JPS609126A - Projection type exposing device - Google Patents

Abstract

PURPOSE:To enable to perform a stabilized detection of wafer pattern without generation of an interferential phenomenon on the surface of photoresist by a method wherein the scattered light only coming from the stepped and tilted part of a wafer is made to irradiate invertedly into a projection lens using the illumination having a tilt angle. CONSTITUTION:Wafer patterns 35a and 35b are lighted up by the illuminating devices 200a and 200b having a tilt angle. The stepped inclination parts on the right and left sides of a wafer pattern are illuminated by illumination lights 104a, 104b, 104c and 104d. When the wafer patterns are illuminated as above, no interference fringe pattern is generated. Of the scattered lights 71b and 75b lorated at the left and the right tilted parts to be illuminated, the scattered light 75b only is invertedly made to incident on a projection lens 2, and the scattered light 71b has no relation with the detection of the wafer pattern because it is not invertedly made to incident. The scattered light 75b is image formed in magnified form on a reticle pattern 7b, the enlarged image-formation 35b is turned into the reenlarged image-formation 35b by the relay lens 37b, and it is formed on a slit plate 39b.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The invention relates to a reduction projection type exposure apparatus, particularly to a wafer pattern detection process.

[Background of the invention]

A conventional reduction projection exposure apparatus will be specifically explained based on FIGS. 1 to 8. That is, the conventional reduction projection type exposure apparatus is an apparatus that repeatedly and sequentially exposes a reduced image of a circuit pattern 5 on a glass dry plate (reticle 1) onto a chip 8 on a wafer 3 using a projection lens 2.

The reticle 1 is first aligned with the apparatus optical axis 64 using the reticle alignment mark 6. This is visually observed using a microscope (not shown) whose relative position to the optical axis 64 is fixed.

Further, the relative position of the mark 6 and the circuit pattern 5 is accurately drawn and worked on the reticle 1.

In order to roughly align the wafer 3 and the optical axis 64, a three-point bin 29 is used to fix the wafer outer position.

In addition, if the rough positioning accuracy using only the three-point pin 29 is not sufficient, perform rough positioning using a separate station (not shown) using a microscope, and check the rough positioning accuracy at the exposure station (see Figure 1). It is also possible to hold the

(For example, JP-A-53-105377S). In this way, the wafer 3 is approximately ±10 μm with respect to the optical axis 34.
Rough alignment is performed, and the wafer is placed on the XY stage 300.
mounted on top.

Next, a method for aligning the wafer 3 and the reticle circuit pattern 5 in the θ direction will be described. The position detector 4 detects the amount of positional deviation Δy1 in the y direction between the upper alignment pattern 55b and the reticle alignment pattern 7b.
Detected by 5b. Next, the stage 28α is moved and the chip BXn is placed below the optical axis 34. Therefore, the amount of positional deviation ΔyrL between the pattern 35 and the reticle pattern 7b is similarly detected.

- From the above, θ between the wafer 6 and the reticle circuit pattern 5
The amount of directional deviation Δθ is calculated by the following formula.

The θ stage (28C) corrects Δθ.

Δ02(ΔcarΔyn)/ILp...Equation (1) (
In 1), np represents the distance between chips 8X and 8XrL. That is, P is the chip spacing (10 mm), and n is the number of chips between the chips eX, , sXn.

Next, each chip 8. and reticle circuit pattern 5
Perform directional alignment. For example, chip 8X.

In the alignment, the alignment pattern 35α and the reticle patterns 7α, 35h, and 7b are aligned. 1Ao
In the case of the reduced-light beam, the area of the reticle circuit pattern portion 50 is 1007L771'' for the chip area i0 mm''. In this case, as shown in FIGS. 2 and 6, a transparent window-like pattern 7α,
A 10 times enlarged image of the linear patterns 35α and 35b is formed within the projection lens 76 by the projection lens 2. For example, the size of the S window-like patterns 7α and 7b is 4001gILO1 (the enlarged image size on the reticle is 600μm in length and 50μm in width), and the length of patterns 35α and 554 is 601LmI.
If the rough alignment is within ±10 μm using a width of 5 μm, window-like patterns 7α, 7b inner busy patterns 35α, 3
5b is reliably formed as shown in FIGS. 3 and 4. Here, the amount of deviation Δ between the center line of the window pattern 7α, 7> and the center line of the linear pattern 65a, 55b is:
T,,Δy.

are detected by the detectors 45a and 45b. The deviation amount Δ"11Δy is corrected by moving the stages 28α and 28h.

