JPH06160019A - Positional shift measuring device - Google Patents

Abstract

PURPOSE:To measure positional shift with high accuracy of nm order by eliminating the error of an optical detection system even when the mutual fluctuations of the in-plane rotation of an object to be inspected are generated between the optical detection system and the object to be inspected. CONSTITUTION:Diffraction gratings 21, 22 to be inspected are provided on a wafer in order to calculate the mutual pattern shift of the gratings. The diffraction gratings 21, 22 correspond to the N-th and (N+1)-th baked patterns of a reticle or mask at the time of the formation of a semiconductor. Further, reference diffraction gratings 25, 26 are formed in the vicinity of the diffraction gratings 21, 22 to be inspected and a mask or reticle for backing is preliminarily formed as an EB image so that the mutual shift quantity of the gratings become zero. At this time, the measured values of the diffraction gratings 25, 26 are set to reference values at each time setting the shift quantity to a zero value to suppress the generation of the error generated by the in-plane rotary component of the water to realize highly accurate pattern measurement.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an alignment apparatus for an exposure apparatus for manufacturing semiconductors, or a positional deviation measuring apparatus used for measuring the overlay accuracy of semiconductor patterns.

[0002]

2. Description of the Related Art Conventionally, in order to measure the shift between two diffraction gratings to be measured, coherent light is made incident on each of the grating patterns, and the obtained diffracted light is used to form interference light. There is known a method of measuring the phase difference between two diffraction gratings based on the phase difference between two beat signals obtained from interference light including.

[0003]

However, in the relative relationship between the incident light fluxes on the two measured patterns and the patterns, the in-plane rotation amount and the phase difference Δφ of the wafer on which these measured patterns are formed are determined. There are the following problems regarding fluctuations. This will be described with reference to FIGS. 13 and 14. 13 and 14 are explanatory views showing a state in which a light flux enters the test pattern, where the distance between the centers of the diffraction grating patterns 1 and 2 to be measured is d, the center point of the pattern 1 is P1, and the pattern 2 is Pattern 1 with the center point as P2
2 and 2, a light beam L1 (angular frequency ω1) obliquely enters the paper from the right side of the drawing, and a light beam L2 (angular frequency ω2) similarly enters from the right side of the drawing, causing diffracted light beams emitted from the same grating to interfere with each other. To obtain a 2-beat signal.

Then, as shown in FIG. 13, there is shown a case where the incident angle of the light flux on the patterns 1 and 2 is deviated from the lattice arrangement direction by θz as viewed from the z direction. At this time, the point
Let H be the intersection of the perpendiculars drawn from P1 to the ray L2 entering point P2. If ray L2 is a plane wave, then line segment HP1
Is included in the equiphase surface. If the ray of light flux L2 entering point P1 is L3, the ray of light flux L2 entering point P2 is L4, and the direction cosines of the traveling directions of light rays L3 and L4 are (α, β, γ), the line segment H
The length of P2 is as follows. HP2 = β ・ d (1)

Further, the light ray entering the point P1 of the light flux L1 is L5,
If the ray entering the point P2 is L6, the complex amplitude display of the first-order diffracted ray of ray L3 is as follows. Ea '(f1) = A · exp {i (ω1 · t + φ1 + φa)}… (2)

Similarly, the complex amplitude display of the diffracted lights of the light rays L5, L4 and L6 is as follows corresponding to the equations (2), (3) and (4). Ea '(f2) = B · exp {i (ω2 · t + φ2-φa)} (3) Eb' (f1) = A · exp [i {ω1 · t + φ1 + φb + (2π / λ) β · d}] … (4) Eb '(f2) = B · exp [i {ω2 · t + φ2 −φb + (2π / λ) β · d}] (5)

Due to the interference between Ea '(f1) and Ea' (f2), the beat signal becomes as follows. Ia '= | Ea' (f1 ) + Ea '(f2) | 2 = A 2 + B 2 + 2A · B cos {(ω1 -ω2) t + (φ1 -φ2) + 2φa} ... (6)

