JP2007201298A - Focus measurement method, exposure system and mask for focus measurement - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To measure focus information related to a projection optics system through the use of simply configured marks. <P>SOLUTION: A focus measurement method of measuring projection optics system focus information comprises the steps of: using illumination light, in which principal ray is inclined relative to an optical axis AX of the projection optics system PL, to illuminate the marks 50A, 50B such that each center of barycenter of two diffraction lights from the marks 50A, 50B deviates in the X direction on the pupil surface PLP of the projection optics PL, and thereby to project images of marks 50A, 50B disposed by projection optics PL on a wafer W; obtaining the spacing in the X direction of the projected marks 50A, 50B; and also obtaining defocus amount at the locations where the mark images are projected based on the obtained spacing. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a focus measurement technique and an exposure technique of a projection optical system, for example, for transferring a pattern onto a substrate in a lithography process for manufacturing various devices such as a semiconductor integrated circuit, a liquid crystal display element, or a thin film magnetic head. This is applicable when measuring the imaging state of the projection optical system of the exposure apparatus used.

  For example, to transfer a pattern formed on a mask onto a wafer (or glass plate or the like) coated with a resist as a photosensitive substrate through a projection optical system in a lithography process for manufacturing a semiconductor integrated circuit. In the projection exposure apparatus to be used, with the further miniaturization of the integrated circuit, the demand for the resolution of the projection optical system is increasing.

In general, when the resolution of the projection optical system increases, the depth of focus becomes shallower and the allowable defocus amount decreases. The defocus amount also varies depending on imaging characteristics such as the curvature of field of the projection optical system, autofocus performance of the exposure apparatus, and flatness corresponding to the uneven distribution on the wafer surface. Therefore, specially asymmetric diffraction grating marks having different diffraction efficiencies of the + 1st order and -1st order diffracted light have been conventionally used in order to evaluate the imaging characteristics of the projection optical system and the focus information such as the flatness of the wafer in advance. And illuminating a reference mark composed of a light shielding pattern with illumination light having a symmetric light distribution around the optical axis of the illumination optical system, and an image of the asymmetric diffraction grating mark and the light shielding pattern on a resist-coated wafer. The shape of the resist image is measured by projection, and the defocus amount is obtained from the measurement result (see, for example, Patent Document 1).
Japanese Patent No. 3297423

As described above, in the conventional defocus amount measurement method using a special asymmetric diffraction grating mark, it is difficult to form a mark for setting the ratio of the diffraction efficiency of + 1st order and −1st order diffracted light to a predetermined value. Therefore, there is a disadvantage that the manufacturing cost of the evaluation mask is increased.
In addition, the defocus amount may be measured at a plurality of measurement points in the field of view (exposure region) of the projection optical system. For this purpose, the defocus amount is measured at positions on the evaluation mask corresponding to the plurality of measurement points. It is necessary to form a special mark for each. However, it is more difficult to form a plurality of special marks with a defined diffraction efficiency ratio.

SUMMARY OF THE INVENTION In view of the above, an object of the present invention is to provide a focus measurement technique and an exposure technique that can measure focus information of a projection optical system using a mark having a simple configuration.
A further object of the present invention is to provide a focus measurement mask having a simple configuration that can be used when measuring focus information of a projection optical system.

  A focus measurement method according to the present invention is a focus measurement method for measuring focus information of a projection optical system (PL), and the projection is performed using illumination light whose principal ray is inclined with respect to the optical axis of the projection optical system. A first step of illuminating a first mark (50A) that generates two diffracted lights so that a light intensity distribution in which the distance from the optical axis on the pupil plane is different from each other is obtained on the pupil plane of the optical system; Projecting the image of the first mark on a predetermined surface via the projection optical system, the second step of obtaining the position information of the image of the first mark, and the projection based on the position information obtained in the second step And a third step for obtaining focus information of the optical system.

According to the present invention, an image formed by passing two diffracted lights from the first mark through the projection optical system is laterally shifted according to the defocus amount of the projection surface. Focus information can be obtained from quantity information. At this time, a simple mark such as a periodic mark having a predetermined pitch can be used as the first mark.
Next, the exposure apparatus according to the present invention illuminates the first mark (50A) arranged on the object plane side of the projection optical system in the exposure apparatus that transfers the pattern onto the substrate via the projection optical system (PL). The illumination optical system (5) and the first diffracted light generated from the first mark so as to obtain a light intensity distribution such that the distance from the optical axis on the pupil plane of the projection optical system is different from each other. And a control device (1, 11, 12) for controlling the inclination angle of the principal ray of the illumination light for illuminating the mark. According to the present invention, the focus measurement method of the present invention can be used.

  In addition, the focus measurement mask (TR) according to the present invention is a mask that is illuminated under illumination conditions in which the chief ray of illumination light is tilted, and the diffracted light is shifted in different ways and each has a predetermined periodic direction. A symmetrical first mark (50A) and second mark (50B) are formed side by side. According to the present invention, the focus control method of the present invention can be used.

  In addition, although the code | symbol with the parenthesis attached | subjected to the predetermined element of the above this invention respond | corresponds to the member in drawing which shows one Embodiment of this invention, each code | symbol is used for easy understanding of this invention. However, the present invention is not limited to the configuration of the embodiment.

Hereinafter, an example of a preferred embodiment of the present invention will be described with reference to FIGS.
FIG. 1 shows a schematic configuration of a projection exposure apparatus comprising a scanning stepper of this example. In FIG. 1, an ArF excimer laser light source (wavelength 193 nm) is used as the exposure light source 6. As an exposure light source, a KrF excimer laser light source (wavelength 247 nm), an F ₂ laser light source (wavelength 157 nm), a harmonic generator of a solid-state laser (semiconductor laser or the like), a mercury lamp, or the like can also be used.

  Exposure illumination light (exposure beam) IL emitted from the exposure light source 6 at the time of exposure has a cross-sectional shape through a mirror 7, a beam shaping optical system (not shown), a first lens 8A, a mirror 9, and a second lens 8B. After being shaped into a predetermined shape, it is incident on the fly-eye lens 10 as an optical integrator (a homogenizer or a homogenizer), and the illuminance distribution is made uniform. On the exit surface of the fly-eye lens 10 (the pupil plane of the illumination optical system), an aperture stop 13A for dipole illumination (dipole illumination) that distributes the light amount distribution of illumination light in two circular regions symmetrical to the optical axis. An aperture stop 13B and a ring for tilting the principal ray of illumination light on the object plane at a predetermined angle with respect to the optical axis (that is, breaking the telecentricity), which is composed of one circular area away from the optical axis. An illumination system aperture stop member 11 having a band illumination aperture stop 13C, a normal illumination aperture stop 13D having a circular aperture, and the like is rotatably arranged by a drive motor 12. Hereinafter, losing telecentricity is referred to as telecentric collapse. The main control system 1 that controls the operation of the entire apparatus controls the operation of the drive motor 12, and the main control system 1, the illumination system aperture stop member 11, and the drive motor 12 control the tilt angle of the principal ray of the illumination light IL. A control device is configured.

  The illumination light IL that has passed through a predetermined aperture stop in the illumination system aperture stop member 11 passes through a beam splitter 14 and a relay lens 17A having a low reflectivity, and then a fixed blind 18A as a fixed field stop and a movable blind as a movable field stop. Pass through 18B sequentially. The illumination light IL that has passed through the blinds 18A and 18B passes through the sub-condenser lens 17B, the optical path bending mirror 19, and the main condenser lens 20, and the illumination area 21R of the pattern area of the reticle R as a mask has a uniform illuminance distribution. Illuminate.

  On the other hand, the illumination light reflected by the beam splitter 14 is received by the integrator sensor 16 including a photoelectric sensor via the condenser lens 15. The detection signal is supplied to the exposure amount control system 3, and the exposure amount control system 3 uses the detection signal and the transmittance of the optical system from the beam splitter 14 measured in advance to the wafer W as an object. The exposure energy on W is indirectly calculated. The exposure amount control system 3 controls the exposure light source 6 so that an appropriate exposure amount is obtained on the wafer W based on the integrated value of the calculation result and the control information from the main control system 1 that controls the overall operation of the apparatus. Controls the light emission operation. The illumination optical system 5 includes the optical members from the mirrors 7 and 9 and the lenses 8A and 8B to the main condenser lens 20.

