JPH02262320A - Measuring method for intensity of interference pattern - Google Patents

Abstract

PURPOSE:To measure the contrast of an interference pattern without being affected by control precision of exposure amount, spreading irregularity of resist, etc., by obtaining the contrast value of interference fringes from results of a specified measurement for a resist image obtained when a sensed image, wherein the edge of a shielding body is out of focus, is exposed to light and etched. CONSTITUTION:A substrate covered with a resist layer having a specified thickness is arranged on an irradiation surface of a coherent beam; a shielding body is arranged between the irradiation surface and a beam source, and the beam is projected; an image wherein the edge of the shielding body is out of focus is transferred on the above resist layer with a specified exposure amount; thus an image which is not focussed is etched on the resist layer. Next the following distances X1 and X2 are measured. The distance X1 is an interval from a starting point where, in a remaining resist layer corresponding with the above out-of-focus image, the thickness of the resist layer begins to decrease, to a point wherein the thickness of the resist layer becomes firstly nearly equal to zero. The distance X2 is an interval from the above starting point to the most distant point where the remaining of the resist layer can be recognized. Based on the distances X1, X2, the ratio of (X1+X2) and (X2-X1) is obtained as the intensity of contrast of an interference pattern.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention is directed to a beam irradiation device used in an exposure device for semiconductor manufacturing that uses a coherent beam such as an excimer laser beam. The invention relates to a method for measuring the intensity (contrast intensity) of fine illumination unevenness such as an interference pattern called .

[Prior Art] Conventionally, although this type of method for measuring speckle intensity is not necessarily well known, for example, as shown in Fig. 2, the amount of illuminating light is changed by a minute pinch. It has been thought that the exposure illa, in which a small amount of a photosensitive material called resist remains after development, and the exposure 311b, in which a small amount of solder is removed, are used to determine the contrast intensity due to speckles. That is, at the exposure amount Ia in FIG. 2(A), resist remains as shown in FIG. 2(B) at the position of the minimum light amount 1 sin a due to the contrast due to speckles, and in FIG. Now, the maximum light illumination is due to the contrast caused by speckles. The resist will come off at the Mb position as shown in Figure 2 (D).
Assuming I ai++ a = Immx b = It, this is the dark value that determines whether the resist remains or not.

Now, if the contrast intensity due to speckles that we want to find is C, then I saw a Ta1n a Is
my a Ta1n aIwrax a +
1. ta a 2 1 al, □
b-1,,ll b 1. □ b-1,1,bl
, □ b+1−+−b 21b.

Therefore, I sag a + I 5tna ko21 a1++
+ax a 1ain a = 21 Il from a-C
If you delete lmXa and summarize, Ill! II a=r
a ・(I C). Similarly, 1,,,b+1. ．． 1 lb=2 I bl, 1. Xb
−1, tllb=2 T b If I□7 b is deleted from −C and summarized, ■□, b=Ib・(1+C). At the threshold It, Imina=Im□b, so the contrast intensity C (a small number between 0 and 1) becomes Ia+Ib, and the exposure amount 1a is such that the resist images become as shown in FIGS. 2(B) and 2(D). I was looking for more speckle contrast than RB.

[Problem to be solved by the invention]

However, in the conventional techniques as described above, a speckle contrast smaller than the exposure control accuracy is not required. Also, when exposing speckles on a substrate such as a wafer, the exposure position is changed, so the contrast of the speckles at two different exposure positions must be determined, and depending on the unevenness of resist coating on the wafer. , the resist remains, but it becomes unstable to specify the dark value (It in FIG. 2) indicating whether the resist remains or not, resulting in poor accuracy. Furthermore, if the contrast of speckles changes due to exposure control, the difference cannot be determined. Additionally, there were various problems in that the contrast between the projected pattern and speckles could not be measured at the same time.

The present invention has been made in consideration of the above points, and is capable of measuring the contrast of speckles (interference patterns) without being affected by exposure control accuracy, resist coating unevenness, etc. The purpose is to make it possible to simultaneously observe the correlation between the contrasts of the two images.

[Means to solve problems]

In order to solve the above problems, the present invention provides a defocused image (
Disclosed is a method for measuring speckle contrast using a blurred image. This method of measuring speckle contrast using a defocused image allows accurate speckle contrast measurement regardless of exposure control accuracy. Furthermore, since the contrast of speckles can be measured with one exposure, it is less susceptible to uneven coating of the resist, and the influence of speckles on the pattern can be accurately determined.