Similarly, due to the interference between Eb '(f1) and Eb' (f2), the beat signal is given by the following equation. ^{Ib '= A 2 + B 2} + 2A · B cos {(ω1 -ω2) t + (φ1 -φ2) + 2φb - 2 (2π / λ) β · d} ... (7)

Therefore, when θz ≠ 0 according to the equations (6) and (7), compared with the case where θz = 0, that is, β = 0, (4π /
The phase changes by λ) β · d. Than this,
If the error included in the measured value of the amount of deviation between the diffraction gratings 5 and 6 is Δx ′, then Δx ′ = 2β · d (8), and the value corresponding to Δx ′ = 5 nm is d = 100 μm. β = Δx ′ / 2d = 5 nm / {2 × 100 (μm)} = 2.
It becomes 5 × 10 ^-5 .

[0010] By the ^{way, θz = 90 ° - cos -1} β = 5.
When the value of 1z, that is, the value of θz fluctuates by 5.1 ″, the detection value between the grating patterns 5 and 6 deviates by 5 nm.

This means that if there is a relative fluctuation of 5.1 "between the detection optical system and the wafer, which corresponds to the rotation in the wafer plane, it is 5n.
This causes a variation error of m, which is a serious problem in detecting the deviation amount with high accuracy. Usually, when the wafer stage feeds a 6-inch or 8-inch wafer, it is necessary to move a long stroke of about 200 mm, and a complicated monitor function and control are required to prevent rotation on the order of seconds. To do. Further, when attempting to achieve stage feed with a simple configuration, rotation on the order of seconds, that is, yawing occurs, and a shift detection error in the pattern corresponding to this occurs. In a stage with a stroke of 200 mm, yawing of about 10 ″ to 20 ″ is usually generated, and a measurement error of 10 nm to 20 nm is generated accordingly. Therefore, it is not possible to achieve the printing accuracy evaluation of 5 nm or less, which is suitable for the desired 0.25 μm line and space pattern (equivalent to 256 MDRAM). Also, to reduce the absolute value of yawing,
It is conceivable to control the stage with high precision, but in this case the system becomes complicated and expensive.

Although the conventional problems have been described with reference to FIGS. 13 to 14 regarding the displacement measurement between a plurality of patterns on the same object, the same problem is also encountered in the displacement measurement between patterns on two objects. is there.

An object of the present invention is to eliminate the error between the optical detection system and the object to be inspected, even if there is mutual variation in in-plane rotation of the object to be inspected, and to provide a highly accurate position on the order of nm. It is to provide a deviation measuring device.

[0014]

A position shift measuring apparatus according to the present invention for achieving the above object is an apparatus for measuring a position shift between a plurality of diffraction grating patterns provided on one object or two objects. A coherent light source, a system for injecting light into the diffraction grating, a system for interfering the diffracted light from the diffraction grating with at least one of the interference lights, and a phase shift of the diffraction grating is measured based on an interference signal. Means for moving at least a part of the photodetection system and the object relative to each other, of the diffraction grating pattern for which a displacement amount is known in advance on the object and displacement measurement should be obtained. By measuring a reference pattern provided in the vicinity by the photodetection system, an error due to mutual in-plane rotation between the detection system and the object in the lattice plane of the displacement measurement object is reduced.

[0015]

