JPS63298102A - Position aligning method - Google Patents

Abstract

PURPOSE:To correct the deflected amount of position alignment, by inputting S or P polarized laser light vertically with respect to a position aligning mark, measuring the rotary angle of the polarization plane of the reflected laser light, and computing the asymmetry of the mark. CONSTITUTION:An optical system is composed of a position aligning mark 11 on a wafer, P or S polarized laser flux 12, an analyzer 13 and a photodetector 14. The P or S polarized laser light 12, which is emitted from a light source, is vertically inputted into a positioning lattice 11. The rotary angle of the polarization plane of the light, which is diffracted with the positioning lattice 11 on the wafer, is detected by using a light polarizing element 13, which is provided in the parallel direction and in the vertical direction with respect to the positioning lattice 11 on the wafer. The correcting value based on the asymmetry of the positioning lattice 11 is computed by using the rotary angle of the polarization plane. The wafer is moved with respect to the correcting value, and the position is aligned.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention is a micropattern with a fine pattern.

1. Prior Art - Semiconductor devices have recently become more and more densely packed, and the dimensions of the fine patterns of each element are now 1 micron or less.

Conventionally, the alignment of the photomask 1 and the LSI wafer during LSI manufacturing was done by overlapping the mark on the wafer with the mark on the 5 wafer 1, but the alignment accuracy was ± It is about 0.3 microns, and when forming submicron elements, the alignment accuracy is poor and it is not practical. Also, S. Austin, [Abrid Physics Letters (Appl.
ied Physics Letter) Vol 3
147P, 428.1977), as shown in FIG. 6, an incident laser beam 1 is made incident on a photomask 2, and a grating 3 formed on the photomask 2 is The diffracted light is diffracted again by the grating 6 formed on the wafer 4'', thereby producing diffracted light 6. A. Obtain 8... When this diffracted light is expressed as a binary representation of the diffraction order at the photomask and the diffraction order at the wafer, one diffracted light 6 is (o,
1), diffracted light 7 is (1,1), diffracted light 8 is (-1,2)
It can be expressed as... This diffracted light is collected at one point by a lens and the light intensity is measured.

The diffracted light has a light intensity at a position symmetrical to the incident laser beam 1, and the photomask 2 and the wafer 4 can be aligned by matching the intensities of the diffracted light observed on the left and right sides. In this method, the alignment accuracy is said to be several 10oA.

However, in this method, since the alignment between the photomask 2 and the wafer 4 is greatly influenced by the spacing between the photomask 2 and the wafer 4, accuracy in the spacing is required. Furthermore, it is necessary to bring the photomask 2 and the wafer 4 close to each other and align them while maintaining an accuracy of five intervals, which complicates the apparatus and poses a problem in practical use.

In addition, for alignment of elements with submicron line width,
There is a method of observation based on secondary electron emission from the device, but since it cannot be handled in the atmosphere, the throughput for manufacturing LSI becomes small, which poses a practical problem.

In addition, the conventional example shown in FIG.
transon E, D,) ED-26, 4, 1974
, 723, GijBouwhu is :) In this case, the -2 beam is incident on the Fourier transform surface of the microlens shown by the two lens systems Ll and L2, and it is formed on the wafer via the lens L2. Illuminate the beam l onto the grating and apply the spatial filter SF
Only the ±1st-order light diffracted from the grating by the lens system L2. L
1 onto the reticle R, generates an interference pattern in the vicinity of the reticle R, and detects the light passing through a grating provided on the reticle with a photodetector to align the wafer W and the reticle R. Illustrated. In the configuration shown in Figure 6, it is stated that it is difficult to correct the position of the asymmetric grating formed on the wafer W, and this can be achieved in all processes and marks in the alignment mark manufacturing method. It is impossible and has not been put into practical use.

Problems to be Solved by the Invention As described above, the conventional alignment method has the problem that alignment cannot be performed if the alignment marks are asymmetrical.

In view of these conventional problems, the present invention has been developed to provide a method for aligning marks formed on a wafer, even if the alignment marks are asymmetrical.
The purpose of this invention is to provide an alignment method that can detect and correct such asymmetry.

Means for Solving the Problems In order to solve the above problems, the present invention makes S or P polarized laser light perpendicularly incident on the alignment mark, and rotates the plane of polarization of the laser light reflected from the wafer surface. The asymmetry of the alignment mark is calculated by measuring the angle with a polarizer, and the amount of misalignment based on the asymmetry of the alignment mark is corrected.

Function Conventionally, it was not possible to correct asymmetry when forming alignment marks, and two-position alignment was carried out in a state that included the asymmetry of the alignment marks.
The alignment error was large. With the alignment method according to the present invention, it is possible to correct the asymmetry of alignment marks, and achieve alignment with higher precision.

