JPH07130622A - Shaping method for electron beam section at position of aperture for angle stop of electron beam exposure system - Google Patents

Abstract

PURPOSE:To conform, in a short time, the section of an electron beam to a circle whose diameter is equal to a set value, by obtaining a first driving signal for making the section circular and making the section diameter nearly equal to an aperture diameter, on the basis of the sectional form of the charged particle beam, and supplying the signal to a corrector. CONSTITUTION:An electron beam EB which has passed an aperture 34a is cast and converged on a semiconductor wafer 10, and detected as a current I by a current detector 40. An astigmatism correction circuit outputs a signal proportional to a driving signal to be supplied to an astigmatism corrector 36, in order to make the transversal section of the electron beam EB at the position of an angle stop plate 34 circular. A signal proportional to a driving signal to 'be supplied to an image deflection corrector 38 is outputted, in order to make the transversal section diameter of the electron beam EB at the position of the angle stop plate 34 nearly equal to the diameter of the aperture 34a.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for shaping an electron beam cross section at the position of an aperture for an angle diaphragm of an electron beam exposure apparatus using a stencil mask.

[0002]

2. Description of the Related Art With the miniaturization of LSI circuit elements, an electron beam exposure apparatus using a stencil mask is being developed. In this device, an electron beam passing through the optical axis is swung by a two-stage mask entrance side deflector to pass through a selected passage hole pattern on a stencil mask, and further two stages of a mask exit side deflector are used to swing back the electron beam. To pass on the optical axis, and then through the aperture for angle diaphragm, and then converge on a semiconductor wafer to expose. According to such an exposure apparatus, it is possible to draw a fine pattern on a wafer with an electron beam having an extremely small number of shots as compared with an exposure apparatus having a variable rectangular electron beam cross section.

[0003]

However, the crossover image, which is a cross section of the electron beam at the position of the aperture, generally becomes elliptical as shown in FIG. 7A due to astigmatism and field curvature, and the aperture is reduced. The amount of electron beam passing through is limited, which increases the exposure time. In view of such a problem, an object of the present invention is to provide an electron beam cross-section shaping method capable of making a cross section of an electron beam at a position of an aperture for an angle diaphragm into a circular shape and a set diameter in a short time. is there.

[0004]

[Means for Solving the Problem and Its Action] FIG. 1 shows the principle of the present invention. The present invention relates to a charged particle beam cross-section shaping method at the position of the aperture 1a for an angle diaphragm of a charged particle beam exposure apparatus, in which the charged particle beam exposure apparatus forms charged particles in a selected passage hole pattern on the stencil mask 2. The cross section of the charged particle beam EB is shaped through the beam EB, and the aperture 1a formed in the diaphragm plate 1 is converged and irradiated onto the sample 3, for example, a semiconductor wafer or a mask through the charged particle beam EB, and the aperture is also formed. A deflector 4 for deflecting the charged particle beam EB and a corrector 5 for shaping the cross section of the charged particle beam at the position of the aperture 1a by an electric field or a magnetic field in response to the first drive signal S1 are arranged above 1a. Has been done.

In the present invention, (A2) while changing the second drive signal S2 supplied to the deflector 4 to scan the charged particle beam EB on the diaphragm plate 1, it is blocked by the diaphragm plate 1 or passes through the aperture 1a. The cross-sectional shape of the charged particle beam EB is measured by detecting the electric current I of the charged particle beam EB, and (A5) the cross section is made circular based on the cross-sectional shape, for example, from a previously given relational expression. A first drive signal S1 for making the cross-sectional diameter substantially match the diameter of the aperture 1a is obtained, and (A6) the obtained first drive signal S1 is supplied to the corrector 5.

In the present invention, the cross-sectional shape of the charged particle beam EB at the position of the aperture is measured, and based on this, the first driving for making the cross-section circular and making the cross-sectional diameter substantially coincide with the diameter of the aperture 1a. Since the signal S1 is obtained, the number of times of the iterative processing in the correction is relatively smaller than that in the method of repeating the trial and error many times based on only the current I without obtaining the cross-sectional shape of the charged particle beam EB. The cross section of the charged particle beam at the position of the aperture can be made circular and have a set diameter in a relatively short time.

In the first aspect of the present invention, from (A3) the cross-sectional shape of the charged particle beam EB, three characteristic quantities L1 and L of the shape as shown in, for example, FIG.
2 and L3 are extracted, and (A4) the first drive signal S1 is changed, and the change amount ΔS1 of the first drive signal S1 and the feature amounts L1 and L
2, the relationship with the amount of change of L3 is obtained, and (A5) the first drive signal S1 for making the cross section circular and making the cross section diameter substantially coincide with the diameter of the aperture 1a is obtained from the relationship.