In the position shift measuring device having the above-mentioned structure, a reference pattern whose shift amount is known in advance is provided in the vicinity of the diffraction grating pattern pair on the object to be inspected, and the corner test is performed by referring to the measured value of the reference pattern. This is a position shift detection device that measures the shift within the diffraction grating pattern set, which is generated when the optical detection system and the wafer are moved relative to each other on a stage or the like in order to measure the pattern widely provided within the wafer. The in-plane rotation component of the wafer, that is, the occurrence of an error due to yawing is suppressed.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail with reference to the embodiments shown in FIGS. FIG. 1 is an explanatory view of the first embodiment, and reference numerals 21 and 22 are diffraction gratings to be inspected provided on a wafer for obtaining mutual pattern shifts. Further, test diffraction gratings 23 and 24 are provided at positions separated by Dx in the −x direction of the test diffraction gratings 21 and 22 by baking or the like for the same evaluation. Therefore, for example, the diffraction gratings 21 and 22 to be inspected correspond to the Nth and N + 1th printing patterns of the reticle or mask at the time of manufacturing a semiconductor, and when this portion is enlarged, it becomes as shown in FIG. Similarly, the inspected diffraction gratings 23 and 24 are the Mth and M + 1th printing patterns of the reticle or mask, and M = N.
But it's okay. In addition, in the -y direction of the diffraction gratings 21 and 22 to be tested.
The reference diffraction gratings 25 and 26 are formed at positions apart from Dy as shown in FIG. 3, and a mask or reticle for printing is prepared by EB drawing so that the mutual deviation is 0 in advance. It

ΔX is a pattern printing deviation amount of the diffraction gratings 21 and 22 to be inspected printed on the wafer. Here, the pitch of the diffraction grating is P. In FIG. 1, S1,
Each of S2 and S3 indicates a spot when the projection light of the photodetection system is incident on the pair of the test diffraction gratings 21 and 22, the pair of the test diffraction gratings 23 and 24, and the pair of the reference diffraction gratings 25 and 26. ing.

Therefore, when the photodetection system shown in FIG. 5 and the wafer to be inspected are moved relative to each other, the diffraction gratings 21 and 2 to be inspected are moved.
From the state of projecting a light spot on 2, the stage on which the wafer under test is supported is driven by -Dx in the x direction, and the light is projected onto the diffraction gratings 23 and 24 under test, and the spot is spotted.
To move to the position of S3, it is sufficient to drive the stage by -Dy in the y direction.

FIG. 4 is a perspective view showing a case where the wafer to be measured is set on the stage for measurement, and the diffraction grating 2 is used.
1 to 26 are formed on a wafer, and a photodetection system 31 is provided above the wafer. The wafer is placed on the y-axis stage 32 and the x-axis stage 33, the y-axis stage 32 is connected to the y-axis stage drive mechanism 34, and the x-axis stage 33 is connected to the x-axis stage drive mechanism 35. Further, the light detection system 31, the y-axis stage drive mechanism 34,
The x-axis stage drive mechanism 35 is connected in a circuit so as to be controlled by the control means 36.

FIG. 5 shows details of the inside of the photodetection system 31. A mirror 42 is provided on the optical path of the two-frequency linearly polarized laser light source 41, and a collimator lens 43 and a prism 44 are sequentially arranged in the reflecting direction of the mirror 42.
A wafer is arranged in the emitting direction of the prism 44. For the purpose of explaining the principle, the wafer is a wafer 47 having diffraction gratings 45 and 46 as shown in FIG. A mirror 48 is provided in the reflection direction of the wafer 47,
In the reflection direction of the mirror 48, a lens 49, a Glan-Thompson prism 50, an edge mirror 51 having a glass portion 51a and a mirror portion 51b as shown in FIG.
2. The photodetectors 13 are sequentially arranged, and the output of the photodetector 53 is connected to the calculator 55 via the phase difference meter 54. Further, the lens 5 is arranged in the reflection direction of the edge mirror 51.
6. The photodetector 57 is arranged, and the output of the photodetector 57 is connected to the phase difference meter 54.

The S-polarized light of frequency f1 and the P-polarized light of frequency f2 emitted from the dual-frequency linearly polarized laser light source 41 are deflected by the mirror 42 and enter the prism 44 through the collimator lens 43. As shown in FIG. 8, the polarization beam splitter section 44a of the prism 44 is set to the frequency of S-polarized light.
The light flux of f1 is reflected, and the light flux of P-polarized frequency f2 is transmitted,
The light is deflected by the reflecting portions 44b and 44c, emitted from the prism 44 as shown in FIG. 9, and enters the diffraction gratings 45 and 46 on the wafer 47 to become diffracted light.