  Under the illumination light IL, the pattern in the illumination area 21R of the reticle R is coated with a resist at a projection magnification β (β is 1/4, 1/5, etc.) via a bilateral telecentric projection optical system PL. And projected onto a slit-like exposure area 21W elongated in the non-scanning direction on one shot area SA on the wafer W. Hereinafter, in FIG. 1, the Z axis is taken in parallel to the optical axis AX of the projection optical system PL, and X in the non-scanning direction orthogonal to the scanning direction of the reticle R and the wafer W during scanning exposure in a plane perpendicular to the Z axis. A description will be given by taking the axis and taking the Y axis in the scanning direction.

  First, the reticle R is held on the reticle stage 22, and the reticle stage 22 moves on the reticle base 23 at a constant speed in the Y direction, and finely moves in the X, Y, and rotational directions so as to correct the synchronization error. Then, the reticle R is scanned. The position of the reticle stage 22 is measured by a movable mirror (not shown) and a laser interferometer (not shown) provided on the reticle stage 22, and based on the measured value and control information from the main control system 1, a stage drive system 2 controls the position and speed of the reticle stage 22 via a drive mechanism (not shown) such as a linear motor. A reticle alignment microscope (not shown) for reticle alignment is disposed above the periphery of the reticle R.

  On the other hand, the wafer W is held on the wafer stage 28 via the wafer holder 24. The wafer stage 28 moves on the wafer base 27 at a constant speed in the Y direction, and moves in steps in the X and Y directions. And a Z tilt stage 25. The Z tilt stage 25 performs focusing of the wafer W based on a measurement value of a position in the Z direction of the wafer W by an auto focus sensor (not shown). The position of the wafer stage 28 in the XY plane and the rotation angles around the X, Y, and Z axes are measured by a laser interferometer (not shown), and the measured values and control information from the main control system 1 are used. Based on this, the stage drive system 2 controls the operation of the wafer stage 28 via a drive mechanism (such as a linear motor) (not shown).

  An aerial image measurement system 29 is fixed in the vicinity of the wafer holder 24 on the wafer stage 28, and the illumination light IL is applied to a surface set at substantially the same height as the surface of the wafer W on the aerial image measurement system 29. An elongated slit 30 is formed in the Y direction for passing through. A photoelectric sensor that condenses and photoelectrically converts the illumination light that has passed through the slit 30 is installed inside the aerial image measurement system 29, and a detection signal of the photoelectric sensor is supplied to a signal processing unit in the main control system 1. ing. The aerial image measurement system 29 can be used to measure the best focus position of the projection optical system PL at a predetermined measurement point in the exposure area 21W (field of view) of the projection optical system PL.

  Specifically, for example, a test reticle on which an evaluation mark made of a line and space pattern (hereinafter referred to as an L & S pattern) having a predetermined pitch in the X direction is loaded on the reticle stage 22 and the evaluation is performed. The mark image is scanned in the X direction by the slit 30 of the aerial image measurement system 29, the detection signal of the photoelectric sensor is taken in correspondence with the coordinate in the X direction of the wafer stage 28, and the contrast of the detection signal is obtained. Then, while changing the Z position of the slit 30 (which can be measured by the autofocus sensor), the operation is repeated and the Z position when the contrast of the detection signal is maximized is obtained. This is the best focus position at the projected measurement point.

Further, an off-axis type alignment sensor 32 for wafer alignment is disposed above the wafer stage 28, and the main control system 1 aligns the wafer W based on the detection result.
At the time of exposure, the reticle stage 22 and the wafer stage 28 are driven so that the reticle R and one shot area on the wafer W are synchronously scanned in the Y direction while being irradiated with the illumination light IL, and the wafer stage 28 is driven. Then, the step of moving the wafer W stepwise in the X and Y directions is repeated. Thereby, the pattern image of the reticle R is exposed to each shot area on the wafer W by the step-and-scan method.

  Next, the focus information related to the projection optical system PL of this example includes imaging characteristics such as field curvature and astigmatism of the projection optical system PL, and the flatness of the wafer to be exposed. In this case, the curvature of field of the projection optical system PL can be obtained, for example, by obtaining the best focus position at a plurality of measurement points in the exposure region 21W using the aerial image measurement system 29 described above. However, since the flatness of the wafer cannot be measured by the method using the aerial image measurement system 29, in the following, the defocus amount with respect to the projection optical system PL is measured at a predetermined measurement position in the exposure area 21W using a predetermined mark. Thus, an example of a method for measuring the flatness of the wafer will be described. By this measurement, for example, a plurality of lots of wafers can be divided into a plurality of ranks according to whether the flatness is good or not, and the wafers can be selectively used according to the application. Further, by using a reference wafer that is known to have good flatness in advance, the defocus amount is measured at a plurality of measurement positions in the exposure area 21W, thereby measuring the curvature of field of the projection optical system PL. It is also possible to do.

  2A shows a state in which the test reticle TR used as a focus measurement mask in this example is placed on the reticle stage 22 of FIG. 1, and in FIG. 2A, the pattern region of the test reticle TR is shown. 2B is representative of five mark formation regions 52A, 52B, 52C, 52D, and 52E to which a light shielding film (for example, a metal film of chromium or the like) at the center and four corners in 51 is deposited. As shown by the enlarged mark formation region 52A, the first mark 50A and the second mark 50B are arranged along the X direction. The first mark 50A has a plurality of opening patterns arranged at a predetermined pitch along the X direction, and the second mark 50B has a pitch different from the pitch of the first mark 50A along the X direction. It has a plurality of opening patterns arranged. Note that the pitch of the opening pattern of the first mark 50A and the opening pattern of the second mark 50B may be the same. That is, the marks 50A and 50B are marks obtained by patterning a light shielding film such as chromium. The pattern area 51 of the test reticle TR is substantially the same size as the illumination area 21R of FIG. 1, and the center of the mark formation area 52A at the center is located on the optical axis AX of the projection optical system PL. Note that only the central mark formation region 52A may be formed in the pattern region 51, or an arbitrary number of mark formation regions of two or more may be formed. Further, the pitch of the marks 50A and 50B is about 100 nm to several hundred nm at the stage of the projected image by the projection optical system PL.

  In this example, an unexposed wafer (wafer W) coated with a resist (photosensitive material) is loaded on the wafer stage 28 in FIG. 1 to measure flatness, and the wafer stage 28 is moved in the X direction. While stepping in the Y direction, as shown in FIG. 2 (C), a large number of shot areas SAi (i = 1, 2,..., I) on the wafer W are sequentially applied in a stationary exposure manner to FIG. The image 51P of the pattern region 51 of the test reticle TR is exposed (first step). The image 51P includes the images 52AP to 52EP of the mark formation regions 52A to 52E in FIG. Although the projection optical system PL of FIG. 1 actually projects a reverse image, the following description will be made assuming that an erect image is projected for convenience of explanation.

  FIG. 2D is an enlarged view representatively showing an image 52AP among the images 52AP to 52EP of the mark formation regions 52A to 52E of FIG. 2A. In FIG. 2B, the image 50AP of the first mark 50A and the image 50BP of the second mark 50B are included. In this example, the distance DXA between the center of the image 50AP and the center of the image 50BP in the X direction (the periodic direction of the images 50AP and 50BP) is measured (second step). As will be described later, in the first step described above, in this example, the principal ray from the illumination optical system 5 in FIG. 1 is used with the illumination light inclined with respect to the optical axis AX of the projection optical system PL, as shown in FIG. The first mark 50A and the second mark 50B are illuminated. With this illumination, two diffracted lights are generated from the first mark 50A. The two diffracted lights pass through different positions in the pupil plane of the projection optical system PL from the optical axis AX to the X direction. That is, light intensity distributions are formed in different regions from the optical axis AX to the X direction on the pupil plane of the projection optical system PL.

Also, this illumination generates two diffracted lights from the second mark 50B. The two diffracted lights pass through a position shifted in the X direction with respect to the position through which the two diffracted lights from the first mark 50A pass. In other words, a light intensity distribution is formed in a region different from the light intensity distribution by the two diffracted lights from the first mark 50A in the X direction from the optical axis AX on the pupil plane of the projection optical system PL. In the following, the position of each diffracted light on the pupil plane of the projection optical system PL means the position of the center of gravity of the light intensity distribution of each diffracted light.
Accordingly, the first mark image 50AP and the second mark image 50BP in FIG. 2D have the imaging positions in the X direction (period direction) in accordance with the defocus amount from the focal position of the projection optical system PL. Since the lateral shift amount (shift amount) differs, the relationship between the defocus amount and the lateral shift amount is obtained in advance by simulation or actual measurement, and the distance DXA in the X direction between the image 50AP and the image 50BP is measured. The defocus amount at the measurement point on the wafer W on which these images are projected can be obtained (third step).