In the case of this configuration, the first drive signal S1 is changed to change ΔS1 of the first drive signal S1 and the characteristic amount L1,
Since the relationship with the amount of change in L2 and L3 is obtained, it is not necessary to previously give a complicated empirical formula in which the driving signals to the ambient temperature and various deflectors and magnetic field lenses are used as parameters, and the first driving can be performed easily and accurately. The optimum value of the signal S1 can be obtained.

In the second aspect of the present invention, the deflector 4 is disposed above the stencil mask 2 and the charged particle beam EB is used.
To pass through the selected passage pattern with the stencil mask 2 substantially perpendicular to the surface, and (A1) supply the deflector 4 so that the amount of the charged particle beam EB passing through the aperture 1a becomes a maximum value. The second drive signal S2 is adjusted (A2 to A6), and then the process of the present invention is performed.
(A7) Next, the charged particle beam E passing through the aperture 1a
The second drive signal S2 supplied to the deflector 4 is readjusted so that the amount of B becomes the maximum value.

In the case of this configuration, the optimum values of the first drive signal S1 and the second drive signal S2 which are related to each other can be obtained efficiently, easily and quickly. Conversely, if the optimum values of the first drive signal S1 and the second drive signal S2 that are related to each other are to be obtained simultaneously, the processing becomes extremely complicated and the required time becomes long.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 4 shows the main configuration of the electron beam exposure apparatus. The semiconductor wafer 10 as a sample is a scanning X-
It is mounted on the Y stage 12. XY stage 1
2 above the magnetic field lenses 14, 16, 18, 20 along the optical axis C perpendicular to the upper surface of the XY stage 12.
And 22 are arranged, the electron beam becomes a parallel bundle by these magnetic field lenses as shown by the chain double-dashed line, and the semiconductor wafer 10 is converged and irradiated by repeating convergence and divergence.

A stencil mask 24 is provided between the magnetic field lenses 14 and 16 through which the collimated electron beam passes.
The stage 25 is attached and arranged. The area on the stencil mask 24 is the area E1 as shown in FIG.
1 to E55, and through-hole patterns (not shown) are formed in a matrix in each area. Magnetic field lens 1
4 and the stencil mask 24, two stages of mask incident side deflectors 26 and 28 are arranged, and between the stencil mask 24 and the magnetic field lens 16 two stages of mask emitting side deflectors 30 and 32. Are arranged. The electron beam EB passing on the optical axis is deflected by the deflector 26, is deflected by the deflector 28, is made incident on the stencil mask 24 substantially perpendicularly, passes through the selected passage hole pattern, and is deflected by the deflector 30. It is swung to the axis C side, is swung back by the deflector 32, and passes substantially on the optical axis C.

The passage hole patterns selectable by the two-stage mask entrance side deflectors 26 and 28 are within one area shown in FIG. 9B, and the passage hole patterns in other areas are X-.
It can be selected by driving the stencil mask 24 on the Y stage 25 at the pitch of the area arrangement. An aperture 34a is provided between the magnetic field lens 18 and the magnetic field lens 20.
The angle diaphragm plate 34 in which is formed is arranged so that the center of the aperture 34a coincides with the optical axis C. Angle diaphragm plate 3
The crossover image CE1 which is the cross section of the electron beam EB at the position 4 is generally elliptical due to astigmatism and field curvature, as shown in FIG. 7 (A). Astigmatism and field curvature are corrected to make the crossover image CE1 circular, and
By making its diameter substantially equal to the diameter d0 of the aperture 34a, in order to eliminate the underexposure, astigmatism corrector 36 and field curvature correction are provided between the stencil mask 24 and the magnetic field lens 16 as shown in FIG. A container 38 is arranged.

FIG. 5 shows the principle configuration of the astigmatism corrector 36 and the field curvature corrector 38. The astigmatism corrector 36 is
A pair of coils 361 and 362 wound concentrically with the X axis
And a pair of coils 363 and 3 wound concentrically with the Y axis.
The field curvature corrector 38 has a coil 64 that is wound concentrically with the optical axis C. Here, the X axis and the Y axis are axes orthogonal to each other in the plane perpendicular to the optical axis C. For simplicity, the coil 361 and the coil 362 have X = 0, Y
= 0, the directions of the magnetic fields formed by them are opposite to each other and the magnitudes of the magnetic fields are equal to each other, so that equal currents are supplied to the coils 363 and 3 similarly.
The same currents are supplied to 64 so that the directions of the magnetic fields formed by X = 0 and Y = 0 are opposite to each other and the magnitudes of the magnetic fields are equal to each other.

