JPS627129A - Exposure and exposure system with projection exposure device - Google Patents

Abstract

PURPOSE:To alleviate the reduction in the accuracy of lapping due to the distortion, by obtaining the absolute value of the deflection in the amount of the distortion that is obtained in the arbitrary, plural image height positions within the specified exposure range and by adjusting the one of the magnifying power of projection out of the first and second projection exposure devices so that the error will be minimum. CONSTITUTION:After the exposure of the first layer onto a wafer W has been performed with a stepper A, the specified process E for forming the first layer is performed to perform the lapping exposure for the second layer with a stepper B. For this, a main control device CNT2 for the stepper B feeds the data DS1 about the distortion of a projection lens PL1 from a main control device 1 of the stepper A. The data D21 is stored in the main control device CNT1 as the inspection data at the time of the production of the stepper A beforehand. Similarly, the data DS2 concerning the distortion of the projection lens PL2 of the stepper B is stored in the main control device CNT2 and fed to a main control device CNT1 as the data when the magnifying power of the stepper A is to be adjusted.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field of the Invention) The present invention relates to (4) a method for performing step-by-step processing using a plurality of units (at least two units) during overlapping exposure between different layers on a substrate (semiconductor wafer, etc.) for IJ lithography; The present invention relates to an exposure method and an exposure system when an and-repeat type exposure device and a stepper are used together.

(Background of the invention) In recent years, reduction projection exposure equipment has been used in the production of VLSIs.
It has achieved great success in improving productivity.

This type of projection exposure equipment is used to increase the precision of overlapping between different layers on a semiconductor wafer during exposure, as the line width of the pattern on the reticle to be projected exposed becomes finer and the pattern itself becomes more highly integrated. It is required to improve accuracy.

This is especially true between different devices. Recently, various types of projection exposure apparatuses with different numerical apertures, magnifications, or exposure areas (projection fields) have been developed, and in VLSI manufacturing factories, the required resolution and throughput are taken into account during the process of manufacturing one VLSI. Therefore, it is increasingly common to use separate devices for exposure between different layers. The so-called mix and match is the mixing of exposure devices with different projection magnifications and projection fields, or devices with different exposure methods such as refractive systems or reflective systems in the photolithography process of semiconductor manufacturing. Ma
tch) In case of "4", overlay accuracy is particularly important. Distortion (image distortion) of the projection optical system is cited as a factor that affects overlay accuracy.

General K.

The current situation is that even projection optical systems with the same structure have slightly different distortion characteristics (aberration curves), and even more so between projection optical systems with different magnifications and fields of view.
Distortion characteristics may vary significantly. Therefore, even if mix-and-match is performed using such devices, sufficient overlay accuracy may not necessarily be obtained. This poor overlay causes serious problems such as decreased productivity.

(Objective of the Invention) An object of the present invention is to provide an exposure method and an exposure system that solve the above-mentioned problems and reduce the reduction in overlay accuracy due to distortion between different projection exposure apparatuses.

(Overview of the Invention) The present invention provides distortion characteristics (aberration -11i1) within a predetermined exposure range from the field center (optical axis) of a first projection exposure device to a predetermined image height position, and mix-and-match For example, when comparing the distortion characteristics within the predetermined exposure range of the second projection exposure device used for overlay exposure in such cases, while relatively changing the projection magnification of the projection optical system of both devices, the predetermined exposure range Find the absolute value of the deviation of the amount of distortion (deviation amount from the ideal grid point) obtained at any number of image height positions in In other words, the projection magnification of at least one of the first and second projection exposure devices is set to vI4 so that the distortion error is minimized.
The technical point is to make sure that the

(Example) Next, an exposure method according to an example of the present invention will be explained based on FIG. FIG. 1 is a perspective view showing the schematic configuration of two projection type II optical devices &1 (hereinafter referred to as steppers) A and H having different projection field sizes.