Next, the configuration of the detector 45α, (45h) will be explained.

Generally, the projection lens 2 is designed for projection exposure ultraviolet light (G&I), and light other than Y-rays causes "chromatic aberration" in which the imaging position changes.

The pattern detection button also uses 9 lines. ! Turn illumination light 70a (7) of i-line illumination fiber 56a (56b)
0b) is reflected by half mirror 12α (12'6),
Lens 18α (18b), mirror 11α (11h), window pattern 7α (7b)'a? Pass through the projection lens 2
It is incident on . The transparent part 400 μmo of the J-shaped part (turn 7a (7b)) is reduced to 40 μm by the projection lens 2.
The pattern 55Q (55b) is illuminated with μmO. Reflected scattered light 7 from pattern 354 (35b)
5a.

(75b) enters the projection lens 2 inversely and forms an enlarged image 5.
Create 5α (55h) in the ninja pattern 7a (7b). Window pattern 7α (7b) and enlarged image 35α (35
b) are superimposed, mirror 11cL (11b), lens 18α (1Bb), relay lens 37αC37h
), the slit plate 3 via the mirror 38z (58b)
An enlarged image is again formed on 9α (39b). Slit 23α (2sb) on slit plate 39α (59b)
) is supported by leaf spring 44α (4416), and gal/(-
41α (41b) and lever 42a (42
b) The connecting rod 43α (43h) vibrates and scans the magnified image of the two. Above the slit 26α (23,6).

Condensing lens 40α, (40b), photoelectric conversion element 25
Since α (25h) is installed, as shown in Fig. 4.

Superimposed enlarged image of both 7a (7b), 35a
(35b) When the slit 23α (23h) passes over the photoelectric conversion element 25α (25b), the output of the photoelectric conversion element 25α (25b) is
) can be obtained as follows. Here, window-like pattern 7α (7b)
Since the positions of the left and right ends of are detected as M, , M from the origin 600 of the slit 23α (2!1,6), the center position of the window is calculated as (M, 1 & 2)/2. on the other hand,
The center position of the turn 35α (55b) is detected as W. The difference Δx and (Δyx) between the two is expressed by the following equation.

Δ-Z"+ (Δ3'l)=CMl+Jf2)/2-F
=121 The principle of pattern detection in a reduction projection exposure apparatus has been explained above.

Next, the interval point in the above detection will be explained. As shown in FIG. 6, the pattern 35α (65b) is composed of linear steps as described above. For example, a layer (Sin, ) 3h constituting a pattern step is formed on a substrate (Si 2 ) 6C, and a photoresist 3α is coated on the pattern step. - Since the grain pressure photoresist 3α is transparent, an interference fringe pattern is generated near the pattern step by illumination with the one-line illumination light 7oα (7ob) as shown in FIG. This is caused by an interference phenomenon (for example, Dit
trick F, Widmann: “QLLant
if, ativt Evalu, atior of ph
otorasist pattern? in, the
1-pm Range J April 1975/
Vol, 14.7f64 Applied, 0pt
ics p931～. (see page 934). When the photoresist is uniformly coated symmetrically with respect to the center of the pattern step, the output waveform of the photoelectric element 25α (25b) becomes symmetrical with respect to the center of the step, as shown in FIG. The center of pattern 35α (35) can be easily and accurately detected by a method (Japanese Patent Application Laid-Open No. 53-69065) of detecting symmetrical tJ, ) by utilizing the symmetry of the output waveform.

However, as shown in FIG. 8, if the photoresist coating is asymmetrical and uneven with respect to the center of the step, it is natural that the photoelectric element 2
5α (The output waveform of 25 is asymmetric. Figure 8 (α)
In the example shown in Fig. 8, since the slope of the step difference is different on the left and right sides, the photoresist 3α is applied asymmetrically, and it is difficult to detect the center of one step difference with the output waveform as shown in Fig. 8(A). , wafer pattern 35α (55b) cannot be detected.

The causes of such asymmetric application of photoresist are (1) in addition to the above-mentioned non-lead step slope, (2) vibration of the rotary coating device (spinner) used in the photoresist coating process, and (3) variation in the viscosity of the photoresist solvent. Appropriate etc.