In FIG. 4, the electron beam EB which has passed through the aperture 34a is converged and irradiated onto the semiconductor wafer 10 and detected as a current I by a current detector 40 connected between the semiconductor wafer 10 and ground. On the one hand, the detected current I is transmitted through the differentiator 42 to the signal processing circuit
To the signal processing circuit 44 on the other hand. Further, in order to detect the irradiation position of the electron beam EB on the semiconductor wafer 10, a secondary electron detector 46 is arranged above the semiconductor wafer 10, and its output is supplied to the signal processing circuit 44.

In FIG. 4, a blanking deflector 47 is for oscillating the electron beam EB and for pulsing the electron beam EB in relation to the angle diaphragm plate 34. The scanning sub-deflectors 49 are for scanning the electron beam EB within a predetermined large area and small area on the semiconductor wafer 10, respectively. These deflectors 47, 48 and 49 are the deflectors 26, 28 and 3 to be described later.
0, 32, they are not used when adjusting the drive signals to be supplied to the correctors 36 and 38 and are turned off.

Deflectors 26, 28, 30, 32, corrector 3
The drive signal generation circuit for 6 and 38 is shown in FIG. A pattern code Ci for selecting the i-th through hole pattern on the stencil mask 24 is supplied to the address input end of the pattern coordinate memory 50. The coordinate (Xi, Yi) of the i-th passing pattern is stored in the address Ci of the pattern coordinate memory 50. The coordinates (Xi, Yi) read from the pattern coordinate memory 50 in response to the pattern code Ci are the mask area correction circuit 5
2 and is converted into signals (Xi1, Yi1).
The mask area correction circuit 52 is the X-axis of the electron beam exposure apparatus.
-When the Y coordinate system and the XY coordinate system of the stencil mask 24 match, it functions as a unit matrix,
Xi1 = Xi and Yi1 = Yi.

The signals (Xi1, Yi1) are converted by the conversion circuit 5
4, the astigmatism correction circuit 56 and the field curvature correction circuit 58 are supplied to a pair of input ends. The conversion circuit 54 uses the signal (X
In response to i1, Yi1), the electron beam EB should be supplied to the deflectors 28, 30 and 32 in order to pass the electron beam EB substantially perpendicularly to the i-th through hole pattern of the stencil mask 24 and swing back on the optical axis C. A signal proportional to the drive signal (Xi2,
Yi2), (Xi3, Yi3) and (Xi4, Yi4)
Is output. The astigmatism correction circuit 56 uses the signal (Xi1,
In response to Yi1), the astigmatism corrector 3 is used to make the cross section of the electron beam EB at the position of the angle diaphragm 34 circular.
A signal (SXi, SY) proportional to the drive signal to be supplied to
i) is output. The field curvature correction circuit 58 uses the signal (Xi
1, Yi1), the cross-sectional diameter of the electron beam EB at the position of the angle diaphragm plate 34 is determined by the diameter d of the aperture 34a.
A signal Fi proportional to the drive signal to be supplied to the field curvature corrector 38 in order to substantially match 0 is output.

The signals Xi1 and Yi1 are amplified by power amplifiers 61X and 61Y, respectively, and become driving voltages VXi1 and VYi1 which are supplied to the X-direction portion and the Y-direction portion of the deflector 26. The signals Xi2 and Yi2 are amplified by the power amplifiers 62X and 62Y, respectively, and the drive voltage VXi2
And VYi2 are supplied to the X-direction portion and the Y-direction portion of the deflector 28. The signals Xi3 and Yi3 are amplified by the power amplifiers 63X and 63Y, respectively, and the drive voltage VX
i3 and VYi3 are supplied to the X-direction portion and the Y-direction portion of the deflector 30. The signals Xi4 and Yi4 are amplified by power amplifiers 64X and 64Y, respectively, and become driving voltages VXi4 and VYi4, which are supplied to the X-direction portion and the Y-direction portion of the deflector 32. The signals SXi and SYi are amplified by the power amplifiers 65X and 65Y, respectively, and converted into electric currents, and become driving currents ISXi and ISYi which are supplied to the X-direction portion and the Y-direction portion of the astigmatism corrector 36. The signal Fi is amplified by the power amplifier 66 and converted into a current, which is supplied to the field curvature corrector 38 as a drive current IFi.