Embodiment FIG. 1 shows an embodiment that briefly illustrates the principle of the present invention. The optical system of the present invention has an alignment mark 1 on a wafer.
It is composed of a 1% P or S polarized laser beam 12, an analyzer 13, and a photodetector 14. In the following explanation, for the sake of simplicity, the zero-order term, which is specularly reflected light from the pattern surface, of the diffracted light from the alignment mark 11 will be explained. 1st
As shown in the figure, the incident light AI is
2, the reflected light B15, and the incident angle φ are given by the following equation.

B = A -2 (AoN) N ・--
---(1) A -, N= l A l ・l N 1
cos (yr - φ) = -cos φ... (2) Also, the change in polarization that occurs upon reflection is due to Fresnel's law of reflection, and due to reflection at the incident angle φ, the polarization E8 perpendicular to the incident plane
The polarized light Ep parallel to , becomes R8 and Rp after reflection, respectively.

IR,l=l S(φ)IIE,I...
...(3)lRpl=IP(→IIEpl
... sum) where the coefficient S (<6)? p
(→ is given by the following formula using the refractive index n of the pattern.

”(1'1=-6in(φ-φ')/sin (
φ＋φ勺・・・−−−(5)P(→−
jan (φ−φ勺／jan(φ＋φ勺...
...(6) Nsinφ' = sinφ
... Song... (7) Also, what is the relationship between polarized light E and H?

Rx=S(<6)EB! -P(φ) (α to Epx-2)
・・・・・・・・・(8) R A S (A E8 A P (φ)
(β to Ep lure) Decoy...Medium (9) R2: s(φ)
E8z-P(φ) (γ to Epz-2)・Song・(10)
Here, E=E8+Ep=[E8x, Esa, EIl□]
+ [Ep, Ep, Ep,] R = [R, x, Ry, R
2], to=Epxa+”pyβ+Ep2γN=[α,β
, γ].

Therefore, when an S-polarized laser beam is irradiated, the polarized light R becomes Rx = ``(nail 2/ (α2 + γ2)-
2-1)/(α2+γ2) ・・・・・・・・・・・・
・・・・・・・・・・・・(11) Ra=2Q→αβ
...Curved surface/during the song...(12) R2 knee S(φ)αγ
/(α2+γ2) -P(→αγ(2β2-1)/(α
2+γ2) ・・・・・・・・・・・・・・・−・
(13) Now, when the laser beam is perpendicularly incident on the alignment mark as shown in Fig. 1, the normal vector of the pattern surface is Co, o, 1], and if the inclination angle is φ in a portion where the pattern cross section is inclined, the normal vector of the inclined 9-Heno plane becomes (sinφ, o, cosφ). Therefore, the polarization of the reflected light from the alignment mark depends on the inclination angle of the reflective surface of the alignment mark, and if it is not inclined, Rx=S(φ) Ry=o R2=o −−−−−−
-(14) When the inclination angle is φ, S('4 j p(<s) is calculated by Fresnel's law of reflection, 5°, and equation 6.7, so the reflected light from the alignment mark is By detecting only the polarized light components in two directions with a polarizer, the inclination angle φ of the reflective surface of the alignment mark can be calculated.Therefore, the correction value for the amount of misalignment based on the asymmetry of the alignment mark is Depth of Alignment Mark Next, the procedure for applying the method of correcting the asymmetry of the alignment mark on a wafer according to the present invention to actual alignment will be described. This is a positioning method that can effectively implement the correction method.The light 21 emitted from the light source is incident on the entrance pupil of the microlens through the illumination optical system.

17th-+) 2- Place the reticle 23 at the front focal point f1 of the conversion lens 22. The reticle 23 has a pattern to be projected and a period P1 formed within the pattern.
It consists of a phase grating. As shown in FIG. 2, when parallel incident light 21 enters the reticle 23, the light is diffracted into a plurality of light beams by the phase grating inside the pattern, and the light beams enter the first Fourier transform lens 22, and each light beam enters the rear focal plane of the first lens. A Fourier spectrum image corresponding to the diffracted light is formed. Second
As shown in the figure, a spatial filter 5F24 is arranged on this Fourier spectral plane to block 0th-order and ±2nd-order or higher diffracted light, and allow only ±1st-order light to pass through.

This diffracted light passes through the second Fourier transform lens 25 and is projected onto the wafer 26 placed at the rear focal plane f2.The image projected onto the wafer approximately focuses the image of the aperture on the reticle and ± The first-order light components interfere with each other, and a new 1/1/゛ interference fringe is formed within the open pattern at a pitch of P2.