  Further, in order to obtain the reference value of the distance DXA, the aperture stop 13A for dipole illumination may be used in the illumination system aperture stop member 11 of the illumination optical system 5 in FIG. When the aperture stop 13A is used, the two marks 50A and 50B in FIG. 2A are illuminated with illumination light symmetric in the X direction without being telecentric. Therefore, the two marks 50A and 50B in FIG. The distance DXA between the mark images 50AP and 50BP has the same value as the distance at the best focus position, and can be used as a reference value for the measurement value when telecentric collapse is performed. However, when the influence of coma aberration cannot be ignored, the method described later may be used.

  Then, by obtaining the defocus amounts in the images 52AP to 52EP (measurement points) of all the mark forming areas 52A to 52E in all the shot areas SA1 to SAI of the wafer W in FIG. The flatness (local flatness) that is the distribution state of the unevenness amount on the entire surface can be obtained. In this example, as an example, after exposure of the entire surface of the wafer W, the resist is developed, and the distance DXA between the image 50AP and the image 50BP in FIG. For example, the measurement is performed by a mark measuring device including a registration measuring machine (overlay measuring machine) provided with an optical microscope type sensor and a laser interferometer type coordinate measuring system. The interval DXA may be measured at the latent image stage without developing the resist.

However, a scanning electron microscope may be used as the mark measuring device, and the developed wafer is returned to the wafer stage 28 of FIG. It is also possible to measure the distance DXA between the image 50AP and the image 50BP in FIG.
Further, the shape of the image may change due to the influence of asymmetry at the ends in the X direction of the image 50AP and the image 50BP in FIG. Therefore, in order to eliminate the influence of the asymmetry, as described later, for example, an image of a mark (hereinafter also referred to as a trim mark) in which only the central portion of the marks 50A and 50B in FIG. The image at the end in the X direction of the image 50AP and the image 50BP in FIG. 2D may be erased by double exposure. For this purpose, as an example, a positive resist may be used as the resist on the wafer W. Such exposure for erasing the end portion or part of the image of the mark image can also be called trim exposure. In addition, it is preferable on a measurement that the mark after trim exposure shall be about 6 micrometers pitch. In addition, when the line width of the image of each opening pattern constituting the image 50AP and the image 50BP is too fine and the measurement accuracy is lowered depending on the sensor, the image 50AP and the image 50BP (or after trim exposure) The image) may be regarded as an image of one opening pattern as a whole, and the position in the X direction (period direction) may be measured.

Next, diffracted light generated from the marks 50A and 50B in FIG. 2B will be described.
FIG. 3A shows diffracted light generated from the second mark 50B of FIG. 2A. In FIG. 3A, the second mark 50B is in the X direction of the aerial image by defocusing. This is a mark in which the amount of shift is substantially zero. In this case, the illumination light IL from the illumination optical system 5 in FIG. 1 is a test reticle in a state where the principal ray is inclined by an angle θi in the X direction with respect to the optical axis AX of the projection optical system PL, that is, in a telecentric state. The TR is irradiated. Hereinafter, the angle θi is also referred to as telecentric collapse amount. In order to perform telecentric collapse of the illumination light IL in this way, an aperture stop 13B having one opening may be selected in the illumination system aperture stop member 11 of FIG.

The coherence factor (σ) of the illumination light IL of the illumination optical system 5 is preferably as small as possible, for example, 0.3 or less in terms of the numerical aperture iNA of the illumination optical system 5. The relationship between the coherence factor (σ) and the numerical aperture iNA of the illumination optical system 5 is as follows when the numerical aperture of the projection optical system PL is NA.
iNA = NA · σ
Further, the pitch P1 is determined from the equation (1) by using the telecentric collapse amount θi, where P1 is the pitch of the mark 50B and λ is the wavelength of the illumination light IL. It should be noted that the following mark pitch values (mark pitches) indicate values in a projection image by the projection optical system PL. At this time, the exit angle θ1 of the 0th-order light D0 from the mark 50B in FIG. 3A is θi. The telecentric collapse amount θi is preferably as large as possible within the range satisfying the expression (2) because the lateral shift amount at the time of defocusing of the first mark 50A described later becomes large.

λ / P1 = 2sin θi (1)
sin θi + iNA ≦ NA (2)
By satisfying the expression (1), in FIG. 3A, the first-order diffracted light D1 from the mark 50B is emitted symmetrically with the 0th-order light D0 with respect to the optical axis AX at the emission angle θ1, and therefore the 0th-order light D0. The mark image formed on the wafer W by the first-order diffracted light D1 via the projection optical system PL does not shift laterally in the X direction even when the wafer W is defocused.

  The simulation results at the best focus position and 100 nm defocused position of the light quantity distribution of the aerial image of the second mark 50B under the condition satisfying the expression (1) are shown by solid lines in FIGS. 5A and 5B, respectively. 5A and 5B, the horizontal axis represents the X coordinate (μm) of the aerial image, and the vertical axis represents the light amount (relative value) I of the aerial image. This is the same in FIG. The simulation conditions are P1 = λ / (2 sin θi) = 121 (nm), the number of aperture patterns is 24, NA = 0.92, λ = 193 (nm), iNA = 0.125, sin θi = NA−iNA , And. In FIGS. 5A and 5B, the dotted line indicates the ideal light amount distribution of the second mark 50B at the best focus position. The simulation conditions of the example of FIG. 5 other than the pitch P1 are the same in the examples of FIGS. 6, 7, and 8 below.

  In the example of the defocused image in FIG. 5B, asymmetry occurs in the width of about 600 nm from the end of the mark image, and the lateral shift of the mark image occurs in that portion. There is no lateral shift of the mark image. Therefore, in order to eliminate the influence of the asymmetry, as an example, the positive resist on the wafer is first exposed with the aerial images of FIGS. 5A and 5B to expose the positive resist. Next, the second exposure is performed with the above-described trim mark, and a resist within 600 nm from the end of the pattern image is exposed. Thereafter, by developing the resist, the resist image within 600 nm from the end of the pattern image disappears, and only the resist image near the center where no lateral shift occurs can be left.

  Next, in the aerial image of the mark whose pitch does not satisfy the expression (1), a lateral shift occurs due to defocusing. Therefore, as the mark 50A in FIG. 2B, a mark whose pitch P1 does not satisfy the expression (1), such as the first mark 50A in FIG. 3B or the first mark 50C in FIG. 4A. use. 3 (C) and 4 (B) show the light quantity distribution on the plane (pupil plane) of the pupil PLP of the projection optical system PL of the diffracted light from the first marks 50A and 50C, respectively.

  As the first mark 50A in FIG. 2B, the larger the lateral shift amount of the mark image generated by defocusing, the more advantageous in sensitivity. For this purpose, it is better that the 0th order light and the 1st order diffracted light from the first mark 50A are incident as close as possible on the pupil PLP of the projection optical system PL. Further, as the first mark 50A, use of a mark having the same width of the light shielding area and the width of the light transmission area, that is, a mark having a duty of 50% is advantageous in terms of sensitivity because even-order diffracted light is not generated. . In the first mark 50A in FIG. 3B, the second-order diffracted light is not generated, and the third-order diffracted light D3 goes out of the pupil PLP as shown in FIG. It can be confirmed by simulation that the best performance can be obtained by using which mark on the pupil PLP the first-order diffracted light passes.

For example, in the case of the first mark 50A in FIG. 3B and the first mark 50C in FIG. 4A, the positions where the respective first-order diffracted lights D1 and D (+1) are generated are the zero-order light in the mark 50A. Although it is close to D0 and is advantageous in terms of sensitivity, it is necessary to check whether there is an exposure amount that can be transferred simultaneously with the mark 50B in FIG. 3A because the pitch increases and the amount of first-order diffracted light decreases. is there.
Here, as in the case of the first mark 50C in FIG. 4A, when the first-order diffracted light D (+1) is incident on the projection optical system PL perpendicularly, the best focus position and the defocus amount are 100 nm. The simulation results of the aerial image are shown in FIGS. 6A and 6B, respectively. The pitch P2 of the marks 50C was 226 nm. In the defocused aerial image shown in FIG. 6 (B), asymmetry is generated at a portion of about 600 nm from the end portion, but a uniform lateral shift amount is generated on the inner side. In the example of FIG. 6 as well, only the resist image near the center that is uniformly laterally shifted can be left by the double exposure of the trim mark as in the example of FIG.