[Purpose of the invention]

SUMMARY OF THE INVENTION In view of the above-mentioned problems of the prior art, it is an object of the present invention to provide a reduction projection exposure apparatus that prevents the occurrence of interference fringe patterns caused by non-uniformity of photoresist coating and enables stable wafer pattern detection. It is in.

[Summary of the invention]

The present invention uses illumination with an inclination angle of M.

A detection method has been used in which only the scattered light from the wafer step inclined portion is made to enter the projection lens back, and the center of the left and right step inclined portions is defined as the center of the wafer pattern. That is, since the present invention uses illumination having an inclined angle, interference phenomena on the photoresist surface do not occur, and wafer pattern detection can be performed stably.

[Example of invention]

Hereinafter, embodiments of the present invention will be explained in terms of tools based on FIGS. 9 to 26. In other words, the lighting devices 200α and 200b have the inclination angles shown in FIGS. 9 and 10.

Wafer pattern 35α, 35 by 200C9200d
We show how to illuminate h.

Lighting device 200α (200b, 200c, 200d
) is y-line monochromatic source illumination 7 fiber 103c (105b
, 103c, 1'05d).

C? Ivy 105a (105h, 105c, 105
d, ), v:ys102(Z(102h,
102C, 102ct, ), slit 1
01α (101A, 101c, 101d), lens 100α (100b).

100C, 100d) K. The left and right stepped slopes of the wafer pattern 65α (55b) are illuminated by illumination light 1.
04.6, 104d (104α, 104c) K
more illuminated. At this time, patterns 65α, 3
It is important to place 5h at the circuit periphery (scribing area) 8X; Also, the illumination range is 9oα.

(90b) needs to be limited only to the vicinity of the pattern a7.
) for, pickpocket ) 101b-, 1otd (101(Z
, 101c) was used. In addition, when the mercury lamp 54 exposes the top of the pattern 5 chip 8X after pattern detection 1 and alignment, the shutter 105α (105A, 1
osc.

105d,) must be closed.

The pattern 556 is shown in FIGS. 11 and 12 with the illumination light 10.
4α and 104c are shown. In the case of this illumination, the interference fringe pattern seen in FIG.
．． ．． It does not occur at 52 points. Here, the illumination light 104α (10
Scattered light 71 on the left (right) slope illuminated by 4C)
Of A and 75b, only 75h enters the projection lens 2 in a reverse direction, and 71h does not enter the projection lens 2, so it does not participate in wafer pattern detection.

Figure 13 shows the principle of detecting scattered light 7575, Figures 14 to 1
A six-time pressure detection method is shown. Further, in FIGS. 13 to 16, for the sake of simplicity, detection of only the pattern 35b will be described again.

In FIG. 13, the scattered light 75b from the pattern 55b is
A state in which an enlarged image is formed on the reticle pattern 7b is shown.

The enlarged image 35b on the reticle pattern 7b is formed by the lens 18.
z, a re-enlarged image 35b is formed on the slit plate 596 by the relay lens 37b. The above re-enlarged image 65
When the slit 23b scans over b, the detected waveform is as shown in Fig. 14 (
, 6). However, with this method, the window end positions M, , M2 of the reticle window pattern 7b cannot be detected. Therefore, a reticle pattern illumination device shown in FIG. 1 (FIG. 13) is used. The PI device uses illumination fiber 4
8b.

It is composed of a lens 47b, a mirror 46b, and a mirror 10b. Illumination light 60b from the device is regularly reflected by the opaque parts (7 chrome deposited parts) forming the reticle window-like pattern 775.1. The light enters the detection optical system 45b. For this reason, the detected output waveform becomes as shown in FIG.

Figures 15 and 16 explain how to determine the window end positions M, , M, and the step center F using an electric circuit. The output (b) of the photoelectric conversion element 25b is sent to a binarization circuit 80.in which a slice level % is set. input to the binarization circuit 90 in which the signal is set. The rise and fall position calculation circuits 81 and 82 calculate the window end positions M, , M by the output (C) of the binarization circuit 80.
, is calculated. The average value circuit 86 is <M1+ut)/
2, and the output (d) is sent to the position deviation calculation circuit 84.
send to On the other hand, the output of the binarization circuit φ is output from the gate circuit 92.
is input to. The other input of the gate circuit 92 is a position M1+Δ obtained by adding a margin value Δ (50 μrrL on the reticle) to the window left end position M1, and this is the output of the adder circuit 91. In the end, the interval S in which the gate circuit 92 is opened is approximately 600 μm, with the center coincident with the center of the reticle window (signal (
e)).