In the present invention, the defocused image refers to an image in which the shadow of the edge of the shielding object is largely blurred on the resist layer, and does not necessarily mean that the image is defocused in a projection exposure method using a projection lens, etc. In other words, even with a method such as the proximity method in which the mask and photosensitive substrate are brought close to each other, if the gap between the mask and the photosensitive substrate is intentionally widened, a similar blurred image can be created. In either the projection type or the proximity type, a blurred image can also be created by inserting the edge of another shielding plate directly above the photosensitive substrate with the mask (or reticle) removed.

[For production]

Here, the manner in which speckles occur in an illumination device using excimer laser light or the like will be briefly explained with reference to FIG.

FIG. 3(A) shows the beam LB from a KrF (lipton fluoride) laser light source 100 expanded to a square cross section by a beam expander optical system lO2 including a cylindrical lens, and a fly-eye lens (the number of elements is about 10XIO). ) 104, and a large number (100) of divergent light from the secondary light source images SP (here, just spots) are formed on the exit surface of the fly-eye lens 104, and the condenser lens 106 superimposes the divergent light from the SP. A mask (reticle) or a photosensitive substrate is placed on the irradiated surface IP, which represents a system that uniformly irradiates the surface IP.

Generally, this type of laser light source can be broadly classified into three oscillation methods: a stable resonance method, an unstable resonance method, and an injection locking method. In these systems, the spatial coherency of the beam that is oscillated and output increases in order, and a simple stable resonance system with only a rear mirror and a front mirror has the lowest spatial coherency, and at the same time, the temporal coherency is also low, making it extremely multiplexed. mode.

The properties of the light emitted from such a multimode laser light source are similar to the spectrum from a known mercury lamp. Therefore, it is expected to be applied to semiconductor lithography. However, the beam from this type of stable resonance type laser light source places an extremely heavy burden on the projection lens used in projection exposure. That is, KrF
Since the laser beam is in the far ultraviolet region, only quartz can be used as a practical glass material for the projection lens, making it difficult to correct chromatic aberration. Therefore, if the oscillation spectrum has a wide range, the desired resolution cannot be obtained due to chromatic aberration. Therefore, inside the stable resonator, an etalon,
A method has been considered in which a wavelength selection element (spectroscopic element) such as a darating or a prism is provided to make the spectral width of the oscillated beam extremely narrow (eg, about 0.003 nm in full width at half maximum).

It was thought that by using a laser light source adopting this method, it would be possible to obtain intense ultraviolet pulsed light with high temporal coherency but extremely low spatial coherency. However, it has been confirmed that elements used to narrow the wavelength spectrum can increase spatial coherency.

Of course, other unstable resonance methods and injection locking methods have much higher spatial coherency.

Such a beam is a fly beam as shown in Figure 3(A).
When incident on the eye lens 104, if the pitch of the lens elements of the fly eye lens 104 is smaller than the coherence distance, for example, Mitsuoka light from adjacent secondary light source images SP interfere with each other. , noticeable interference fringes occur on the irradiated surface IP. These interference fringes may be only one-dimensional or may be two-dimensional. As shown in FIG. 3(B), the illuminance distribution on the irradiated surface IP is at a uniform level on average, but fine ripple-like fluctuations due to interference fringes are superimposed. The contrast intensity C of this interference fringe is much larger than the standard for uniformity of illuminance unevenness of ±1% required for this type of apparatus, and cannot be used as is for line width resolution in the sub-micron region.

Therefore, in the present invention, in order to quantitatively measure the contrast of interference fringes, a light-shielding pattern is inserted, for example, between the condenser optical system 106 and the irradiated surface IP.
A blurred edge image of the light-shielding pattern is created on P, and this is transferred onto a bare silicon wafer covered with a resist layer.
Shine. When the beam has no coherence, the light intensity distribution of an image with blurred edges is a straight line that drops to zero with a gentle slope. However, when the beam has coherence, the linear intensity distribution has interference fringes. The contrast becomes superimposed.