  If the image of the mark 50B in FIG. 3A and the image of the first mark 50C in FIG. 4A are simultaneously exposed on the wafer and the interval in the period direction after the double exposure of the trim mark is measured, the defocus amount Is required. In this case, when the reference value of the distance DXA between the second mark 50B and the first mark 50C is obtained in advance, it is better to expose the two marks 50B and 50C at the same time. It can be seen from FIGS. 5 and 6 that the appropriate exposure threshold for transferring two marks simultaneously is about 0.3. The defocus amount detection sensitivity (= lateral deviation amount / defocus amount) at this time was 44 nm per 100 nm defocus amount.

  Next, as shown by the first mark 50A in FIG. 3B, when the best focus position and the defocus amount are about 100 nm for the mark in which the third-order diffracted light is substantially in contact with the pupil PLP but does not enter the pupil PLP. The aerial image simulation results are shown in FIGS. 7A and 7B, respectively. The condition of such a mark pitch P3 is as follows, and the value in this example is 289.5 nm.

3 · λ / P3 = 2NA (3)
The detection sensitivity known from the simulation result of FIG. 7 is 58.8 nm per 100 nm defocus amount, which indicates that the sensitivity is higher than that of FIG.
Next, the mark pitch was further increased, and the conditions under which the third-order diffracted light was completely incident on the pupil PLP were examined. The pitch P4 in this case was as follows.
P4 = P3 (1 + 2σ) = 368 (nm)
FIGS. 8A and 8B show the simulation results of the aerial image when the best focus position and the defocus amount are 100 nm for this mark, respectively. As can be seen from FIG. 8B, the light amount distribution of the aerial image has a large asymmetry, and the distribution is such that the lateral deviation amount changes depending on the exposure amount. Therefore, the defocus amount is obtained from the lateral deviation amount of the aerial image. Measurement errors may occur. That is, when the mark pitch is increased and up to the third-order diffracted light is incident on the pupil PLP, an accurate defocus amount cannot be measured. This is because the amount of mark misalignment caused by coma during dipole illumination becomes the same amount between the second mark 50B and the first mark 50C and is canceled (details will be described later).

From the above examination, it was found that marks formed by two-beam interference (for example, marks 50A, 50B, and 50C) are desirable as marks for measuring the defocus amount. As a result, a preferable mark pitch Ps is expressed by the following equation (4) from the condition that it is greater than or equal to P1 in equation (1) and less than or equal to P3 in equation (3).
λ / (2 sin θi) ≦ Ps ≦ 3λ / (2NA) (4)
Next, the double exposure method will be briefly described. FIG. 9 shows a cross-sectional view of the resist on the wafer W and the distribution of the light quantity I (relative value) at the position X of the mark image (for example, corresponding to the image 50AP in FIG. 2D). , The exposure amount of the aerial image is adjusted so that the photosensitive level Eth of the positive resist PR crosses the light amount I indicated by the solid line at the center. This is adjusted by the transmittance of an ND filter (not shown) installed in the illumination optical system 5 of FIG. 1 and the number of integrated exposure pulses of the illumination light IL. When the exposed resist PR is developed, the concavo-convex pattern as shown in FIG. 9 should remain. However, the development is not performed immediately here, and the development is performed after the second exposure as shown in FIG. To do.

  In FIG. 10A, a region on the resist PR that is exposed by the first exposure is indicated by a white portion. Next, as an example, each of the above-described distances ΔY (values at the stage of the projection image of the projection optical system PL) are respectively provided in the vicinity of the Y direction of the mark formation regions 52A to 52E of the test reticle TR in FIG. Trim marks are formed, the wafer stage 28 is moved in the Y direction by ΔY, and a second time is applied to the regions 53A and 53B including images of several opening patterns at both ends of the mark image in FIG. Perform exposure.

  At this time, in the illumination optical system 5 of FIG. 1, for example, the aperture stop 13D is used so that telecentric collapse is not performed. In this case, the tilt angle θi of the chief ray of the illumination light IL in FIG. Further, the coherence factor (σ) of the illumination light IL is set as large as possible so that a lateral shift of the aerial image during defocus does not occur due to the interaction between defocus and coma. When the resist PR after the double exposure is developed, a resist pattern 54 as shown in FIG. 10C is formed.

  FIG. 11 shows resist patterns 541, 542, 543, and 544 similar to the resist pattern 54 in FIG. 10C, and the pitches of the resist patterns 541 to 544 in the X direction are P3, P1, P3, and P1 alternately. It shall be. Then, the distance between the resist patterns 542 and 544 having the pitch P1 and the resist patterns 541 and 543 having the pitch P3 is measured by, for example, a registration measuring machine. Since the registration measuring device usually has an imaging performance with a numerical aperture of 0.5 and a wavelength of about 530 nm, and marks with a pitch of 200 nm to 300 nm cannot be resolved, the resist patterns 542, 544 and 541, 543 in FIG. Each is measured as an aggregate of a plurality of marks having pitches P1 and P3.

  In order to improve the measurement reproducibility, the second trim exposure is performed on the regions 53A, 53C, 53D, 53E, and 53B having an interval of about 0.5 μm in FIG. The obtained pattern may be separated into resist patterns 54A, 54B, 54C, and 54D arranged at a pitch of about 1 μm in FIG. Also in this case, the second exposure does not cause telecentric collapse with a large σ value (θi = 0 in FIG. 3A).

Next, FIG. 13 shows again the angular relationship of the diffracted light of the first mark 50A of FIG. In FIG. 13, the numerical aperture NA of the projection optical system PL is sin θ ₀ , the 0th-order light position (aperture half-angle) on the pupil plane of the projection optical system PL is sinθ ₁ , the first-order diffracted light position is sinθ ₂ , and illumination light. Where iNA = σ · NA, iNA is the radius of diffracted light of each diffraction order on the pupil plane. At this time, the defocus amount detection sensitivity (= lateral deviation amount / defocus amount = Δx / Δz) is expressed by the following equation.
Δx / Δz = tan {(θ ₁ + θ ₂ ) / 2} (5)
Further, since the third-order diffracted light is substantially circumscribed by the pupil PLP of the projection optical system PL but does not enter the pupil PLP, the following relationship is established with the pattern pitch being P.
₂ sin θ ₂ = 3λ / P (6)
sin θ ₂ −sin θ ₁ = λ / P (7)
The following formula is obtained from formulas (6) and (7).

3 sin θ ₁ = sin θ ₂ (8)
Since the expressions (5) and (8) are monotonous in the range of 0 ≦ θ1 and θ2 ≦ 90 °, the detection sensitivity increases as θ ₁ and θ ₂ increase and the iNA decreases.
By using so-called halftone marks instead of the marks 50A to 50C in the above embodiment, the image contrast is improved, the resist pattern shape change amount with respect to the exposure amount error is reduced, and the exposure amount error is reduced. Increased tolerance. The defocus amount detection sensitivity is the same as when the marks 50A to 50C are used if the mark pitch is the same.

Next, an example of a simulation result of the light amount distribution by trim exposure for erasing the end portion of the mark image will be described with reference to FIG.
14A and 14B, the horizontal axis is the X coordinate (μm) and the vertical axis is the light amount I (relative value). In FIGS. 14A and 14B, distributions 55C and 55D are trim exposures, respectively. The ideal and actual light amount distributions (through the projection optical system PL) are shown, and the distributions 55A and 55B are the best focus position and the light amount distribution of the periodic mark aerial image and trim exposure when defocused by 100 nm, respectively. The light quantity distribution combined with the light quantity distribution is shown. The simulation conditions are such that the first exposure condition is NA = 0.87, λ = 193 (nm), and iNA = 0.087, and it is advantageous to increase the image contrast in the exposure of the trim pattern. The exposure conditions were NA = 0.92 and iNA = 0.174, and the exposure amount for the second time was set to 50% for the first time. 14A and 14B have different mark pitches.