The diffraction gratings 45 and 46 on the wafer 47 are two adjacent linear diffraction gratings at equal intervals formed on the wafer 47 through separate baking processes, and their pitches are both equal to each other. Between the diffraction gratings 45 and 46, as shown in FIG. 6, there is a positional deviation ΔX during printing in the x direction. At this time, the complex amplitude display of the first-order diffracted light Ea (f1) of the left-side incident light of the frequency f1 and the -1st-order diffracted light Ea (f2) of the right-side incident light of the frequency f2 by the diffraction grating 45 is as follows. . Ea (f1) = A ・ exp {i (ω1 ・ t + φ1 + φa)} (9) Ea (f2) = B ・ exp {i (ω2 ・ t + φ2−φa)} (10)

Here, A and B are amplitudes, ω1 and ω2 are angular frequencies, φ1 and φ2 are initial phases of the light flux emitted from the dual frequency linearly polarized laser light source 41, and φa = 2πXa / p. Xa represents the amount of deviation of the diffraction grating 45 from the reference position on the wafer 47 in the x direction.

Similarly, the complex amplitude display of the first-order diffracted light Eb (f1) of the left-side incident light of the frequency f1 and the -1st-order diffracted light Eb (f2) of the right-side incident light of the frequency f2 by the diffraction grating 56 is expressed by the following equation. become. Eb (f1) = A ・ exp {i (ω1 ・ t + φ1 + φb)} (11) Eb (f2) = B ・ exp {i (ω2 ・ t + φ2−φb)} (12)

Here, φb = 2πXb / p, and Xb represents the amount of deviation of the diffraction grating 56 from the reference position in the x direction.

The light beams diffracted by the diffraction gratings 45 and 46 are deflected by the mirror 48, pass through the lens 49 and the Glan-Thompson prism 50, and the Glan-Thompson prism 50 aligns the polarization planes of the diffracted light and interferes with each other. The intensity changes Ia, Ib of the interference lights Ea (f1), Ea (f2), Eb (f1), and Eb (f2),
In other words, the beat signal is given by the following equation. ^{Ia = A 2 + B 2 +} 2A · B cos {(ω1 -ω2) t + (φ1 -φ2) + 2φa} ... (13) Ib = A 2 + B 2 + 2A · B cos {(ω1 -ω2) t + (φ1 -φ2) + 2φb} (14)

The two diffracted interference light beams may be separated as follows. That is, since the two diffracted interference lights are spatially slightly separated from each other, the edge mirror 51 is used in FIG.
The boundary line L between the glass portion 51a and the mirror portion 51b shown in
The diffracted light from the diffraction grating 45 is set to reach the center between the diffracted lights of the two diffraction gratings 45 and 46, and the edge mirror 5
The light reflected by 1 is guided to the photodetector 57 via the lens 56, and the diffracted light from the other diffraction grating 56 is reflected by the edge mirror 5
1 is transmitted and guided to the photodetector 53 through the lens 52.

The beat signals detected by the photodetectors 57 and 53 are introduced into the phase difference meter 54 to detect the phase difference.
When the phase difference is Δφ, Δφ is given by the following equation. Δφ = (φa−φb) = 4π (Xa−Xb) / p (15)

That is, the relative displacement amount ΔX of the diffraction gratings 45 and 46 in the x direction is as follows. ΔX = Xa−Xb = Δφ · p / 4π (16)

Therefore, by inputting the output Δφ of the phase difference meter 54 to the calculator 55 and calculating the equation (16), the relative shift amount of the diffraction gratings 45 and 46 can be obtained. In this way, the grating pattern that has been baked for the first time as the inspection grating pattern, for example, the diffraction grating 45 and the second grating pattern
The grating pattern that has been baked for the second time, for example, the diffraction grating 5
By obtaining the amount of deviation from No. 6, the alignment accuracy of the semiconductor exposure apparatus, that is, the amount of deviation between the actual element patterns formed by the first and second printings can be detected.