In this example, the stepper person's projection lens PLI has a reduction magnification of 115 and a projection field of view on the wafer W of 20X.
The projection lens PL2 of the stepper B has a reduction magnification of 1/2.5 and the projection field on the wafer W is 30 x 301 m square (a circular area with a diameter of 42 m).
(circular area). A blind BLI is provided as a field diaphragm that illuminates only the pattern area FAI to be exposed on the stepper AKFi reticle R1, and the shape and size of the opening are adjusted by four movable blades. Reticle R1 is mounted so that the center of pattern area PAI coincides with the optical axis (field center) AX of projection lens PLI. Now, the Kueha WJd loaded on stepper A, the stage ST that moves two-dimensionally,
pattern area P by the projection lens PLI.
The projected images of the AI are sequentially exposed onto the wafer W by the step-and-metal method. The projection lens PLI in this embodiment has a pressure for finely controlling the magnification of the projection lens PLI by controlling the pressure (atmospheric pressure)' in the air space between the plurality of lenses constituting the projection lens. A controller BCI is provided and operates to obtain a desired projection magnification according to setting information from the main braking device CNTl. This pressure h
Regarding the specific configuration, operation, and usage of the u unit BCI and the main control i&placement CNT1, please refer to Japanese Patent Application Laid-Open No. 60-28613.
Since it is disclosed in detail in Japanese Patent Publication No. 60-78454, the explanation thereof will be omitted here.

On the other hand, stepper B also differs only in the size of the projection field of projection lens PL2, except for the other basic configurations of blind BL2. Stage ST, Pressure regulator BC2
, the main controller CNT2 is the same as that of stepper A. Further, the pattern area PA2 of the reticle R1 mounted on the stepper B is the same as the pattern area FA of the reticle R1.
It has the same dimensions as I. That is, the size of the pattern area PAI is the maximum field of view (2
If the projection lens PL2 is 100 x 1'00 m square corresponding to 0
3Gm), which will be used in a 20x20m square area. In this embodiment, such a stepper A
and B are used together for overlapping exposure between different layers formed on the wafer W.

For the sake of simplicity, here, after exposing the first layer of the wafer W using the stepper A, a predetermined process E for forming the first layer is performed, and the second layer is superimposed on the wafer W. It is assumed that stepper B performs the combined exposure. For this reason, the main controller CNT2 of the δ chipper B manually inputs the data DS1 from the main controller CNTl of the stepper A to the distone 1 of the projection lens PLI. The data DS1 is stored in advance in the main controller CNTl as inspection data at the time of manufacturing the stepper A. Alternatively, as disclosed in Japanese Patent Application Laid-Open No. 60-18738, a photoelectric sensor with a slit is provided on the stage ST, and cross-shaped marks formed at multiple positions on a reticle for distortion inspection are projected. By scanning the slit within the image plane and determining the projection position of each mark, the amount of distortion at a plurality of points within the projection field of view may be detected and stored as data DSI.

If you use the slit on the stage like this,
At any time during the operation of the stepper, the distortion at that point can be accurately measured, making it possible to deal with changes in the distortion characteristics over time. Similarly, data DS2 regarding the tone 1 of the projection lens PL2 of the stepper B is stored in the main controller CNT2 and sent to the main controller CNTl as data for adjusting the magnification of the stepper A.

In addition, if the main controllers aicNT1 and CNT2 are centrally controlled by a higher-level control device (large computer), the distortion data DSI of the projection lens PLl and the distortion data of the projection lens PL2
The same effect can be obtained even if both the data DS2 and the data DS2 are sent to a higher-level control device and centrally managed there.

Now, FIG. 2 is a chart showing an exaggerated example of the overlay accuracy (hereinafter referred to as matching index) when two steppers A, 132) perform overlapping exposure without performing magnification correction. In the same figure, the solid line is Ste.
This is a chart showing the deviation of bar A from the ideal grid point, and the broken line is a chart showing the deviation of stepper B from the ideal grid point.

The center of the projected field of view of stepper A and the center of the projected field of view of stepper B are the origin P of the orthogonal coordinate system X7.