The mark pitch Psmll in the example of FIG. 14 (A) is as follows.
Psml1 = λ / {2 (NA−iNA)} = 1200.92 (nm) (9)
The trim pattern is set so that the image shields the region of ± 0.973 μm at the center of the mark image. If the photosensitive level of the positive resist is set to, for example, a level of 0.4 in the distributions 55A and 55B formed by the double exposure shown in FIG. 14A, and the resist is developed, lateral shift is almost caused by 100 nm defocus. A resist pattern is formed.

On the other hand, the mark pitch Pelrg in FIG. 14B is as follows.
Pelrg = 3λ / (2NA) = 314.67 (nm) (10)
The trim pattern is set so that the image shields the region of ± 2.532 μm at the center of the mark image. If the photosensitive level of the positive resist is set to a level of, for example, 0.4 in the distributions 55A and 55B formed by the double exposure in FIG. 14B, and the resist is developed, the lateral shift is −60 nm by defocusing of 100 nm. The resist pattern thus formed is formed.

  Actually, in order to cancel various lateral deviation errors such as alignment errors, the distance (relative distance) between the mark image with the pitch Psml in FIG. 14A and the mark image with the pitch Pelrg in FIG. Is converted into a defocus amount. It is not necessary to measure all the edges of the resist pattern. If only the outermost edge is measured, the distance between the mark image with the pitch Psml and the mark image with the pitch Pelrg can be measured even with a microscope of a normal registration measuring machine. It can be measured. Since the microscope usually has a wavelength of 540 nm and a numerical aperture of about 0.5, the resolution can be up to a sinusoidal light quantity distribution of about 1 μm pitch. When a resist pattern obtained by developing the double exposure results of FIGS. 14A and 14B with a light sensitivity level of 0.4 is observed with such a microscope, 16 pieces in the center of FIGS. 14A and 14B are observed. The patterns cannot be separated and the average position is determined. Due to the limit of the resolution of the microscope, the presence of a pattern edge is recognized around ± 0.9 μm in FIG. In FIG. 14B, the presence of a pattern edge is recognized around ± 2.3 μm. Even in this case, a change in the interval between the mark image having the pitch Psml and the mark image having the pitch Pelrg due to defocusing can be measured without any problem.

  Next, FIG. 15A is a mark that forms the aerial image of FIG. 14A, that is, a mark that substitutes for the second mark 50B of FIG. 3A that does not cause a lateral shift in the aerial image due to defocusing. FIG. 15B shows the distribution of diffracted light from the mark 50D on the pupil PLP of the projection optical system PL. The mark 50D is a phase shift mark in which phase shifters of 0 ° and 180 ° are alternately provided in the transmissive portion in an L & S pattern with a duty of 50%. In this case, since the 0 ° portion, the width, and the 180 ° portion are equal, zero-order light is not generated.

The pitch Ppsm of the mark 50D is
Ppsm = 3λ / {2 (NA−iNA)} (11)
Only the first-order diffracted light D (+1) and the third-order diffracted light D (+3) are incident on the pupil PLP, and an image is formed by these two-beam interference. The average optical axis axa of the first-order diffracted light D (+1) and the third-order diffracted light D (+3) is the 0th order when the first mark 50A in FIG. 3 (B) and the first mark 50C in FIG. 4 (A). Since the average optical axis axa of the light and the first-order diffracted light is on the opposite side of the optical axis AX of the projection optical system PL, the moving direction of the mark image by defocusing is opposite. Therefore, it is also possible to measure the defocus amount by measuring the interval between the image of the mark 50D with the phase shifter of this example and the image of the mark 50A or 50C.
When the first mark is a mark (chrome mark or the like) obtained by patterning the light shielding film and the second mark is a phase shift mark, the pitch of both marks may be the same.

  An example of the simulation result of the mark 50D in FIG. 15A is shown in FIGS. 16 to 17, the horizontal axis represents the X coordinate, and the vertical axis represents the light amount I (relative value). The simulation conditions are: NA = 0.874, λ = 193 (nm), iNA = 0.087, sin θi = NA−iNA, and in this case, the pitch Ppsm = 362.8 (nm). It is.

  FIGS. 16A and 16B show the light amount distribution of the aerial image of the mark having the pitch Ppsm at the best focus position and the position defocused by 100 nm, respectively. Note that the rectangular waveforms in FIGS. 16A and 16B indicate the amplitude distribution of the illumination light passing through the corresponding marks (the values are inverted to indicate that the phase is 0 ° and 180 °). It is. From FIG. 16B, it can be seen that a lateral shift of +35 nm occurs in the mark image defocused by 100 nm. The change rate (detection sensitivity) of the distance between the image of the first mark 50A (pitch Pelrg is 314.67 nm) in FIG. 3B and the image of the mark 50D in FIG. 95 nm.

Next, FIG. 17 shows a light amount distribution after trim exposure is performed on the mark images of FIGS. 16A and 16B, and distributions 55C and 55D are an ideal and actual light amount distribution of trim exposure, respectively. Distributions 55A and 55B show the light amount distribution in the best focus position and 100 nm defocused state, respectively.
The trim pattern is set so that the image shields the region of −3.085 to +3.084 μm in the center of the mark image. Since it is advantageous to increase the image contrast in the exposure of the trim pattern, the second exposure condition is set to NA = 0.92 and iNA = 0.087. The exposure amount for the second time was set to 50% for the first time. If the positive resist exposure level is set to 0.4, for example, of the light amount distribution formed by double exposure and the resist is developed, a resist pattern laterally shifted by +35 nm is formed by defocus of +100 nm. The defocus amount may be obtained from the interval between this resist pattern and the resist pattern shown in FIG.

18A shows a mark 50E that can be used instead of the second mark 50B in FIG. 3A, and FIG. 18B shows the mark 50E on the pupil PLP of the projection optical system PL. The distribution of diffracted light is shown. The mark 50E is a phase shift mark in which the 0th-order light disappears. Only the first-order diffracted light D (+1) and the second-order diffracted light D (+2) are incident on the pupil PLP of the projection optical system PL, and a mark image is formed by two-beam interference. Is done. The average optical axis axa of the first-order diffracted light D (+1) and the second-order diffracted light D (+2) is the 0th-order light of the first mark 50A in FIG. 3B and the first mark 50C in FIG. And the average optical axis axa between the first-order diffracted light and the opposite side of the optical axis AX of the projection optical system PL, the movement direction of the mark image by defocusing is opposite. Therefore, it is possible to measure the defocus amount with higher sensitivity by measuring the interval between the image of the mark 50E and the image of the mark 50A or 50C.
In particular, when the reference value of the distance DXA between the second mark 50E and the first mark 50C is obtained in advance, it is better to expose the two marks 50E and 50C at the same time. Note that the influence of the drawing error can be reduced by combining and exposing the two marks. This is because the amount of mark misalignment caused by coma during dipole illumination becomes the same amount between the second mark 50B and the first mark 50C and is canceled (details will be described later).

  A configuration example of the mark 50E is shown in FIG. FIG. 20B shows an amplitude distribution 56A corresponding to the position X of the illumination light transmitted through the mark 50E, and assuming that the pitch is P, the portion of the width Mnsl = P / 8 is 180 from the left side of FIG. The phase shift portion 56Ab, the portion of the next width Pls1 = P / 4 is the transmission portion 56Aa of the phase 0 °, the portion of the next width Mns2 = 3P / 8 is the 180 ° phase shift portion 56Ab, the next width Pls2 = P / 4 is a transmission part 56Aa having a phase of 0 °, and a transmission part having a phase of 0 ° and a phase shift part having a phase of 180 ° are alternately repeated in this order. In this case, since all the widths per pitch of the transmission part having a phase of 0 ° and the phase shift part having a phase of 180 ° are the same, zero-order light is not generated. The structure of the mark 50E is determined by simulation so that the amplitudes of the first-order diffracted light and the second-order diffracted light are the same. When the pitch Ppsm12 of the mark 50E is given by the following equation, the detection sensitivity of the defocus amount is maximized.

Ppsm12 = λ / (NA−iNA) (12)
An example of the simulation result of the mark 50E in FIG. 18A is shown in FIGS. In each of FIGS. 19 to 22 (excluding FIG. 20B), the horizontal axis represents the X coordinate, and the vertical axis represents the light amount I (relative value). The simulation conditions are NA = 0.874, λ = 193 (nm), iNA = 0.087, sin θi = NA−iNA in the first exposure condition. In this case, the pitch Ppsm12 in the equation (12) is 242. (Nm).