As shown in FIGS. 1 and 4, the deviation amount 0 is previously set.
The reference diffraction gratings 25 and 26 set in step 2 are provided in the vicinity of the test diffraction gratings 21 and 22 and the test diffraction gratings 23 and 24 to be measured, and the measurement values of the reference diffraction gratings 25 and 26 are set to values of zero deviation each time. By setting it as a reference value, the distances Dx and Dy are 2
If the value is sufficiently smaller than 00 mm, desired high precision measurement can be achieved.

That is, first, the photodetection system 31 is used to perform measurement with the reference diffraction gratings 25 and 26 to obtain a measured value. Here, as the measured value, the deviation amount of 0 should be originally obtained. However, if a value other than 0 is obtained as the measured value at this point, it can be determined that this value is an error component due to yawing. Therefore, this value is stored in the calculator 55 as a reference value. Next, the stage is moved to such an extent that the stage movement error does not pose a problem, the adjacent diffraction gratings 25 and 26 are arranged at the measurement position instead of the reference diffraction grating, and the photodetector 31 is used to perform this detection. Measure the diffraction grating. The measured value obtained here should include the error component due to the yawing obtained during the measurement of the reference diffraction grating described above. Therefore, the stored reference value is subtracted from this measured value, and the obtained measured value is stored in the calculator 55 as a correct measured value. As a result, it is possible to always obtain a highly accurate measured value with the influence of yawing removed.

For example, if the distances Dx and Dy are both about 20 mm or less, the yawing during a stroke of 200 mm is 1
Even if a stage that has 0 "to 20" exists and an error of 10 nm to 20 nm is generated in a full stroke, the error is limited to an average of 1 "to 2" due to yawing, and this error occurs. That is, the measurement error due to the stage movement from the reference diffraction grating to the test diffraction grating is 1 nm.
The value is as small as 2 nm. Therefore, the measured value of the reference diffraction grating can be sufficiently treated as a reference value when measuring the test diffraction grating.

A pair of lattice patterns simultaneously printed with one mask (reticle) whose displacement amount is known is appropriately arranged on the entire wafer, and the displacement of the test pattern to be evaluated is originally determined. While the grid is being developed and printed on the front surface of the wafer, a grid serving as a reference is provided, and in the process of sequentially measuring the grid of each pattern to be inspected by the step-and-repeat operation, the reference grid placed between the grids is measured. Then, by obtaining the reference value for the next measurement grating measurement, it is possible to suppress the error of the yawing component when measuring the measurement pattern formed in the vicinity of 10 mm to 20 mm of the reference pattern.

FIG. 10 is a pattern layout showing this state, in which a plurality of one-shot sizes 62 are formed on the wafer 61. Further, inside the one-shot size 62, a test diffraction grating 63 and a reference diffraction grating 64, which are printed by using an Nth reticle and an Mth reticle (usually N + 1th reticle), are formed. A reference diffraction grating 65 is formed. By thus providing the reference pattern between the patterns to be inspected, the accuracy can be improved.

Although FIGS. 1 to 4 show only the displacement in the x-axis direction, the displacement in the biaxial directions of x and y can also be measured by arranging them on the x and y axes. is there.

FIG. 11 and FIG. 12 show the second embodiment, and show the relationship between the diffraction grating pattern and the beam when the displacement measurement of the patterns provided on the two objects of the mask and the wafer is performed. 11 and 12 show an example of detecting the deviation in the x-axis direction, in which the inspection diffraction grating 72 is formed on the mask 71 and the inspection diffraction grating 74 is formed on the wafer 73. Is a diffraction grating pattern for detecting the positional deviation between the mask and the wafer. The diffracted light and the subsequent detection system are all omitted. The photodetection system 31 for detecting the light flux diffracted by the diffraction gratings 72 and 74 to be tested is shown in FIG.
The system shown in FIG. 5 is used.