In the first quadrant of the coordinate system , the projection points of the ideal grid points by stepper B are p, b
.

p wb , p , b . . . . As is clear from this example, two steppers are centered (
Optical axis) When you mix and match, the center P.

Although sufficient matching accuracy can be obtained in the vicinity, the center Po
Points away from, for example, points Pea and p, b, or point pea
and Plb, the relative amount of deviation may become so large that it cannot be ignored. This amount of deviation depends on the detone characteristics between the two projection lenses, and does not necessarily increase at each point in the periphery.

FIG. 3(a) shows an example of the distortion characteristic of the projection lens PLI, which is stored as data DSI in the main controller CNTl, and FIG. 3(b) shows an example of the distortion characteristic of the projection lens PL2.
An example of the distortion characteristics of l), main controller CN
It is stored in T2 as data DS2. Figure 3 (a
) and (b), the vertical axis represents the amount of distortion, and the horizontal axis represents the body height (the distance in the radial direction from the optical axis, that is, the center Po). The sign of the distortion amount is negative when the projection point of the ideal lattice point is shifted toward the center PaK, and positive when the opposite is the case. Although the distortion characteristic curves of both projection lenses have similar overall trends, the amount of distortion at the same image height position differs greatly depending on the image height. In this type of projection lens, distortion correction is performed so that the maximum value of the distortion in the positive direction and the maximum value in the negative direction of the distortion are approximately equal. This is to improve (distribute) the distortion evenly over the entire exposed area.

FIG. 4 shows the above two distortion characteristics superimposed on each other with their centers aligned, and it is only necessary to consider the area within the projection field of view (20 x 20 inches) of stepper A. ag
In Figure 4, stepper A.

It is assumed that the projection magnification in both B is at a standard value (initial value). As is clear from this figure, the absolute value of the deviation between the distortion characteristics DSI and DS2 is extremely large from the point where the image height is about 9 degrees to the periphery, and the absolute value of the deviation ΔD is 1ΔD1 at the image height position Px. It is maximum. This deviation 1
The amount of ΔL)l is the distortion characteristic DSI or DS2
The maximum value of the amount of distortion for any of the characteristics may also be larger. Therefore, for patterns located at an image height of approximately 9 m1E or more, sufficient matching accuracy cannot be obtained. The case shown in FIG. 4 corresponds to the chart shown in FIG. 12.

Therefore, in this embodiment, the distortion characteristic DSI of stepper A is left unchanged, and the distortion characteristic DS2 of stepper B is changed by finely adjusting the magnification, and when the two distortion characteristics are superimposed, the deviation that occurs within the projection field of view is The maximum absolute value of ΔD is made to be the minimum.

Therefore, a mathematical analysis for superimposing the two distortion characteristics and maximizing the matching accuracy will be described below. The following analysis is based on stepper B's main controller CNT.
2 or - can be easily executed by a computer such as a higher-level control device.

Assuming that the image height is r, the distortion curve HtfCr), and the coefficient of magnification adjustment is tC, the amount of deviation ΔCγ) from the ideal grid point obtained by adjusting the magnification is expressed as in equation (1).

Δ(γ)−J(γ)+C−γ・・・・・・・・・
(1) Corresponding to this equation (1), the deviation from the ideal lattice point for steppers A and H is calculated by vector (ΔXa
When expressed as lΔYa) and (Δxb, ΔYb), respectively,
Equation (2) for stepper A and Equation (3) for stepper B are as follows (ΔXa, Δy a )
= (f a (JF Takumi) +・・・・・・・・・(
2) (Δxb, ΔYb) = (fb(Jσ'?) +...
...(3) However, subscripts a and b correspond to steppers A and B, respectively.

Therefore, the final distortion deviation vector (ΔX, ΔY) when the steppers A and B are aligned in the center and subjected to overlapping exposure is expressed as in equations (2) and (3)) and as in equation (4).

