 is a schematic diagram showing a part of the processing circuit shown in the figure, and FIG. 18 is a diagram showing the calculation flow of the calculation means 2.2' shown in FIGS. 4 and 17. DESCRIPTION OF SYMBOLS 1...Laser light source, 10.12...Beam splinter, 3...Array sensor, 4...Wafer, 5-mountain processing circuit, 7...Stage, 8...Reducing lens, 9...Reticle, 81...Lighting system, 800...Alignment system 3rd jρ, Sha, Yria 2 Kamoshi Otsuko Ryo Kyouzu Akira Ku/θ Koku/2 Akira 74 Ku Akira 150 1st Otsuguchi Jjl)/7 country

Claims (1)

[Claims] 1. Irradiates a measured object with coherent light, and obtains interference fringe information generated by the reflected light from the measured object and a reference light that interferes with the coherent light. Collect information based only on the reflected light from the object to be measured, use this collected information to correct the detected interference fringe information, and calculate the pitch or interference fringe information from this corrected interference fringe information. An inclination or height detection method characterized by calculating information on the pitch and phase, and detecting the inclination or height of the object to be measured based on the calculated pitch of interference fringes or the information on the pitch and phase. . 2. The method for detecting inclination or height according to claim 1, characterized in that information from only the reference light is collected, and the collected information is used to perform the correction. 3. Claim 1, characterized in that the above correction is performed using the following formula:
The described inclination or height detection method. I_x_c=(I_x-O_x-R_c)/(2√{O
_x×R_c}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on reflected light from the object to be measured, and R_c is a predetermined value. 4. Claim 2, characterized in that the above correction is performed using the following formula:
The described inclination or height detection method. I_x_c=(I_x-O_x-R_x)/(2√{O
_x×R_x}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on the reflected light from the object to be measured, and R_x is information based only on the reference light. 5. The information on the pitch and phase of the interference fringes is calculated based on spectral information obtained by Fourier transforming the corrected interference fringe information. method for detecting the inclination or height of 6. A first irradiation means that irradiates the object to be measured with coherent light, and a second irradiation means that irradiates a reference light that interferes with the coherent light.
irradiation means, and interference fringes for detecting interference fringe information generated by the reflected light obtained from the measured object by the light irradiated from the first irradiation means and the reference light irradiated from the second irradiation means. an information detection means, a collection means for collecting information based only on the reflected light from the object to be measured, and correcting the interference fringe information detected by the interference fringe information detection means using the information collected from the collection means. a correction means, and calculates the pitch of the interference fringe or information on the pitch and phase from the interference fringe information corrected by the correction means, and calculates the pitch of the interference fringe or the information on the pitch and phase to be measured based on the calculated pitch of the interference fringe or the information on the pitch and phase. 1. A tilt or height detection device comprising a tilt or height detection means for detecting the tilt or height of an object. 7. The collecting means further includes means for collecting information based on only the reference light, and the correcting means further includes means for performing the correction using the information collected by the means only on the reference light. The inclination or height detection device according to claim 6. 8. The inclination or height detecting device according to claim 6, wherein the first irradiation means is configured such that the angle of incidence of the coherent light onto the object to be measured is 85 degrees or more. 9. The inclination or height detecting device according to claim 6, wherein the correction by the correction means is performed using the following equation. I_x_c=(I_x-O_x-R_c)/(2√{O
_x×R_c}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on reflected light from the object to be measured, and R_c is a predetermined value. 10. The inclination or height detecting device according to claim 7, wherein the correction by the correction means is performed using the following equation. I_x_c=(I_x-O_x-R_x)/(2√{O
_x×R_x}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on the reflected light from the object to be measured, and R_x is information based only on the reference light. 11. Calculation of information on the pitch of the interference fringe or the pitch and phase in the inclination or height detection means is performed based on spectral information obtained by Fourier transforming the corrected interference fringe information. The inclination or height detection device according to claim 6, 7, 8, 9 or 10. 12. In a projection exposure method in which a circuit pattern formed on a mask is projected onto a substrate using a projection optical system, coherent light is irradiated onto the substrate, and the reflected light from the substrate and the coherent light are Detect the interference fringe information generated by the reference light that interferes with the board, collect information based only on the reflected light from the substrate, use this collected information to correct the detected interference fringe information, and perform this correction. Calculating the pitch of the interference fringe or information on the pitch and phase from the interference fringe information obtained,