In the present invention, we focus on the resist image obtained when the blurred image is exposed and etched, and start from the point where the resist thickness begins to decrease, and from there to the point where the resist thickness first becomes zero ( In other words, the distance Xl to the point where the resist has disappeared) and the distance X2 from the starting point to the longest point where the resist remains are observed. and (X+ +Xt
) and (XtXI) as interference fringes (interference pattern)
It is now calculated as the contrast value of

〔Example〕

FIG. 1 is a flowchart schematically showing a method according to a first embodiment of the present invention, FIG. 4 is a perspective view showing a schematic configuration of a projection exposure apparatus (stepper) suitable for this embodiment, and FIG. FIG. 5 is a diagram illustrating the principle of the contrast intensity measuring method in this embodiment.

First, the configuration of the excimus chin bar will be explained using FIG. 4. In FIG. 4, the beam emitted from the excimer laser light source 10 enters the optical system 11 including a cylindrical lens via the ultraviolet reflecting mirrors MMt, Ms, and M4.
The beam LB having a rectangular cross-sectional shape is shaped into a substantially square beam LB. The beam LB is an ultraviolet reflecting mirror M.

The beam is bent and incident on a beam expander (or zoom lens) 15, where it is expanded to a predetermined cross-sectional size.

The parallel beam with an approximately square cross section emitted from the beam expander 15 passes through the fly's eye lens 3 and lens 4 in FIG.
The light is incident on the scanning mirror 17 via. The scanning mirror 17 is connected to a vibration source (deflection source) 19 such as a galvano, piezo, or torsional oscillator.

Next, the beam deflected by the scanning mirror 17 passes through the lens 21 and enters the second fly's eye lens 5.
After condensing as a large number of tertiary beams 1IIIil (spot beams), the beams diverge and are superimposed on the reticle blind 26 as an illumination field stop by the condenser lens 25A. The transmitted light of the blind 26 passes through the imaging lens 25A, is bent by the ultraviolet reflecting mirror 27, and enters the main condenser lens 2.

The light from each of the many tertiary light sources appropriately focused by the main condenser lens 2 is superimposed on the reticle R again and illuminates the reticle R with a uniform illuminance distribution. The reticle blind 26 has four movable sides and is arranged conjugately with the reticle R. As a result, the circuit pattern area on the reticle R is projected and exposed onto the wafer W by the projection lens PL, which is made of quartz and is telecentric on both sides.

A spot group of the tertiary light source of the second fly's eye lens 5 is imaged on the pupil (entrance pupil) ep of the projection lens PL, and the so-called Koehler illumination method is adopted.

Further, the vibration of the scanning mirror 17 is only required to be tilted by a certain angle during the oscillation of the number N of pulses of the excimer laser light necessary for the exposure of one shot on the wafer W. It is disclosed in JP-A-63-159837.

Due to this inclination of the scanning mirror 17, the interference fringes shown in FIG. 3(B) are slightly moved in the pitch direction of the fringes by the pitch/N pulse for each pulse emission, and the result is that the interference fringes shown in FIG. In this case, the interference fringes are averaged, and the unevenness in illuminance is suppressed to about ±1%.

Also, for alignment of reticle R with the device,
Marks RMx and RMy are formed around the reticle R, and in the space between the condenser lens 2 and the reticle R, reticle alignment optical systems 30x and 30y are arranged corresponding to the positions of the marks RMx and RMy. There is. A diagonal mirror 45' is located at the tip of the alignment optical system 30x, 30y, and an objective lens is placed in the optical path horizontally bent by this mirror, and these are fixed in a holding metal fitting. .

The alignment optical systems 30x, 30y may be configured to be movable in response to changes in mark position, as disclosed in, for example, Japanese Patent Application Laid-Open No. 57-142612.

On the other hand, the wafer W is placed on a stage STG that moves in a step-and-repeat manner. Stage STG BaX
, and move the wafer W by stepping in the Y direction, and also slightly move the wafer W in the Z direction for focusing. This fine movement stroke in the Z direction is around 1 m, but the depth of focus of the projection lens PL is about ±2 μm depending on the numerical aperture (N, A, Positioning accuracy of about ±0.5 to ±1 μm is required. Further, a reference slit mark plate FM is provided on the fine movement stage in the Z direction, and a photoelectric element for receiving light transmitted through the slint is embedded in the lower surface of this mark plate FM. The specific structure and usage of the mark plate FM and the photoelectric element are disclosed in detail in, for example, Japanese Patent Laid-Open No. 60-26343. The slit of the mark plate FM is, for example, elongated in the Y direction and has a width of about 1 to 2 μm. Furthermore, an off-axis alignment microscope 40 for observing marks on the wafer W is fixed to the stepper in FIG. 4 in close proximity to the projection lens PL. This alignment microscope 40 (hereinafter referred to as off-axis alignment system 40) uses a broadband wavelength light source such as a halogen lamp as a self-illumination system in order to observe minute marks on the wafer W as a high-resolution image. We are prepared. The broadband light has a wavelength distribution ranging from green to red, with only the wavelength band that exposes the resist layer being cut by a color filter. In addition, the color filter should be replaceable as appropriate.
It is preferable to be able to select the wavelength band of illumination light.