The gate circuit 920 output Cf) is the wafer pattern 35/)
This is the position of the left and right step portions, and the center position calculation circuit 93°94
The step center position (, V) r * stone is obtained by . The wafer position calculation circuit 95 calculates (tens of digits)/2 (=F) to obtain an output (A). The positional deviation calculation circuit 84 inputs the above-mentioned signals (d) and Ch) and outputs the amount of positional deviation.

17, 18 and 19, the relationship between the tilt illumination angle θ and the detection accuracy of the step tilt center position will be explained. Fig. 17 shows the case where the stepped slope part is illuminated with a relatively large inclination angle θ (=
: θ2). For example, the binarization circuit 90
The output waveform (peak value VMA
When X is 4, the difference # (=4-stone) between the slope center position cage and the position stone according to the detection output signal (!I) is the first
Figure 7.

The result is as shown in FIG. Generally, the debris above the sloped part of the step has a rounded shape, so the experimental results shown in Figure 19 (
Experimental conditions φ-35°m A+ = 500QJ' (S
L02 n = 1.46), ht = 1000OA (
As shown by rL=1.65), the smaller the inclination angle θ, the smaller the detection error ΔF. This error is caused by the wafer position t (F
l+'x)/2, or if the left and right step shapes are different as shown in Figure 8 (α), the inclination angle may affect the detection accuracy of the wafer position grating 'z)/2. It is desirable to have a configuration in which θ is made as small as possible.

Figures 20 and 21 show the inclination angle θ and the output waveform (A) at 5l.
Explain the relationship of IV ratio (contrast) (experimental conditions: = aluminum evaporation 10000λA, = 10000,4
(it=1.65) A1=5000. (). FIG. 20 shows an example in which the step is made of an aluminum vapor-deposited layer 3d. In this case, since the aluminum vapor-deposited surface is rough, the scattered azimuth angle of the scattered light 71A becomes wide, and a component that enters the projection lens 2 in the opposite direction is also included. For this reason, the output waveform (Rino Sl
The N ratio (contrast=Vs/Vo) is significantly lower than in the case of FIG. FIG. 21 shows an experimental example of measuring the LeW ratio of the output waveform at a step formed by the aluminum vapor deposited layer 3d, and it can be seen that in this case as well, it is desirable to have a configuration in which the inclination angle θ is made as small as possible.

In the above, the inclined illumination has been explained as y-line monochromatic light, but even if other grass-colored light (for example, C-scarlet) is used,
Wafer pattern detection becomes possible if a device like "3f" shown in the figure is taken.

In the example shown in FIG. 22, the pattern 65h is
- The illumination lights 704α and 704C use e-win monochromatic illumination. In this case, the photoresist is not exposed due to e-rays, so the slits 101a and 101C are unnecessary. Also, due to the chromatic aberration of the projection lens 2, the repattern stage i3
The enlarged image 5A is formed on seven upper surfaces of the reticle pattern 7b. Therefore, in order to shift the focal point of the lens 18b from the pattern 7b to the A plane, the lens 18A is moved to the position 18 along the original axis of the detection system 45A. By this operation, a re-enlarged image of the pattern 554 is formed on the slit plate 39b. When this operation is performed as described above, the result will be as shown in FIG. 23 (α).

The reticle pattern 7h and the wafer pattern 55b are not detected at the same time, and the reticle pattern 7b's cutting output (23rd
Shown in Figure (A). ) and the detection output of the wafer pattern 35b (shown in FIG. 26(d)) are detected at the excitation time. However, both pattern positions (J/
l+ &! )/2 c shown in Figure 26(c). ), F
(As shown in FIG. 23(g)) can be easily calculated by storing the amount of positional deviation in an electric circuit. At this time,
In detecting the wafer pattern position W, it is necessary to take into consideration the fact that the imaging magnification at point A is slightly larger than 10 times when calculating the amount of positional deviation.

Furthermore, the present invention is not limited to the above-mentioned /io reduction projection exposure.
It is obvious that this method can also be applied to 1/1 equal-magnification projection exposure.