The tilt or height of the substrate is detected based on the calculated pitch of the interference fringes, or information on the pitch and phase, and at least one of the mask or the substrate is moved as described above based on the information on the detected tilt or height of the substrate. A projection exposure method characterized in that leveling control is performed by finely moving the projection optical system around two axes perpendicular to the optical axis and along the optical axis so that the imaging plane of the circuit pattern of the mask and the substrate surface coincide. . 13. The projection exposure method according to claim 12, wherein information based only on the reference light is collected and the correction is performed using the collected information. 14. The projection exposure method according to claim 12, wherein the correction is performed using the following equation. I_x_c=(I_x-O_x-R_c)/(2√{O
_x×R_c}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on reflected light from the object to be measured, and R_c is a predetermined value. 15. The projection exposure method according to claim 13, wherein the correction is performed using the following equation. I_x_c=(I_x-O_x-R_x)/(2√{O
_x×R_x}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on the reflected light from the object to be measured, and R_x is information based only on the reference light. 16. The information on the pitch and phase of the interference fringe is calculated based on spectrum information obtained by Fourier transforming the corrected interference fringe information. projection exposure method. 17. In a projection exposure apparatus that projects and exposes a circuit pattern formed on a mask onto a substrate using a projection optical system, at least one of the mask or the substrate is tilted around two axes orthogonal to the optical axis of the projection optical system, and a moving means that moves along the optical axis; a first irradiation means that irradiates the substrate with coherent light; and a second irradiation means that irradiates the substrate with reference light that interferes with the coherent light. , an interference fringe information detection means for detecting interference fringe information generated by the reflected light obtained from the substrate by the light irradiated from the first irradiation means and the reference light irradiated from the second irradiation means; a collection means for collecting information based only on reflected light from the substrate; a correction means for correcting the interference fringe information detected by the interference fringe information detection means using the information collected from the collection means; Calculating the pitch of the interference fringe or information on the pitch and phase from the corrected interference fringe information,
An inclination or height detection means for detecting the inclination or height of the substrate based on the calculated interference fringe pitch or information on the pitch and phase; and an inclination or height detection means detected by the inclination or height detection means. 1. A projection exposure apparatus comprising: a leveling control means for controlling the moving means so that the imaging plane of the circuit pattern of the mask and the substrate surface coincide with each other based on information. 18. The projection exposure apparatus according to claim 17, characterized in that the sampling means includes light shielding means. 19. The collecting means further includes means for collecting information based on only the reference light, and the correcting means further includes means for performing the correction using the information collected by the means only on the reference light. The projection exposure apparatus according to claim 17. 20. The projection exposure apparatus according to claim 17, wherein the first irradiation means is configured such that the angle of incidence of the coherent light onto the object to be measured is 85 degrees or more. 21. The projection exposure apparatus according to claim 17, wherein the correction by the correction means is performed using the following equation. I_x_c=(I_x-O_x-R_c)/(2√{O
_x×R_c}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on reflected light from the object to be measured, and R_c is a predetermined value. 22. The inclination or height detection device according to claim 19, wherein the correction by the correction means is performed using the following equation. I_x_c=(I_x-O_x-R_x)/(2√{O
_x×R_x}) where I_x_c is corrected interference fringe information, I_x is detected interference fringe information, O_x is information based only on the reflected light from the object to be measured, and R_x is information based only on the reference light. 23. Calculation of information on the pitch of the interference fringe or the pitch and phase in the inclination or height detection means is performed based on spectral information obtained by Fourier transforming the corrected interference fringe information. 23. The projection exposure apparatus according to claim 17, 18, 19, 20, 21, or 22. 24. Fast complex Fourier transform is performed on the signal strength information that changes at an approximately constant period obtained from the interference fringes, and data A_j_ with the maximum absolute value near the spectral position corresponding to the period of the complex Fourier-transformed discrete spectrum information is obtained.
o (_j_o is an integer, A_j_o=R_o+iI_o)
and data A_j_o_-_1 (=R_-+
iI_-), A_j_o_+_1(=R_++iI_+
) to calculate the true spectral peak J_R using
)S_o=R_o^2+I_o^2 S_+=R_o・R_++I_o・I_+ S_−=R_o・R_−+I_o・I However, R represents the real part and I represents the imaginary part. The pitch P of the period of the signal strength information is calculated from the relationship P=N/j_R for the number of sample data points N (2:n is an integer) of the complex Fourier transform. Stripe detection method. 25. Fast complex Fourier transform is performed on the signal strength information that changes at an almost constant period obtained from the interference fringes, and data A_j_ with the maximum absolute value near the spectral position corresponding to the period of the complex Fourier-transformed discrete spectrum information is obtained.
o (_j_o is an integer, A_j_o=R_o+iI_o)
and data A_j_o_-_1 (=R_-+
iI_-), A_j_o_+_1(=R_++iI_+
) using Δ=(S_+-S_-)/(S_+-2S_o+S_-
)S_o=R_o^2+I_o^2 S_+=R_o・R_++I_o・I_+ S_−=R_o・R_−+I_o・I However, R represents the real part and I represents the imaginary part. The interference fringe detection method is characterized in that an accurate phase value φ_2 is calculated using the calculated Δ and the phase value φ_o of A_j_o as follows: φ_z=φ_o+π(N-1)/NΔ.