Inside the off-axis alignment system 40, a color imaging element such as a CCD or iTV that draws an image of the wafer surface obtained through a microscope objective lens is incorporated, and the image is displayed on a color monitor cathode ray tube (not shown). A color image of the wafer surface is displayed.

Next, referring also to FIG. 1, the procedure for measuring the contrast of an interference pattern will be described. In this embodiment, a test reticle R for testing resolution and the like is used at the same time, and the edges of the distal ends of the alignment systems 30x and 30y are used as shielding objects.

The metal tips of the alignment systems 30x and 30y are located approximately 10 to 30 mm above the position of the reticle R.
When the magnification of the projection lens PL is 115, the edge at its tip has a blur width of about 200 μm on the wafer surface. It is also assumed that the edge of the metal tip extends linearly in the same direction as the direction of interference fringes on the image plane (reticle plane). That is, the width direction of the image with blurred edges is made to coincide with the pitch direction of the interference fringes.

First, in step 50 in FIG. 1, the stepper is initialized, the test reticle R is set and aligned,
The opening of the reticle blind 26 is fully opened, the wafer W is carried in from the coater developer (C/D), and the stage S
Pre-alignment on the TG and setting of the dose amount to obtain an appropriate exposure amount (light tv4 adjustment of each pulse from the excimer laser light source 10, etc.) are performed.

The reason why the reticle blind 26 is fully opened is that the alignment system 30x, 3. This is because the edge of the metal tip of Oy is located, and the purpose is to widen the illumination field so that the edge of the tip is also irradiated with illumination light.

Further, during normal exposure, the scanning mirror 1 is
Initial settings are also required to synchronize the deflection angle of 7 and the pulse oscillation timing of the excimer laser, but in this example, the contrast of the interference pattern is measured with the scanning mirror 17 fixed, so synchronization is not necessary. First 3II for
I setting is not performed.

Next, in step 52, after focusing the projected image of the test reticle R on the wafer W (leveling adjustment as necessary), one area on the wafer W is exposed. In addition, at this time, the alignment system 30x, 30y
The shadow of the leading edge of the test reticle is set so that it passes exactly through the transparent area around the test reticle. In addition, in normal exposure, the minimum number of excimer pulses required is determined in order to reduce the visibility of interference fringes on the image plane due to the vibration of the scanning mirror 17, but in this exposure mode, such a limit is not imposed. Therefore, the appropriate exposure amount may be obtained with a smaller number of pulses.

Next, in step 54, it is determined whether or not to step the wafer W. If another area on the wafer W is to be similarly exposed, in step 56, the stage STG is
Alternatively, after stepping in the Y direction, the process returns to step 52. If only one shot is to be exposed on the wafer W, or if all the shots have been exposed, step 58
In step 60, whether or not to perform a developing process is selected based on the properties of the resist. Currently, the resists used in excimer steppers are mostly coater-developing bars (C
70) etc. is required.

However, in recent years, self-developing type resists have begun to be developed in which when exposed to ultraviolet light, only the exposed areas react with the ultraviolet light and are automatically removed. Therefore, in the case of a self-developing type resist, the resist image (etched resist N) is immediately measured without being sent to C70.

Next, in step 62, a choice is made between off-line measurement using another measuring device or on-line measurement using a stepper. In the case of off-line measurement, a combination of an optical microscope that self-illuminates with white light and a length measuring device (or scale plate) that can measure the distance between two points on the resist image on the order of microns is desirable.

In the case of on-line measurement, in step 64, the portion of the resist image with blurred edges is moved under the off-axis alignment system 40, and set to observation and measurement mode using the imaging element.

Then, in step 66, the positions of the feature points of the blurred resist image are actually measured, and the contrast of the interference pattern is determined from the positions. The specific measuring method will be explained using FIG. 5.