〔Effect of the invention〕

As described above, according to the present invention, there is no interference phenomenon on the photoresist surface, and the wafer pattern can be detected stably.

[Brief explanation of drawings]

Figure 1 is a diagram showing the principle of a reduction projection exposure system and a conventional pattern detection method, Figure 2 is a diagram showing a conventional alignment method between a chip and a reticle circuit pattern, and Figure 3 is a diagram showing the position of Figure 2. Diagram showing the relationship of alignment patterns in the alignment method, No. 4
The figure is a diagram showing a detected output waveform of the alignment pattern of FIG. 3. FIG. 5 is a diagram showing the relationship between the wafer pattern and the interference fringe pattern, and FIG. 6 is a diagram showing the AA cross-sectional view of the wafer pattern shown in FIG. 5 and the generation of interference fringes. FIG. 7 is a diagram showing the detected output waveform of the pattern having a stepped surface in FIG. 6, and FIG. Figure 9 is a diagram showing the detection output waveform of Figure 8 (α), Figure 9 is a diagram showing an embodiment of the reduction projection exposure apparatus of the present invention, and Figure 10 is an enlargement of the chip portion shown in Figure 9. Figure 11 is a diagram showing the illumination method using oblique illumination light of the wafer pattern, Figure 12 is a diagram showing the scattered light due to the illumination method of Figure 11, and Figure 13 is a diagram showing the detection principle of the scattered light of the wafer pattern. This is a diagram showing the 16th
Figure (α) is a side view, Figure 13 (A) is a plan view (only the center), Figure 14 is a diagram showing re-enlarged imaging of the wafer pattern and detection signals, Figure 15 is the Hauer pattern and reticle. FIG. 16 is a diagram illustrating a flow for correcting pattern misalignment, and FIG. 16 is a diagram illustrating an electric circuit for implementing the flow shown in FIG. 15. Fig. 17 is a diagram showing the detection error of the center of the step inclination when the inclination angle θ is large, and Fig. 18 is a diagram showing the detection error of the center of the step inclination when the inclination angle θ2 is small.
Fig. 9 is a diagram showing the relationship between the inclination angle θ and the detection error of the step inclination center in a certain experimental example, Fig. 20 is a diagram showing scattered light from a wafer pattern formed with an aluminum vapor deposited layer and its detection output, and Fig. 21 The figure shows the detection output S/N ratio measurement result of Fig. 20 in an experimental example, Fig. 22 shows the detection principle when e#il is used for oblique illumination light, and Fig. 26 shows the detection output S/N ratio measurement result of Fig. 20. It is a figure which showed the detection output in the case of a figure. 1... Reticle 2... Reduction projection exposure lens 34... Optical axis 45α (45b)... Detection optical system - 1icz (11b)... Mirror 18α (18,6)... Lens 37α (57b) Relay lens 48α (48A) Reticle pattern illumination fiber 36α (56b) Wafer pattern illumination fiber 54 Mercury lamp 3α Photoresist coating layer 3b Step formation Layer 3C...Substrate 75α(75,6)...Reflected light from the substrate 200α(
20OA, 200C, 200d)... Inclined illumination device 102α (1025, 102C, 102d)... Lens 101α (101b, 101C, 101d)...
...Slit 100α (100b, 100C, 1oo
d)... Lens 105 (Z (105A, 105
cm, 105d) -Shutter agent patent attorney Takashi
Akio Hashi-'5 Figure 〒7 Figure 11 ÷ Z (t) Araki 10 Figure 11 Figure 5b Kayu 1 δ Figure 19 Figure 20 Figure Kabuto 21 □ θ

Claims (2)

[Claims] (1) A plurality of illumination devices each having an optical axis at an inclination angle of 40 degrees or less in a direction perpendicular to both sides of each pattern on a plurality of linear step patterns of a semiconductor wafer, and scattering light of the illumination light from the illumination devices from the pattern steps. A projection exposure lens optical system for making an enlarged image of the pattern step by inputting the projection exposure lens, and a detection optical system for detecting the position of the enlarged image formed by the projection exposure lens optical system. A distinctive projection exposure device. (2) The reduction projection type exposure according to claim 1, wherein the illumination device uses monochromatic illumination light for wafer exposure, and is configured to limit the wafer illumination range only to the vicinity of the pattern. Device.