(ΔX, ΔY) = (ΔXa1ΔYa) - (ΔXb, ΔY
b)・・・・・・・・・(4) According to equation (4), the absolute value of the deviation ΔX, the absolute value of the deviation ΔY, and the scalar amount of the vector (ΔXlΔY) in the overlapping exposure area (projection field of view) Let the maximum values of Dx and D be respectively
When y%D, the following equations (5), (6), and (7) are obtained.

DX=MAX(lΔXl) ・・・－・・・(5)
D y=MAX(IΔYl) ・・・・・・・・・
(6) B=MAX (J jX”+ jY”)
””・(7) Therefore, these D x s D y s
Minimize any one of D, or D'', D
The magnification adjustment coefficient Ca% is set so as to minimize both y.
By determining cb, the optimum conditions can be obtained. Incidentally, when the magnification correction is performed only by the stepper B as in this embodiment, the coefficient Ca may be set to zero. Therefore, (2) above,
From equations (3) and (4), the deviation vector (ΔX, ΔY) is expressed as shown in equation (8).

(ΔX, ΔY) = (f a (V stone F) − f
b (Mi ear 1) (8) (However, in this embodiment, Ca = 0.) Therefore, the main controller CNT2 determines in advance each of the multiple points within the projection exposure area. Memorized distortion percentage DS
Vector (ΔX, ΔY) and maximum value Dx while changing adjustment coefficient cb using I (function fa) and DS2 (function fb).

D7. Calculate D, and from the accuracy required for the layers to be overlaid and exposed, Dx%D7. Any one of D or DX,
Find a coefficient cb such that both D>' are minimum. Then, the main controller 11cNT2 applies a command value to the pressure regulator BC2 to correct the magnification of the projection lens PL20 by a value corresponding to the coefficient Cb. Pressure regulator BC2
The pressure is controlled by taking into account the projection lens shop PL2's irradiation history and magnification fluctuations due to atmospheric pressure fluctuations, as well as the command value.

In the above analysis, we tried to obtain a precise solution through mathematical analysis in order to improve the matching accuracy over the entire exposure area. You can get the solution without any problem. Therefore, an analysis using a distoshicon curve will be explained with reference to FIG. Figure 5 shows the distortion characteristics DS
It shows how I and DS2 are superimposed while changing the magnification of the projection lens PL2, and is represented by the first ditone of the projection lens PL2 after the magnification adjustment. First, the main controller CNT2 calculates the difference in the amount of distortion between the characteristic DSI before magnification adjustment and DS2 at multiple points (for example, every 0.5 m) from the center P0 to an image height of 151 m. Accordingly, as shown in FIG. 4, the maximum absolute value of the deviation ΔD is found.

In order to reduce the absolute value of this deviation ΔD, the magnification of the projection lens PL2 is corrected as shown in FIG.
It is better to lift (tilt) the image in the positive direction at a specified high point.

Center P in Figure 5. The straight line l passing through corresponds to the original image high axis of the characteristic DS2, and the slope cbFi of the straight line l corresponds to the magnification adjustment coefficient C in equation (1) above.

The main control unit RCNT2 calculates the characteristic D S 2' after magnification adjustment based on (1) 2 with the slope Cb increased by a certain amount from zero. Then again, the absolute ra of the deviation between the corrected characteristic DS'2' and the characteristic DSI is
v Calculate for each image height position, and find the deviation with the maximum absolute value. The above calculation is repeated by increasing the slopes c and b' by a certain amount. Eventually, tlcs
As shown in the figure, the deviation 1ΔD,l near the position RPX is also small (as 1ΔD1 in FIG. 4), but on the contrary, the deviation IΔD,1 at another image height position becomes large. In the case of the characteristics shown in Figure 5, if the characteristic DS2' is raised a little more from the state shown in the figure, 1ΔD, 1>1ΔD
On the other hand, if you lower it a little more, 1ΔDwl <
It becomes IΔD11. Therefore, in the example shown here, the slope cb is set so that 1ΔD1=1Δawl.
It's good to keep it constant. For example, when the inclination is Cb, the image height is 15M! The straight line obtained at the position! The amount of deviation ΔM from the image high axis is the amount of deviation caused by magnification correction on the actual projection image plane. In this way, two distortion characteristics are compared while changing the relative magnification, and at each magnification, the maximum value of the deviation (absolute value) of the amount of distortion is determined, and the magnification that minimizes that maximum value is determined. Accordingly, the best matching accuracy can be obtained over the entire exposure area within an image height of 15 m.