FIG. 5(A) shows the exposure dose distribution of a blurred edge image obtained on a wafer when illumination light does not generate interference fringes,
FIG. 5(B) shows the change in film thickness of the resist image after development obtained in the case of FIG. 5(A). The resist in FIG. 5(B) is a positive type.

FIG. 5(C) shows the exposure amount distribution of the blurred image when interference fringes (speckles) are superimposed, and FIG. 5(D) shows the film of the resist image obtained in the case of FIG. 5(C). Indicates thickness change.

In Figure 5, 10 is the set exposure amount (or appropriate exposure amount)
, It is the dark value exposure corresponding to the threshold value for whether or not the resist is completely peeled off, and X is the distance from the point where the resist thickness starts to decrease to the point where the resist is completely peeled off. be. As shown in FIG. 5A, when a straight edge of the light shield is located at a defocused position, the light amount distribution of the blurred image of the edge has a substantially constant slope in the direction perpendicular to the edge. When the intensity distribution of the interference fringes is superimposed on this, at the set exposure I+, the maximum exposure amount due to speckle contrast is 1 amx, approximately centered on that value.
The distribution fluctuates between the minimum (1 (I ml.). In Fig. 5 (C), I2 is the exposure amount at which the minimum point of the exposure amount due to the interference fringe contrast is the dark value exposure amount It; The maximum point of exposure due to interference fringe contrast is dark value exposure 1
It represents the exposure amount of 11t. When a positive resist is exposed under such a light amount distribution, as shown in FIG. 5(D), the resist remains +1! Starting from the point where the thickness begins to decrease, a point where the resist layer is completely removed for the first time appears at a distance of X1 from there in the pitch direction of the interference fringes (direction perpendicular to the straight edge), and then at a distance of Xg (X
The last remaining resist appears at t>XI).

Here, the contrast value C of the interference pattern is 1, □ +I a+++ as explained above, so the denominator and numerator are set to the exposure amount 1. When divided by , the following equation (1) is obtained.

As is clear from FIG. 5(C), the set exposure amount ■. When , the maximum point of contrast is I saw and the minimum point is I sin, so the minimum point of contrast when the exposure amount is 8 is the darkness value exposure amount! t, and the maximum point of contrast when the exposure amount is 1 corresponds to the dark value exposure amount 1t. Therefore, equation (1) can be replaced with the following equation (2).

Dividing the denominator and numerator of this equation (2) by It gives the following equation.

Here X: I: =Xt: Iz 7! Therefore, equation (3) can be replaced with equation (4).

Therefore, by using this equation (4), the contrast value C
It can be seen that is determined by the distance (width) of the characteristic points of the resist image on the wafer W. Although a positive resist was used here, measurements can be made using exactly the same method for a negative resist.

By using white illumination light or broadband wavelength illumination light, the characteristic points of the resist image (starting point, gap position, longest remaining position) can be easily distinguished by the difference in interference color caused by changes in the resist residual film thickness. can.

Therefore, two representative methods of determining the distances X, , X, using the off-axis system 40 of the stepper will be explained, but first, a method of determining using the number of pixels in the horizontal direction of the image sensor will be described.

The resist image shown in FIG. 5(D) is observed on a color cathode ray tube with the horizontal scanning line of the image sensor parallel to the pitch direction of the interference fringes. At this time, on the cathode ray tube, the color appears to be constant from the starting point to the left side (the part where the remaining resist film thickness is constant), and the color changes sequentially from the starting point to the resist missing position, and at the resist missing position, it becomes a thin band-like color that appears on the wafer. The color of the base begins to appear, and to the right of the longest remaining position, no interference color due to the resist is observed, and only the color of the wafer base appears. Therefore, two cursor lines are displayed vertically on the cathode ray tube via an image signal processing circuit, and one of the cursor lines is moved horizontally by operating a joystick or the like to align it with the starting point. Then, align the other cursor line with the resist gap position. And the distance between these two cursor lines (i.e. the number of pixels in the horizontal direction)
is read as the difference between the setting values of the position setting counter for each cursor line and stored as distance , the zero order is stored as the distance Xt, and the calculation of equation (4) is executed to calculate the contrast value C (0≦C≦1). This result is displayed as a numerical value on a part of the screen of a cathode ray tube, for example.