  Distributions 56B, 56C, and 56D in FIGS. 19A, 19B, and 20A are the best focus position, the 100 nm defocused position, and the −100 nm defocused position, respectively. The light quantity distribution of the mark image of Ppsm12 is shown. Note that the rectangular amplitude distribution 56A in these drawings is the amplitude distribution of illumination light on the corresponding mark. From FIG. 19B, it can be seen that the defocus amount is 100 nm and the mark image has a lateral shift of +50 nm. The detection sensitivity due to the change in the distance between the mark 50E in FIG. 18A and the first mark 50A in FIG. 3B (when the pitch Pelrg is 314.67 nm) is 110 nm per 100 nm defocus.

  Distributions 56B, 56C, and 56D in FIGS. 21A, 21B, and 22 are the light amount distributions of the mark images in FIGS. 19A, 19B, and 20A, respectively. FIG. 23 shows the light amount distribution after performing trim exposure (double exposure) using the trim pattern of the light amount distribution 57A of FIG. 23, and the trim pattern of FIG. 23 has the same amplitude distribution 56A as FIG. A portion corresponding to a portion where the width in the X direction at the center of the phase shift mark is N × P (N is an integer) is a light shielding portion. Further, the edge portion of the light amount distribution 57A of the trim pattern is positioned at the center of the 180 ° phase shift portion 56Ab having a width of 3P / 8, that is, only Sft (= 3P / 16) from the transmission portion 56Aa having a phase of 0 °. Alignment is performed so as to shift the position.

  In FIGS. 21A, 21B, and 22, distributions 57A and 57B indicate ideal and actual light amount distributions at the time of trim exposure, respectively, and trim exposure is advantageous when the image contrast is increased. Therefore, the second exposure conditions were NA = 0.92 and iNA = 0.087. Also, telecentric collapse is not performed so that the trim mark does not shift laterally due to defocusing. Further, the exposure amount for the second time was set to 50% for the first time. In the example of FIG. 21B, when the positive resist photosensitive level is set to 0.4, for example, in the light amount distribution after double exposure and the resist is developed, a +50 nm defocus occurs due to +100 nm defocus. A resist pattern is formed. Therefore, by measuring the interval between the image of the first mark 50A and the mark image of the distribution 56B in FIG. 3B and converting it to a defocus amount, detection sensitivity is obtained in which the interval changes by 110 nm due to 100 nm defocus. It is done. Similarly, also in the mark image defocused by −100 nm in FIG. 22, the lateral shift amount can be accurately measured by setting the resist photosensitive level to 0.4.

  Next, another example of the embodiment of the present invention will be described with reference to FIGS. In this example, in the projection exposure apparatus shown in FIG. 1, a pattern extending along the Y direction on the object plane of the projection optical system PL (hereinafter referred to as V line) and a pattern extending along the X direction (hereinafter referred to as H line). The defocus amount of the image by the projection optical system PL is evaluated separately. This can be used, for example, when astigmatism of the projection optical system PL is obtained from the difference between the best focus position (image plane) of the V-line image and the best focus position of the H-line image. For such an application, it is convenient if there is a mark that can simultaneously measure the defocus amount of the pattern image corresponding to the V-line and the H-line independently.

  FIG. 24 shows such a mark. In the mark formation region 52A of the test reticle TR of FIG. 24, two X-axis marks 58X and 59X made of L & S patterns formed at different pitches in the X direction are in the X direction. Are arranged at predetermined intervals, and two Y-axis marks 58Y and 59Y made of L & S patterns formed at different pitches in the Y direction are arranged at predetermined intervals in the Y direction. The defocus amount of the pattern image corresponding to the V line can be obtained from the interval between the images of the X-axis marks 58X and 59X, and the pattern image corresponding to the H line can be determined from the interval between the images of the Y-axis marks 58Y and 59Y. Can be obtained. The illumination light in this case is irradiated to the mark forming region 52A in a state where telecentric collapse is performed along the incident surface 60 that intersects the X axis at 45 °.

  FIG. 25 is a perspective view showing the entrance surface 60 of FIG. 24. In FIG. 25, the center of the pupil PLP of the projection optical system PL is the origin of the XYZ orthogonal coordinate system. Further, the X ′ axis and the Y ′ axis are set in parallel to the X axis and the Y axis with the intersection point between the + Z side surface of the test reticle TR and the Z axis as the origin. The chief ray of the illumination light IL passes through the incident surface 60 inclined by 45 ° with respect to the X ′ axis, and the 0th-order light (straight-ahead light) D0 within the pupil PLP of the projection optical system PL from the optical axis to the X axis. The incident angle θi of the principal ray is set so as to pass through a position shifted in the intersecting direction at 45 °. The incident angle θi is set to the following relationship using the numerical aperture NA of the projection optical system PL and the numerical aperture iNA of the illumination optical system.

sin θi = NA−iNA (13)
In this case, X-axis marks 58X and 59X and Y-axis marks 58Y and 59Y in FIG. 24 are formed in the illumination area of the illumination light IL in FIG. At this time, unless the mark pitch is increased with respect to the X-axis and Y-axis marks, diffracted light is scattered by an aperture stop (not shown) that defines the pupil PLP. It can be determined geometrically how much it should be expanded.
26A shows the first-order diffracted light D1 (x) and the second-order diffracted light D2 (x) generated from the mark 59X in FIG. 24, and FIG. 26B shows the diffracted light at the pupil PLP of the projection optical system PL. The distribution of is shown. Note that the angle θi ′ in FIG. 26A is an angle obtained by projecting the angle θi in a plane that intersects the X axis at 45 ° onto the XZ plane. As the mark 59X, a phase shift mark such as the mark 50E in FIG. 18 is used. In FIG. 26A, the pitch of the mark 59X may be calculated by converting the numerical aperture NA of the projection optical system PL into an equivalent numerical aperture NAeq as follows.

NAeq = (NA−iNA) / 2 ^1/2 + iNA
= NA / 2 ^1/2 + (1-1 / ^1/2 ) iNA (14)
Pitch = λ / (NAeq−iNA)
Therefore, if the pitch of the one-dimensional mark (for example, the mark 50E in FIG. 18) is multiplied by about ^21/2 , it can be used as the two-dimensional X-axis mark 59X in FIG. Also for the Y-axis mark 59Y in FIG. 24, the X-axis mark 59X is rotated by 90 ° so that the first-order diffracted light D1 (y) and the second-order diffracted light D2 (y) are shown in FIGS. Marks can be used.

Next, with respect to the marks 58X and 58Y in FIG. 24, as shown in FIG. 27, the third-order diffracted lights E3 (x) and E3 (y) from the marks 58X and 58Y substantially circumscribe the pupil PLP of the projection optical system PL. Sensitivity is maximized under conditions that do not enter the PLP.
In FIG. 27, only the 0th order light D0 and the 1st order diffracted light E1 (x) from the mark 58X and the 0th order light D0 and the 1st order diffracted light E1 (y) from the mark 58Y are in the pupil PLP. Therefore, as the mark 58X, a mark having the same pitch as that of the first mark 50A in FIG. 3B can be used, and a mark obtained by rotating the mark 58X by 90 ° can be used as the mark 58Y.

In FIG. 27, the numerical aperture NA of the projection optical system PL is set as follows from the condition that the third-order diffracted light E3 (x) having the distance B from the optical axis AX to the X direction and the distance A to the Y direction circumscribes the pupil PLP. When the equivalent numerical aperture NAeq2 of the equation (14) is also used in terms of the numerical aperture NAeq2 equivalent to, the pitch of the mark 58X (also referred to as P) is as follows.
P = 3 / (NAeq + NAeq2) (15)
B = {(NA + iNA) 2 - (NA-iNA) 2/2} 1/2 ... (16)
NAeq2 = B-iNA (17)
The mark shown in FIG. 24 in this example is a mark used for simultaneously measuring the defocus amounts of the V-line and H-line images, but the diffracted light necessary for image formation is also used when other marks are used. Is determined by the aperture stop of the projection optical system PL or the mark pitch is determined so that unnecessary diffracted light does not leak into the pupil of the projection optical system PL.
Note that the pitch of the marks 58X and 58Y does not depend on the equation (15).
Pitch = λ / (NAeq−iNA)
This is preferable because the influence of coma aberration is offset when measuring the distance between marks by dipole illumination.