FIG. 6 is a chart showing an example of a matching state when the magnification is adjusted using K as described above and overlapping exposure is performed using stepper BK. In the figure, the solid line is for stepper AK, and the broken line is for stepper B. Compared to the chart shown in Figure 2, the projection points of the ideal grid points (P, ll and pub, p=a and Ptbs
It can be seen that the respective deviation amounts of Pj' and P3b) are small on average, and the matching accuracy is improved.

As shown in FIG. 1, linear deformation (expansion/contraction) may occur in the wafer W that has been subjected to the process E tube after exposure with the stepper A due to thermal or chemical treatment. Therefore, when overlapping exposure is performed by stepper B, it is desirable to perform magnification correction in consideration of the deformation. To do this, for example, the positions of alignment marks formed at multiple points on the wafer W are measured using an alignment device equipped with a stepper BK, and each position a is compared with the design value.
The amount of deformation is determined.

Next, other embodiments of the present invention will be described. If the stepper A shown in Fig. 1 does not use the entire 20m square projection field as the exposure area, or if you want to accurately align the absolute magnification tube of the exposure area on the wafer W, use the stepper BK.
In consideration of overlapping exposure, it is preferable to adjust the magnification of the stepper A in advance. Although this is not a practical method of use, for example, suppose that the exposure area of the first layer by a stepper is narrowed down to an image height of 11 m.
If the magnification of only stepper B is adjusted as in the previous embodiment, the degree of matching on the detone characteristic will be biased in the negative direction as a whole, as shown in FIG. This becomes a big problem when trying to manage the absolute magnification on the wafer W.

In Fig. 7, the image height positions where the absolute value of the difference in the amount of detone between characteristics DS2' and DS1 is maximum are a point around 8u and a point around 11m, and the deviation in this vicinity is If the absolute values of ΔD4 and ΔDa are made equal, the absolute value of the maximum deviation amount is minimized. However, when overlapping within the exposure area at a height of about 11 mt, the distortion eye is biased in the negative direction, so the characteristics DSL and D
If S2 is raised in the positive direction by the same value at a certain image high point, it becomes possible to distribute the detone 1 into positive and negative directions within that exposure area.

Fig. 8 shows that from the superimposed state as shown in Fig. 7, 1ΔD
ll = llD41 and the characteristics under the optimum conditions where boost-silane distribution is performed. In Figure 8, straight line no! is the same as the straight line l in Figures 5 and 7, and is a straight line! , represents the original high axis tilted by a coefficient of 01 due to the magnification adjustment of stepper A, and the characteristic DS
L' represents the detone of the projection lens PLI after magnification adjustment. In order to obtain such a characteristic, first, while keeping the characteristic DS1 fixed, the characteristic DS2 is raised, that is, the magnification is adjusted so that the absolute values of both the deviations ΔD8 and ΔD4 become equal, in the same manner as in the previous embodiment. Find the coefficient cb.

Then, on the superimposed distortion characteristics, draw a straight line so that the maximum value of the positive distortion amount and the maximum value of the negative distortion amount are equal in absolute value! 1
, h have the same slope.

Change only the amount. The inclination C& of the straight line determined in this way is the magnification adjustment coefficient of the stepper A. Therefore, when exposing the first layer, for example, at an image height of 10 m, +Δ
It is sufficient to use the stepper A with a magnification offset of MJL, and when performing overlapping exposure of the second layer, it is sufficient to use the stepper B with a magnification offset of +ΔMb at the image height of 10g+1. Performing such magnification correction not only reduces the relative distortion difference between different layers, but also reduces the absolute destage amount (magnification error) of the pattern area formed on the wafer W. can also be reduced.