Another way to find it is to use an index mark provided in the off-axis alignment system 40 or a single cursor line. First, the stage STG is manually positioned so that the index mark (or cursor line) matches the starting point position on the monitor screen, and this position is read from the length measuring device (laser interferometer) of the stage STG.
D. Move the 0-order stage STG horizontally on the monitor screen to match the index mark (cursor line) with the resist missing position, and read the position D1. Furthermore, the stage STG is moved to match the index mark (cursor line) with the longest remaining position, and the position is read. And,X,=D,−D, ,X,=
After calculating D and -D, the contrast value C is determined using equation (4).

It is sufficient if the contrast measurement described above is performed for only one shot on the wafer W, but it is also possible to calculate the average of the contrast values C measured for each of a plurality of shots (for example, 3 to 5). .

As described above, according to this embodiment, the speckle contrast can be measured regardless of the exposure control accuracy. That is, in the conventional method shown in FIG. 2, the speckle contrast C is determined from Ia-1b C=Ia+Ib. This is determined by the exposure amount of two shots, and the deviation from the set exposure amount of Ia and Ib directly appears as an error.

That is, it becomes impossible to measure speckle contrast smaller than the accuracy of the set exposure amount. In this example, the formula (
3) corresponds to the same formula as the conventional method, but the exposure amount is 1t,
I+ is defined during one shot exposure,
The error from the set exposure amount I0 can be canceled out.

That is, in equation (4), even if X2 changes due to an error in the set exposure amount I0, Xl moves in the direction of canceling it out, and accurate speckle contrast can be measured.

In addition, changes in the dark value exposure amount It due to uneven coating (thickness) of the resist also occur due to minute positional differences on the wafer W (200°
(on the order of μm), so it has almost no effect in practice. For example, when the beam is expanded by adjusting the zoom ratio of the beam expander 15 in FIG. becomes stronger.

In this embodiment, the speckle contrast is measured by burning one shot, so even if the exposure amount is controlled by expanding the beam cross section using the beam expander 15 (adjusting the peak level of the pulsed light), one shot is enough to measure the speckle contrast. It is possible to accurately measure the speckle contrast of In the conventional method, if the spare contrast C changes depending on the set exposure amount, an error occurs. Furthermore, even if a light-shielding pattern is engraved on the reticle R as described in this embodiment, if there is a glass part on the outer periphery of the reticle R that is not light-shielded, a blurred image can be created through it and the spec. Since it is possible to measure the reticle contrast, it is possible to print the blurred image and the reticle pattern at the same time using the appropriate exposure amount for the resist, and to accurately understand the correlation between the influence of speckle on the reticle pattern and the speckle contrast. It is.

This method can be similarly implemented with an actual reticle for device manufacturing, and it is possible to more quantitatively grasp resolution failures due to speckles in fine patterns on actual devices.

In the above embodiment, the reticle R is also exposed at the same time, but if only the contrast value C is to be measured, the reticle R may be removed.

Next, a second embodiment will be described with reference to FIGS. 6, 7, and 8.

FIG. 6 shows the configuration between the condenser lens 2 and the projection lens PL, and shows the arrangement of a plate 31 or 32, such as a brass plate or an acrylic plate with high ultraviolet absorption rate, as a light-shielding object. The plate 31 is located approximately 20 mm above the reticle R, and the maximum range of the pattern area of the reticle R is P, , P,
is fixedly installed on the outside of the In FIG. 6, ILb represents the imaging light beam from the edge of the reticle blind 26, and R5 is the reticle stage. In addition, in the case of the plate 32, between the reticle R and the projection lens PL,
They are placed about 0 mm apart, and a similar defocused image can be obtained. In order to project the images of these plates 31 and 32,
The position of the opening edge of the blind 26 may be expanded to the outside of P+ and Pt.

FIG. 7 is a plan view in which plates 31 are provided at four locations above the reticle R, and plates 31a, 31b, 31c, and 31d having straight edges are fixedly provided at each position on the four sides of the reticle R. In FIG. 7, a circular area (2) F represents the maximum field of view of the projection lens PL, and a light-shielding band SB of a constant width is formed on the reticle R so as to fit within this area. The outside of the light-shielding band SB is transparent, and the opening edges on the four sides of the reticle blind 26 are imaged in accordance with the light-shielding band SB.