Therefore, if a grating 27 with a pitch that is an integral multiple of the pitch P2 of the interference fringes is formed on the wafer, diffracted light can be obtained from the grating by wavefront-splitting each of the incident ±1st-order lights U, , U2. With this light, light intensity information indicating the parallelism between the interference fringes of the two light beams and the grating 27 and the relative positional relationship in the pitch direction is obtained.

FIG. 3 is a diagram schematically showing misregistration due to differences in the shape of the alignment grating on the wafer when the above-mentioned alignment method is used. When the grating grooves are symmetrical as shown in Figure 3a, the center of the interference fringes and the center of the grating groove are aligned, but when they are asymmetrical as shown in Figure 3a, the alignment position is Since the position S is shifted from the center T of the designed position H, this results in a large alignment error.

Therefore, according to the present invention, a P- or S-polarized laser beam is incident perpendicularly to the alignment grating. Then, the rotation angle of the polarization plane of the light diffracted by the alignment grating on the wafer is detected using optical photometric elements installed in a direction parallel to and perpendicular to the alignment grating on the wafer. Then, a correction value based on the asymmetry of the alignment grating shown in FIG. 3 is calculated based on the rotation angle of the polarization plane, and the wafer is moved and aligned with respect to this correction value.

Next, a second embodiment of the present invention is shown in FIG.

S- or P-polarized laser light 63 is obliquely incident on alignment grating 67 on the wafer. Then, the light reflected in the direction perpendicular to the alignment grating is detected by an analyzer 64.
The S-wave component and the P-wave component are separated and detected by the detector 66.

And, as shown in the above embodiment, the P wave of the diffracted light.

Asymmetry of the alignment mark on the wafer from the S-wave component,
That is, the inclination angle of the alignment mark is calculated to determine the amount of correction for misalignment, and the substrate is moved relative to this correction value to achieve highly accurate alignment.

Next, a third embodiment of the present invention is shown in FIG.

S or P polarized laser light 73 is incident on the alignment grating 72 on the wafer 1 from diagonally symmetrical directions. and the light 7 reflected vertically by the alignment grating.
6 is separately separated into a P wave component and an S wave component by an analyzer 73 and detected by a detector 74. As shown in the embodiment, the asymmetry of the alignment mark on the wafer is determined from the P wave and S wave components of the diffracted light, that is, the inclination angle θ1 of the alignment mark. θ2 is determined, the amount of correction for misalignment is calculated, and the substrate is moved in accordance with this correction value to achieve highly accurate alignment.

Effects of the Invention As described above, according to the present invention, the reticle pattern is aligned on the wafer with high precision using interference fringes, and
Furthermore, the shape of the grating formed on the wafer is estimated, and if the shape is asymmetrical, the position can be corrected to achieve more accurate alignment.

[Brief explanation of the drawing]

Fig. 1a is an enlarged perspective view of the X portion of Fig. 1a, and Fig. 2 is an enlarged perspective view of the X section of Fig. A configuration diagram of the alignment method that is effectively applied, Figure 3a
FIG. 3 is an explanatory diagram showing the alignment state of interference fringes and grooves when the groove shape of the grating according to the present invention is symmetrical, and FIG. An explanatory diagram showing the alignment state, FIG. 4 is a second configuration diagram of a device for correcting misalignment of the present invention, and FIG. 6 is a third configuration diagram for correcting misalignment of the present invention. 6A and 6B are diagrams showing the principle of alignment using the conventional double grating method, and FIG. 7 is a diagram showing the configuration of the conventional alignment between a reticle and a wafer using interference fringes. 11...Positioning mark, 12...S
or P-polarized laser incident light A, 13...analyzer,
14...Detector, 16...Reflected light B0 Name of agent Patent attorney Toshio Nakao and 1 other person

Claims (4)

[Claims] (1) S- or P-polarized laser light is irradiated perpendicularly to the alignment mark on the wafer, and the plane of polarization of the laser light reflected from the alignment mark on the wafer is rotated using an optical polarizer. The inclination angle of the cross-sectional slope of the alignment mark is determined by determining the angle, and the amount of alignment deviation based on the asymmetry of the alignment mark is calculated from the step of the alignment mark and the inclination angle, and the correction value is calculated. An alignment method in which alignment is performed by moving the board against the substrate. (2) The alignment method according to claim 1, wherein a light polarizing element is arranged in parallel and perpendicular directions to the alignment grating, and the rotation angle of the polarization plane of the reflected light from the alignment grating is determined. (3) The alignment method according to claim 1, wherein the S- or P-polarized laser beam is incident on the alignment mark on the wafer from an oblique direction. (4) S- or P-polarized laser light is irradiated onto the alignment mark on the wafer from two directions, and the rotation angle of the polarization plane of each laser beam reflected from the alignment mark is determined. 2. The alignment method according to claim 1, which corrects the amount of alignment deviation based on the asymmetry of the upper alignment mark.