In each of the embodiments of the present invention described above, mix-and-match of steppers having different projection field sizes was considered. However, even if the projection lenses have the same projection field size, their distortion characteristics usually differ slightly due to differences in lens construction, manufacturing errors, and irregularities. Therefore, an example in which the sizes of the projection fields of view are the same will be explained with reference to FIG. 9. FIG. 9(a) shows a characteristic in which the distortion characteristic DS2 of one of the two steppers and the distortion characteristic DS3 of the other are superimposed without adjusting the magnification. Here, it is assumed that the characteristic DS2 has poor distortion among the same steppers, and the characteristic DS3 has good distortion. For example, if these two steppers are used for an exposure area with an image height of 16 am, the absolute value of the difference in the amount of distortion is 11 + s for the image height.
Near the maximum value ΔD, the distribution of distortion becomes somewhat unbalanced. Therefore, in such a case, consider the distribution of distortion,
The best matching accuracy can be obtained by raising both the characteristics DS2 and DS3 in the positive direction to form DS2' and DS3' shown in FIG. 9(b). The characteristic DS2 is changed to the characteristic DS2' by adjusting the magnification corresponding to the slope of the straight line It.
The characteristic DS3 changes as shown in the characteristic DS3' by adjusting the magnification corresponding to the slope of the straight line Im. In this way, the point with an image height of 16 inches and the point near 8n can have the maximum deviation, but by minimizing the maximum absolute value of these deviations, the relative difference between the two strippers The difference in destination amount is the 9th
Each stage is smaller than the case in Figure (a).

Although each embodiment of the present invention has been described above only with respect to a stepper equipped with a reduction projection lens, the present invention can be similarly implemented in combination with a reflective projection type stepper using a mirror. Furthermore, in each of the above embodiments, matching was only considered for two steppers, but matching can be similarly performed for any n number of steppers more than that. In this case, the matching accuracy is calculated on a host computer using the distortion characteristics of each of the n steppers, and each stepper is Just adjust the magnification. Furthermore, if the object (reticle) side of the projection lens is a non-telecentric optical system, the magnification can be adjusted by finely adjusting the distance between the reticle and the projection lens, and the same effect can be obtained. In this case, a driving means for vertically moving a reticle stage holding a reticle by a minute amount along the optical axis is provided as a magnification adjusting means.

Further, in the embodiment described above, two distortion characteristics are compared, for example, every 0 and 5 taxes on the image high axis, and the maximum absolute value of the deviation is determined. However, in this case, within a predetermined exposure range, 0.
Since the maximum value is found after determining the deviation at all image high points for every 5 images, the accuracy is low and there is some error with respect to the true maximum value. For this reason, if the points to be compared are made even finer by 0.5 points, the accuracy will increase considerably, but the calculation processing time will increase proportionately.

Therefore, a method for quickly obtaining an accurate maximum value from two distortion curves through mathematical analysis will be briefly described below.

Now, if the respective distortion characteristics of the two steppers are fa(r) and fb(γ), then the magnification adjustment coefficient CJI
Distortion function Fa(r), F considering %Cb
b(r) is expressed by equations (9) and (10), respectively.

Fa(r)=fa(r)+Ca11r ==・-(
9) Fb(r)=fb(r)+Cb-r ------
・Cl0) Here, the characteristics fa(γ) and fb(γ) are approximated by Na-order curves and Nb-order curves, respectively, and are controlled by the main control device of the chipper or the upper-level combination. It is assumed that it is memorized. Characteristics fa(r), fb(r
) is usually designed so that when r=oo, fa(r)=0, fbcγ)=0 so that the detone characteristics of this type of projection lens are 71 line symmetrical with respect to the optical axis. Therefore, the general equation can be expressed as equation (11), and there is no constant term.

f(r)=Knor"+Kn-□sr"-'+-・-・
−+,・r,・r・・・・・・・・・(
1F) Here, Kn, Kn-4+...K1, K,
is a constant.