For example, if a stepper has a projection field of 21.2 an φ on the wafer W (the diameter of !
On the reticle, the pattern area is 75II, 11 squares), so there is +
The edges of L31a, 31b, 31c, 31d can be fixedly arranged. In addition, there are cases where a rectangular circuit pattern area is set instead of a 15-angle circuit pattern area, but in this case, the edges of the tips of the plates 31a, 31b, 31c, and 31d are made as short as possible, and the edges are set as close to the outer periphery within the viewing area IF. There is no problem if you place it in

Note that the distance of the plates 31 and 32 from the reticle R is set at 20 m+, but this is because the distance Hxt shown in FIG. It is sufficient to set the value to a level that can be recognized, and it is not a particularly limited value. Also, during actual exposure of a device, etc., the reticle blind 26 is placed at the position P of the light shielding bands S and B.
Since it is closed up to i and P, the existence of plates 31 and 32 does not pose any problem. Therefore, when it is desired to measure the contrast value, trial printing can be performed freely by simply opening and closing the blind 26.

Also, when measuring speckle contrast depending on location within one shot, use a fully transparent reticle (full glass reticle) or do not insert a reticle.
All you have to do is put a rough grid pattern on the entire surface 20 degrees above the reticle position.

In each of the above embodiments, it has been described that a blurred image is created on the reticle side, but as shown in FIG. Effects can be obtained. In that case,
The reticle pattern does not allow best focus exposure, but instead the reticle pattern itself functions as a shield in the present invention and the speckle contrast can be measured in the same way, so the speckle contrast of the entire surface (any point within the shot) can be measured. can be easily determined.

Further, in this embodiment, a projection exposure apparatus has been described, but this method can be applied to all illumination optical apparatuses (proximity exposure apparatus, etc.) that use highly coherent light.

In a proximity exposure system, the mask and wafer are faced to each other with a fixed gap (about 10 to 300 μm).
Similar measurements can be made by arranging the shield with a gap larger than the originally set gap.

Note that with an illumination system that combines a fly-eye lens and a scanning mirror, the interference fringes can be accurately smoothed by exposing a defocused image in a mode in which the scanning mirror is oscillated to smooth the interference fringes. You can check whether it is. In this case, if smoothed correctly, X, ζ
It becomes Xt.

Next, a method according to the third embodiment will be explained. Here, the slit of the mark plate FM shown in FIG. Let us describe the case of direct measurement. In this case, Figure 5 (C)
The maximum exposure point shown in 1. □ and the minimum point! 1.0 may be detected by a photoelectric element under the mark plate FM. Therefore, it is necessary that the slit width of the mark plate 1'M is sufficiently smaller than the pitch of the interference fringes. Furthermore, instead of scanning the mark plate FM (moving the stage STG), the scanning mirror 17 may be swung to move the interference fringes themselves relative to the slit.

When using the mark plate FM, the emission trigger of the excimer laser light source 10 is made to respond to a measurement pulse for each unit movement from a laser interference needle for position measurement of the stage STG, or to measure an angle for each unit deflection angle of the scanning mirror 17. All you have to do is respond to the pulse.

Furthermore, if the difference between the maximum point (2) and the minimum point 1+si becomes small, the desired measurement accuracy may not be obtained depending on the linearity characteristics, S/N ratio, etc. of the photoelectric element. In that case, the beam expander 15 shown in FIG. 4 expands the cross-sectional dimension of the beam by a predetermined zoom ratio to increase the spatial coherency of the beam incident on the fly-eye lens 3. . In this way, although the overall illuminance at the image plane decreases, the contrast value of the interference fringes, that is, 1. □-! , 8. Since the value of increases in accordance with the zoom ratio, it becomes easier to maintain measurement accuracy.

Then, by correcting the actually measured contrast value according to the zoom ratio, the original contrast value is obtained. In addition, a glass plate (quartz) is installed at the mark plate FM, and an objective lens (quartz) with several times magnification is embedded beneath it.
The image of the interference fringes formed on the glass plate may be received by an imaging element such as a linear array through an objective lens to determine the contrast. At this time, the beam expander 15
In this embodiment, there is no need to provide a shield in the illumination optical path.

In addition to the embodiments described above, a shield may be made to protrude near the reticle blind 26 shown in FIG. 4. At this time, when the magnification from the reticle blind 26 to the reticle R is, for example, 2, the distance from the reticle blind 26 to the shield in the optical axis AX direction may be about 5 mm. Furthermore, when the reticle blind 26 is movable in the direction of the optical axis AX, a defocused image can also be obtained by using the opening edge of the blind itself as a shield.