In FIG. 11, reference diffraction gratings 76 and 77.
Is a diffraction grating which is present on the mask 71 and which is created by previously setting the shift amount to 0 and baked by the same mask. Further, the reference diffraction gratings 78 and 79 are on the wafer 73, and their mutual deviation amounts are similarly set to zero. The reference diffraction gratings 76 to 79 are relatively close to the patterns of the diffraction gratings 72 and 74 to be tested, and are provided at a distance of, for example, about 10 mm to 20 mm. In order to move the luminous flux irradiation position from the diffraction gratings 72 and 74 to be tested to the pattern of the reference diffraction gratings 76 and 77 or the reference diffraction gratings 78 and 79, the photodetection system 31 described above may be moved. In that case, since the irradiation position may be changed, only a part of the light detection system may be moved.

Therefore, the present invention can be similarly applied to the measurement of the deviation between the patterns provided on the two objects.
Further, even if the deviation amount of the reference pattern is not 0, the object of the present invention can be similarly achieved by setting a specific known deviation amount in advance. As a result, 0.25 μm
DRAM required for 0.1 μm pattern rule
It is possible to evaluate the printing accuracy of a semiconductor such as (Dynamic Random Access Memory) at a desired 5 nm or less.

[0040]

As described above, the position shift measuring apparatus according to the present invention provides a reference pattern with a known shift amount in the vicinity of the diffraction grating pattern on the object to be inspected, and uses the measured value of this reference pattern. By measuring the deviation of each pattern to be inspected, the measurement of a pattern widely provided in a large-diameter wafer can be realized with high accuracy while suppressing the occurrence of an error caused by the in-plane rotational component of the wafer.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of a pattern shift on one object.

FIG. 2 is a detailed explanatory diagram of a lattice unit.

FIG. 3 is a detailed explanatory diagram of a lattice unit.

FIG. 4 is a configuration diagram of an apparatus according to an embodiment.

FIG. 5 is an explanatory diagram of a light detection system.

FIG. 6 is an enlarged view of a lattice pattern.

FIG. 7 is a relationship diagram between an edge mirror and a lattice pattern.

FIG. 8 is a relationship diagram of light incident on a prism and a wafer.

FIG. 9 is an explanatory diagram of a state of incidence and diffraction on a wafer.

FIG. 10 is a pattern layout diagram on a wafer.

FIG. 11 is an explanatory diagram of a pattern shift on two objects.

FIG. 12 is an explanatory diagram of measurement of a pattern shift on two objects.

FIG. 13 is an explanatory diagram of a relationship between yawing and error generation.

FIG. 14 is an explanatory diagram of a relationship between yawing and the occurrence of an error.

[Explanation of symbols]

21, 22, 63, 72, 74 Test diffraction grating 23, 24, 25, 26, 64, 65, 76, 77, 7
8, 79 Reference diffraction grating 31 Case 32 y-axis stage 33 x-axis stage 36 Control means 61, 73 Wafer 71 Mask

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kenji Saito 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc.

Claims (3)

[Claims] 1. An apparatus for measuring a positional deviation between a plurality of diffraction grating patterns provided on one object or two objects, a coherent light source, a system for making light incident on the diffraction grating, and diffracted light from the diffraction grating. Is used for at least one of the interference lights, a photodetection system including a means for measuring the phase shift of the diffraction grating based on the interference signal, and at least a part of the photodetection system and the object relative to each other. And a means for moving, the displacement amount is known in advance on the object by measuring the reference pattern provided in the vicinity of the diffraction grating pattern for which the displacement measurement should be obtained by the photodetection system, thereby measuring the displacement measurement object. A positional deviation measuring device, which reduces an error due to mutual in-plane rotation between the detection system and the object in a lattice plane. 2. The positional deviation measuring device according to claim 1, wherein the reference pattern is provided within 20 mm of the diffraction grating pattern for which positional deviation measurement is to be performed. 3. The position shift measuring device according to claim 1, wherein the coherent light source is a dual frequency light source, and the phase shift measuring means of the diffraction grating is a heterodyne measuring means.