Next, a method for measuring a drawing error (an error with respect to a design value) of an interval (relative distance) between a plurality of marks for focus measurement formed on the test reticle TR will be described. In the following, the drawing error is measured as a value in the projected image.
First, in order to measure the defocus amount relating to the V-line, a test reticle (also referred to as TR) in which the mark 50C in FIG. 4A and the mark 50E in FIG. 18A are formed at a predetermined interval. It is assumed that the pitch of the marks 50C and 50E is represented by the right side of the equation (12). In this case, in order to measure the drawing error of the interval between these marks, the aperture stop 13A is selected in the illumination optical system 5 of FIG. As a result, as shown in FIG. 28A, the marks 50C and 50E are illuminated with the illumination light IL that is symmetrically inclined at an angle θi in the ± X directions, and as shown in FIG. 28B, the projection optical system PL In the pupil PLP, the first-order diffracted light D (+1) from the marks 50C and 50E passes on the optical axis, and the zero-order light D0 from the mark 50C and the second-order diffracted light D (+2) from the mark 50E pass through both ends. Pass through. Then, similarly to the above-described embodiment, the distance between the resist patterns which are images of the marks 50C and 50E is measured after exposure and development.

At this time, the lateral shift amounts of the images of the marks due to asymmetric aberration such as coma aberration that the diffracted light (imaging light flux) from the marks 50C and 50E suffers are the same. The lateral shift due to the asymmetry is canceled when measuring. Therefore, the drawing error δE can be obtained by subtracting the design value (target value) from the measurement result of the interval. Note that symmetric aberration such as spherical aberration does not cause a lateral shift of the aerial image during dipole illumination, and therefore does not cause a measurement error of the relative position of the mark image.
Further, when measuring the defocus amount related to the H-line, marks obtained by rotating the marks 50C and 50E by 90 ° are used, but when measuring the drawing error of these marks, these marks are moved at an angle θi in the ± Y direction. What is necessary is just to illuminate with the dipole illumination using the illumination light IL inclined symmetrically.

Thereafter, in order to perform focus measurement as in the above-described embodiment, the measurement result when illuminating the mark by performing telecentric collapse, that is, the distance between the images of the marks 50C and 50E is obtained by subtracting the design value of the distance. The deviation δT includes the following components.
(A1) Deviation δ1 due to defocus. (A2) Mark drawing error δ2. (A3) Deviation δ3 due to symmetric aberration (spherical aberration, etc.) other than defocus.
Among these components, the drawing error δ2 of the mark (a2) is the same as the drawing error δE obtained as described above. Further, the deviation δ3 in (a3) can be measured by another aberration measurement method such as aberration measurement using an aerial image. In this way, only the deviation δ1 due to defocusing can be separated and extracted from the deviation δT. As an example, exposure can be performed in the best focus state by adjusting the relative distance in the Z direction between the projection optical system PL and the wafer W so that the deviation δ1 is minimized.

Next, a method for measuring a mark drawing error when the defocus amounts for the V-line and the H-line are simultaneously measured using the X-axis marks 58X and 59X and the Y-axis marks 58Y and 59Y shown in FIG. To do. The pitch of the marks 58X and 59X can be expressed by an expression using the equivalent numerical aperture NAequ of the expression (14) as the numerical aperture NA of the expression (12), but the mark 59X is a phase shift mark. The marks 58Y and 59Y are marks obtained by rotating the marks 58X and 59X by 90 °. At the time of measuring the defocus amount, as shown in FIG. 29A, the marks 58X and 58Y are illuminated with the illumination light IL inclined at the angle θi of the expression (13) from the direction intersecting the X axis at 45 °. . The angle θi ′ is an angle when the angle θi is projected onto the XZ plane. As shown in FIG. 29B, among the diffracted lights from the marks 58X and 58Y, only the 0th-order light D0 and the first-order diffracted lights D1 (x) and D1 (y) form the pupil PLP of the projection optical system PL. The second-order diffracted light (positions PD2 (x), PD2 (y)) disappears after passing.
Similarly, as shown in FIG. 30A, the marks 59X and 59Y are also illuminated with the illumination light IL from the direction intersecting the X axis at 45 °. As shown in FIG. 30B, the marks 59X and 59Y The zero-order light (position PD0) of the diffracted light disappears, and only the first-order diffracted lights D1 (x) and D1 (y) and the second-order diffracted lights D2 (x) and D2 (y) are pupils of the projection optical system PL. Pass PLP.

  In order to measure the drawing error of these marks, as shown in FIG. 31 corresponding to FIG. 25, the clock is relative to the X ′ axis on the test reticle TR (the axis parallel to the X axis on the pupil PLP). Assume incident surfaces 60A and 60B that intersect the direction at + 45 ° and -45 °, respectively, and are perpendicular to the reticle surface. For the X-axis marks 58X and 59X, in the incident surfaces 60A and 60B, in the −Y ′ direction (the Y ′ axis is an axis parallel to the Y axis on the pupil PLP), the incident angle θi is the main. Illumination is performed with dipole illumination using two illumination lights having light beams IL1x and IL2x. Thereby, in the pupil PLP of the projection optical system PL shown in FIG. 31, the 0th-order light D10 (x) from the mark 58X, the second-order diffracted light D22 (x) from the mark 59X, and the 0th-order light D20 from the mark 58X. (X) and the second-order diffracted light D12 (x) from the mark 59X pass symmetrically with respect to the Y-axis, and the first-order diffracted lights D11 (x) and D21 (x) from the marks 58X and 59X pass on the Y-axis. Therefore, the lateral shift amount of the mark image due to the asymmetrical aberration becomes the same, and the influence of the asymmetrical aberration on the relative position measurement of the images of both marks can be offset, so that only the mark drawing error can be measured.

  Similarly, for the Y-axis marks 58Y and 59Y, two poles using two illumination lights having principal rays IL1y and IL2y respectively having an incident angle θi in the + X ′ direction in the incident surfaces 60A and 60B of FIG. Illuminate with lighting. Thus, in the pupil PLP of the projection optical system PL shown in FIG. 32, the zero-order light D10 (y) from the mark 58Y, the second-order diffracted light D22 (y) from the mark 59Y, and the zero-order light D20 from the mark 58Y. (Y) and the second-order diffracted light D12 (y) from the mark 59Y pass symmetrically with respect to the X-axis, and the first-order diffracted lights D11 (y) and D21 (y) from the marks 58Y and 59Y pass on the X-axis. Therefore, the lateral shift amount of the mark image due to the asymmetrical aberration becomes the same, and the influence of the asymmetrical aberration on the relative position measurement of the images of both marks can be offset, so that only the mark drawing error can be measured.

In addition to the marks 50E shown in FIGS. 18A and 20B, the phase shift mark shown in FIG. 33 or the halftone mark shown in FIG. Can be used. Since the minimum line width of these marks can be made wider than that of the mark 50E, the test reticle TR can be easily manufactured. The projected images of these marks are equivalent to the image of the mark 50E.
In the phase shift mark represented by the transmitted light amplitude distribution 70A in FIG. 33, a phase shifter portion 70Ab having a phase of 180 ° is formed between the transmitting portion 70Aa having a phase of 0 ° and the light shielding portion 71. If the mark pitch P is 16A (A is a constant), the width of the phase shifter 70Ab, the transmission part 70Aa, the light-shielding part 71, and the transmission part 70Aa is sequentially increased to 6.5A, 3.25A, 3.. It is formed at 0A and 3.25A. In this case, if the pitch P is 176 nm on the wafer, the minimum line width 3P / 16 is 33 nm on the wafer, and the projection magnification is 1/4 on the reticle to 132 nm.

Next, in the halftone mark having the pitch P represented by the amplitude distribution 72A of the transmitted light in FIG. 34, the intensity transmittance HT is 6%, for example, at a phase of 180 ° between the transmission portions 72Aa of the phase 0 ° of the width PHL. About a phase shift portion 72Ab is formed. When the intensity transmittance HT is used, the width PHL is as follows.
PHL = P × HT ^1/2 / (1 + HT ^1/2 ) (18)
As an example, when HT = 0.06 (6%), the width PHL is P × 0.197. In the example of FIG. 34, as shown by the transmittance distribution 73A, trim exposure is performed so that only the central N pitch portion is exposed.