Also, the stepper alignment system 30x shown in Figure 4
, 30y are movable to arbitrary positions on the reticle R, the contrast value can be measured at any arbitrary point within one shot. Further, in each of the embodiments of the present invention, one-dimensional (or two-dimensional) interference fringes are mainly treated as speckle patterns, but the contrast of interference patterns with little periodicity can also be measured in the same way.

When the contrast of the interference pattern (ic) is determined quantitatively and accurately, based on this contrast value C, the fourth
The optimal way of swinging the scanning mirror 17 shown in the figure and the optimal (minimum necessary) number of pulses can be set, and a more efficient exposure sequence can be constructed.

〔Effect of the invention〕

As described above, according to the present invention, it is possible to accurately measure speckle contrast without being affected by exposure control accuracy or uneven resist coating Lf, and it is also possible to accurately measure speckle contrast with one shot. Therefore, it is possible to simultaneously verify the difference in speckle contrast due to exposure control and the correlation between speckle contrast and a pattern printed with proper exposure. Furthermore, since the speckle contrast can be grasped quantitatively, it is possible to more accurately control mirror scanning and the like for eliminating speckles.

[Brief explanation of drawings]

FIG. 1 is a flowchart explaining the procedure of the method according to the first embodiment of the present invention, and FIG. 2 (A), (B), and (C)
, (D) is a diagram explaining the principle of a conventional contrast measurement method, FIG. 3 (A) is a diagram showing a schematic configuration of an illumination optical system using a coherent beam, Figure (B) is a diagram showing an example of the illuminance distribution obtained by the optical system of Figure 3 (A), Figure 4 is a perspective view showing the configuration of a stepper suitable for carrying out the method according to the present invention, and Figure 5 Figures (A), (B), (C), and (D) are diagrams explaining the contrast measurement method according to the first embodiment; Figure 6 is a diagram showing a partial configuration of a stepper according to another embodiment; FIG. 7 is a plan view showing the arrangement of a reticle and a shield, and FIG. 8 is a diagram illustrating an exposure method according to another embodiment. [Explanation of symbols of main parts] R... Reticle, PL... Projection lens, W... Wafer, 2... Condenser lens, 3.5... Fly's eye lens, 10... Laser light source , 11... Beam shaping optical system, 15... Beam expander 17... Scanning mirror 26... Reticle blind, 30x, 30y... Alignment system, 31.31a,
31b, 31c, 31d, 32... Light shielding plates.

Claims (4)

[Claims] (1) The irradiated surface is illuminated by a beam irradiation device equipped with a coherent beam generation source and an illuminance distribution uniformizing means that shapes the beam into a substantially uniform illuminance distribution on a predetermined irradiated surface. In the method of measuring the contrast intensity of an interference pattern generated on the irradiated surface when irradiated, a substrate coated with a resist layer sensitive to the beam to a predetermined thickness is placed on the irradiated surface, and the arranging a shielding object between the irradiation surface and the source in a non-conjugate relationship with the irradiated surface; irradiating the beam to form a blurred image of the edge of the shielding object on the resist layer; a step of etching a blurred image into the resist layer; and a step of etching a blurred image into the resist layer;
A distance X_1 from the starting point where the thickness of the resist layer starts to decrease to a point where the thickness of the resist layer first becomes almost zero, and a distance X_2 from the starting point to the approximately longest point where the resist layer remains. and (X_1+X_2) based on the distances X_1 and X_2.
) and (X_2-X_1) as the intensity of the contrast. (2) The beam irradiation device is an exposure device that exposes a pattern of a mask onto a photosensitive substrate, and the shielding object is arranged at a certain distance in the optical axis direction from a set position of the mask. The method described in Section 1. (3) The exposure apparatus uses a rare gas halide laser source as the generation source, and the illuminance distribution uniformizing means receives a laser beam from the laser source to form a plurality of secondary light source images. a light flux forming optical system and a condenser optical system that superimposes light fluxes from each of the secondary light source images on the mask, and interference fringes are generated on the mask due to interference of light from each of the secondary light source images. 3. A method according to claim 2, characterized in that the contrast intensity of the image is measured. (4) The exposure device is either a projection exposure device that images the pattern of the mask onto the photosensitive substrate via a projection optical system, or a proximity exposure device that separates the mask and the photosensitive substrate by a minute gap. Claim 2 or 3, wherein a part of the alignment system holding hardware disposed between the mask and the illuminance distribution uniformizing means is used as the shielding object. The method described in section.