Therefore, two distortion functions FIL(γ) and F
If the difference in b(7') is expressed by a function G(γ), then the function G(r
) is expressed as in equation (12).

G(r)=Fb(r) −Fa(r) =fb(γ)−faCr)+(Cb−Ca) ・r・・
・・・・・・・・・(12) Here, in order to find the maximum and minimum points of the function G(γ) with c=c b−c a, the function G(γ) is imaged as shown in equation (13). Differentiate with high r.

C (13) where Kan represents the constant of the nth term of the characteristic fa(γ), and Kbn represents the constant of the nth term of the characteristic fbcr).

Then, the absolute value of the deviation is maximum at either of the maximum and minimum points of the function G(r'). In addition, when Na-Nb are both 3rd order or less, equation (13) becomes 2
The solution can be found analytically as the following equation, but N
If any one of a, N, and b is quartic or higher, the solution cannot be found analytically, so the solution must be found by numerical calculation.

Let γ1, rl, ..., rn and the maximum image high point in the exposure area be rmax, then each value is (12
) Formula % Formula %) Calculate each value of G(γmaw). Then, from among these calculated values, choose the one with the maximum absolute value and set it as Gmax, then Qmax is expressed as in equation (14).

Qmax=Max(IG(7B)l, IG(γ,)1,
...IG(rn)1, IG(rmax)l)
・・・・・・・・・(14) However, among rl, γ, ・・・・・・rn, rma
x) is also large or has a negative value. By performing K in this way, the maximum value of the absolute value of the deviation can be uniquely determined without having to compare every 0.5 degrees on the image height axis. The above calculation is performed using the relative magnification difference C(
The optimum magnification adjustment coefficients Ca and Cbf can be determined by repeatedly performing calculations while changing Cb-Ca) by a minute amount.

Based on the above method, an example in which both Na and Nb are tertiary will be described below. Therefore, the characteristics fa(r), fb(r
) is determined as shown in equations (15) and (16).

j'a(r)=Ka, ・r'+Ka, ・r"+Ka, -
r −・−・・(15) f b (r )=Kbr r
"+Kbc r"+Kbt·r . . . -" (16) Therefore, the function G(r) of the difference between the functions FB(r) and l'b(γ) is expressed as in equation (17).

G (r) = (Kb, -Ka,) ・r"+c Kb
, -Ka,) ・r"+(Kb, -Ka,)-r+(C
b-Ca)-r (17) When formula (17) is differentiated with respect to r, formula (18) is obtained.

G'(r)=3(Kbs Ka,)・7"+2(1(
b, ni&t)・rten(Kb,-Ka,)+(Cb-Ca
)・・・・・・・・・(18) Here, Kb, −Ka, = K,, Kb! -Kal=K
t, ・Kb, -Ka, = K,, Cb-Ca=
C, the solution that satisfies G'(r)=0 is given by the following equations (19) and (20), where r and γ
2.

Therefore, by changing the value of C by a minute amount, (19), (20
) and calculate γ1. Find γ, and then use equation (17)! ! 7. Iteratively calculates the absolute values of G(r,), G(r,), and G(rmax). At this time, γm
, γ is both larger than rmax, or γ
1. γ! When both are negative values or complex numbers, IG(rmax)l is the maximum value (IGmaxl) corresponding to the value of C at that time.

In this way, lQmaxl calculated each time the value of C is changed
Compare all values of l
The CO value corresponding to Qma:cl K is the solution to be found.

Therefore, Ca and Cb may be determined so as to satisfy the value of C.