The present invention is not limited to a scanning exposure type projection exposure apparatus, but a projection optical system using a batch exposure type projection exposure apparatus or an immersion type exposure apparatus disclosed in, for example, International Publication (WO) 99/49504. It can also be applied to measuring focus information. The exposure apparatus of the present invention includes a semiconductor device, a display device such as a liquid crystal display element or a plasma display, an image sensor (CCD, etc.), a micromachine, a thin film magnetic head, a DNA chip, etc., and a mask (photomask, reticle, etc.). The present invention can also be applied to an exposure apparatus for manufacturing.
In addition, this invention is not limited to the above-mentioned embodiment, Of course, a various structure can be taken in the range which does not deviate from the summary of this invention.

  The present invention can be used for measurement of imaging characteristics of a projection exposure apparatus used when manufacturing various devices such as semiconductor devices, measurement of flatness of a substrate to be exposed, and the like.

It is a perspective view which shows the projection exposure apparatus of an example of embodiment of this invention. (A) is a plan view showing an example of a test reticle, (B) is an enlarged view showing a mark formation region 52A in FIG. 2 (A), (C) is a plan view showing an example of a shot arrangement of a wafer, (D () Is an enlarged view showing a mark image projected on the wafer. (A) is a diagram showing diffracted light from the mark 50B, (B) is a diagram showing diffracted light from the mark 50A, and (C) is a diagram showing diffracted light on the pupil plane of the projection optical system PL. (A) is a figure which shows the diffracted light from the mark 50C, (B) is a figure which shows the diffracted light on the pupil plane of the projection optical system PL of FIG. 4 (A). It is a figure which shows the simulation result of the light quantity distribution of the image of the mark 50B. It is a figure which shows the simulation result of the light quantity distribution of the image of the mark 50C. It is a figure which shows the simulation result of the light quantity distribution of the image of the mark 50A. It is a figure which shows the simulation result of the light quantity distribution of the image at the time of enlarging pitch with the mark 50A. It is a figure which shows the light quantity distribution of the resist cross section on a wafer, and the mark image on it. It is explanatory drawing of the trim exposure for erasing the both ends of a mark image. It is a figure which shows an example of the arrangement | sequence of the mark image on a wafer. It is explanatory drawing of the trim exposure for erasing the both ends etc. of a mark image. It is a figure which shows the diffracted light from the mark 50A. It is a figure which shows the example of the light quantity distribution after trim exposure. It is a figure which shows the diffracted light from the phase shift type mark 50D. It is a figure which shows the simulation result of the light quantity distribution of the image of the mark 50D. It is a figure which shows the result of having superimposed the light quantity distribution of trim exposure on the light quantity distribution of FIG. It is a figure which shows the diffracted light from the phase shift type mark 50E. It is a figure which shows the simulation result of the light quantity distribution of the image of the mark 50E. (A) is a figure which shows the simulation result of the light quantity distribution of the image of the mark 50E on another condition, (B) is a figure which shows the amplitude of the transmitted light of the mark 50E. It is a figure which shows the result of having superimposed the light quantity distribution of trim exposure on the light quantity distribution of FIG. It is a figure which shows the result of having overlapped trim exposure on the light quantity distribution of FIG. It is a figure which shows an example of the mark for trim exposure. It is an enlarged view showing an example of a mark that can simultaneously and independently measure the defocus amount of the pattern image corresponding to the V line and the H line. It is a figure which shows an example of the illumination light with respect to the mark of FIG. It is a figure which shows the diffracted light from the one part mark of FIG. It is a figure which shows the diffracted light from the other mark of FIG. (A) is a figure which shows the diffracted light from the marks 50C and 50E, (B) is a figure which shows the diffracted light on the pupil plane of the projection optical system PL of FIG. 28 (A). (A) is a figure which shows the diffracted light from the marks 58X and 58Y, (B) is a figure which shows the diffracted light on the pupil plane of the projection optical system PL of FIG. 29 (A). (A) is a figure which shows the diffracted light from marks 59X and 59Y, (B) is a figure which shows the diffracted light on the pupil plane of the projection optical system PL of FIG. 30 (A). It is a figure which shows an example of the illumination light with respect to the marks 58X and 59X. It is a figure which shows an example of the illumination light with respect to the marks 58Y and 59Y. It is a figure which shows the amplitude distribution of the transmitted light of the phase shift mark of another example. It is a figure which shows the amplitude distribution of the transmitted light at the time of performing trim exposure using the halftone mark of another example.

Explanation of symbols

    DESCRIPTION OF SYMBOLS 1 ... Main control system, 5 ... Illumination optical system, TR ... Test reticle, PL ... Projection optical system, W ... Wafer, 28 ... Wafer stage, 52A-52E ... Mark formation area, 50A-50E ... Mark, 50AP, 50BP ... Mark statue

Claims (11)

A focus measurement method for measuring focus information of a projection optical system,
By using illumination light whose chief ray is inclined with respect to the optical axis of the projection optical system, it is possible to obtain a light intensity distribution on the pupil plane of the projection optical system such that the distance from the optical axis on the pupil plane is different from each other. A first step of illuminating a first mark that generates two diffracted lights;
A second step of projecting an image of the first mark on a predetermined surface via the projection optical system to obtain position information of the image of the first mark;
And a third step of obtaining focus information of the projection optical system based on the position information obtained in the second step.   The third step includes a step of obtaining a defocus amount of the projection optical system based on positional information of the image of the first mark. The intensity distribution of the two diffracted lights on the pupil plane is shifted in a direction orthogonal to a straight line passing through the optical axis on the pupil plane,
The focus measurement method according to claim 1, wherein the second step includes a step of obtaining position information of the image of the first mark with respect to the orthogonal direction on the predetermined plane. The first step generates a plurality of diffracted lights so that the light intensity is distributed at different positions with respect to the position of the light intensity distributed on the pupil plane by the two diffracted lights generated by the first mark. Including the step of illuminating the mark,
The said 2nd process includes the process of calculating | requiring the positional information on the image of the said 1st mark projected on the said predetermined surface, and the image of the said 2nd mark, The Claim 1 characterized by the above-mentioned. The focus measurement method described.   5. The focus measurement method according to claim 4, wherein at least one of the first mark and the second mark is a phase shift type mark in which zero-order light is not substantially generated. In the first step, the light intensity distribution of the illumination light on the pupil plane of the illumination optical system is raised in two regions decentered from the optical axis of the illumination optical system with respect to the central portion, and the first mark And illuminating the second mark,
In the third step, the image of the first mark and the image of the second mark projected on the predetermined surface with respect to a reference value of the interval between the first mark and the second mark on the predetermined surface The focus measurement method according to claim 4, wherein a deviation amount of the interval is obtained. In an exposure apparatus that transfers a pattern onto a substrate via a projection optical system,
An illumination optical system for illuminating a first mark disposed on the object plane side of the projection optical system;
The chief ray of illumination light that illuminates the first mark so that the two diffracted lights generated from the first mark have a light intensity distribution such that the distances from the optical axis on the pupil plane of the projection optical system are different from each other. An exposure apparatus comprising: a control device that controls an inclination angle of the exposure apparatus. The illumination optical system illuminates a second mark different from the first mark arranged on the object plane side of the projection optical system;
The control device generates a plurality of diffracted lights so that the light intensities are distributed at different positions with respect to the positions of the light intensities distributed on the pupil plane by the two diffracted lights generated from the first mark. 8. The exposure apparatus according to claim 7, wherein an inclination angle of a chief ray of illumination light that illuminates the second mark is controlled. A mask that is illuminated under illumination conditions in which the chief ray of illumination light is inclined,
A focus measurement mask characterized in that a first mark and a second mark that are different from each other in the way of diffracted light and are symmetrical in a predetermined periodic direction are formed side by side.   When illuminated under the illumination conditions, the center of the center of gravity of each of the two diffracted lights generated from the first mark and the center of the center of gravity of each of the two diffracted lights generated from the second mark are in the periodic direction. The focus measurement mask according to claim 9, wherein the focus measurement mask is misaligned.   10. The focus measurement mask according to claim 8, wherein at least one of the first mark and the second mark is a phase shift type mark that does not substantially generate zero-order light.

Images