(Effects of the Invention) As described above, according to the present invention, the matching precision during overlapping exposure using two projection exposure devices is improved, and the productivity of semiconductor devices with patterns with finer line widths is improved. The effect of a significant improvement can be obtained. Furthermore, when matching multiple exposure devices, in addition to controlling the distortion characteristics of each individual device, it was also necessary to strictly control the distortion characteristics between the devices from the manufacturing stage to minimize the distortion characteristics between the devices. . However, according to the present invention, adjustment can be made to obtain the best matching accuracy within the projection exposure range used for exposure, so there is no need to limit the use of exposure devices that were made with matching in mind from the beginning. Moreover, it is also possible to obtain the effect that a plurality of devices can be combined relatively freely on a semiconductor device manufacturing line.

[Brief explanation of drawings]

FIG. 1 is a perspective view showing the schematic configuration of two projection exposure devices suitable for the mix-and-match method according to an embodiment of the present invention, and FIG. Chart diagrams exaggerating the deviation of the projected points of ideal grid points when superimposed on the
) are the projection lens disc-71 of the two exposure devices, respectively.
7 curves, Figure 4 is a characteristic diagram when the Dist-717 curves in Figures 3 (a) and (b) are superimposed, and Figure 5 is a characteristic diagram showing the characteristics shown in Figure 4 with magnification adjustment. Figure 6 shows the deviation of the projected point of the ideal grid point when overlapping exposure is performed after adjusting the magnification.
: An exaggerated chart diagram, Figure 7 is a characteristic diagram showing the matching when changing the image height with the characteristics shown in Figure 4, and Figure 8 is a diagram showing the characteristics of two exposure devices based on the characteristics shown in Figure 7. Fig. 9 is a characteristic diagram showing the case where the magnification of both is adjusted to obtain the best result.
FIGS. 7A and 7B are characteristic diagrams for explaining improvement in matching accuracy when exposure apparatuses have the same projection field size. [Explanation of symbols of main parts] A...First projection exposure apparatus, B...
Second projection exposure apparatus, E...Wafer process, R,, R1...Reticle, PLI, PL2...Projection lens, W...
・Wafer, BCI, BC2...Pressure regulator (magnification adjustment section)
, CNTl, CNT2... Main controller, DSI
, DS2, DS3...distortion characteristics, DS1', DS2', DS3'...distortion characteristics after magnification adjustment.

Claims (4)

[Claims] (1) When a projection-type first exposure device and a second exposure device are used for overlapping exposure between different layers on a photolithography substrate, the centers of their projection fields are made to substantially coincide on the substrate. When using them separately, the distortion characteristics within the exposure range from the center of the field of view to a predetermined image height position of the first exposure device and the distortion characteristics within the exposure range of the second exposure device are determined based on the relative relationship between both devices. Projection exposure characterized in that the projection magnification of at least one of the first exposure device and the second exposure device is adjusted so that the overlay accuracy is the best in terms of both distortion characteristics when compared while changing the projection magnification. Exposure method using equipment. (2) When comparing the distortion characteristics of the first exposure device and the second exposure device while changing the relative projection magnification, the difference in the amount of distortion obtained at any plurality of image height positions within the exposure range 2. The method according to claim 1, wherein a maximum absolute value among the absolute values of is determined, and a relative projection magnification is determined so as to minimize the absolute value. (3) In order to keep the absolute magnification of the exposure area on the substrate almost constant, while maintaining the determined relative projection magnification, the projection magnification of the first exposure device and the second exposure device is further adjusted. 3. A method according to claim 2, characterized in that they are adjusted together by approximately the same amount. (4) a projection-type first exposure device that exposes a desired pattern on a substrate for photolithography; and the first exposure device that superimposes and exposes a new pattern on the pattern formed on the substrate. and a projection-type second exposure device different from the first exposure device, in which the projection field centers of both devices are made to substantially coincide on the substrate for overlapping exposure, wherein Distortion characteristics within the exposure range up to the image height position and the second
Comparing the distortion characteristics of the exposure devices within the exposure range, adjusting the projection magnification of at least one of the first exposure device and the second exposure device so that the overlay accuracy is the best in terms of both distortion characteristics. An exposure system characterized by comprising a magnification adjustment means.