The conversion circuit 54 outputs (Xi2, Yi2).
Is approximated by the cubic polynomial of the input (Xi1, Yi1). That is, Xi2 = C2 (1,1) + C2 (1,2) Xi1 + C2 (1,3) Yi1 + C2 (1,4) Xi1Xi1 + C2 (1,5) Xi1Yi1 + C2 (1,6) Yi1Yi1 + C2 (1,7) Xi1Xi1Xi1 + C2 (1,8) Xi1Xi1Yi1 + C2 (1,9) Xi1Yi1Yi1 + C2 (1,10) Yi1Yi1Yi1 (1) Yi2 = C2 (2,1) + C2 (2,2) Xi1 + C2 (2,3) Yi1 + C2 (2 , 4) Xi1Xi1 + C2 (2,5) Xi1Yi1 + C2 (2,6) Yi1Yi1 + C2 (2,7) Xi1Xi1Xi1 + C2 (2,8) Xi1Xi1Yi1 + C2 (2,9) Xi1Yi1Yi1 + C1 (2,1) Y2,10Y2,1 And Here, 20 C (1,1) to C (2,1)
0) is a constant. The expressions (1) and (2) are collectively expressed simply as (Xi2, Yi2) = C2 · (Xi1, Yi1) (3). Where C2 is a constant matrix,
To be exact, (Xi1, Yi1) on the right side is (1, Xi1,
Yi1, Xi1Xi1, Xi1Yi1, Yi1Yi1,
Xi1Xi1Xi1, Xi1Xi1Yi1, Xi1Yi
1Yi1, Yi1Yi1Yi1), but for simplicity, this is expressed as (Xi1, Yi1). This point is the same below.

The outputs (Xi3, Yi3) and (Xi4, Yi4) of the conversion circuit 54, the output (SXi, SYi) of the field curvature correction circuit 54 and the output Fi of the astigmatism correction circuit 54 are input (Xi1, Yi1). ) And the output of the mask area correction circuit 52 (Xi
1, Yi1) is approximated by a cubic polynomial of the input (Xi, Yi), and these are simply expressed as (Xi3, Yi3) = C3 · (Xi1, Yi1) (4) (Xi4, Yi4) = C4 * (Xi1, Yi1) ... (5) (SXi, SYi) = CS * (Xi1, Yi1) ... (6) Fi = CF * (Xi1, Yi1) ... (7) (Xi1, Yi1) ) = CM · (Xi, Yi) (8)

Mask area correction circuit 52, conversion circuit 5
4, astigmatism correction circuit 56 and field curvature correction circuit 58
Can be configured using an adder, a multiplier, and a register for holding a constant, as is clear from the above equation. A constant is set in this register by the signal processing circuit 44 of FIG. It is assumed that the optimum values of the conversion matrices CM, C2, C3, C4, CS and CF at a certain time point (previous time) have already been obtained. After that, the conversion matrix CM,
FIG. 2 shows a process performed by the signal processing circuit 44 when it is necessary to correct C2, C3, C4, and CS. Hereinafter, the numerical value in the parenthesis represents the step identification number in the figure. In addition,
Except for steps 74 and 75, the stencil mask 24
Is processed in the state without.

(70) First, the conversion matrix C2 is corrected so that the crossover image CE1 is obtained as shown in FIG.
As shown in. The correction of the conversion matrix C2 is performed as follows. The conversion matrices CM, C2, CS, and CF are left as the previous proper values, and C3 = C4 =
0 (deflectors 30 and 32 are turned off). Even if the deflectors 30 and 32 are turned off, as is clear from the chain double-dashed line in FIG. 4, the electron beam incident on the magnetic lens 16 in parallel with the optical axis C passes through the aperture 34a.

FIG. 9 in the plane of the stencil mask 24.
The pattern code C corresponding to the point of i = i0 shown in (A)
i0 is supplied to the pattern coordinate memory 50. The signal processing circuit 44 changes the constant term C2 (1,1) of the above equation (1) little by little, for example, five times in the vicinity of the previous value, and each C2 (1,1) is changed.
1) That is, for each Xi2, the current I is detected by the current detector 40, and Xi2 that maximizes the current I is obtained from the graph of Xi2 and I. Similarly, the constant term C2 (2,1) of the above equation (2) is changed, the current I is detected by the current detector 40 for each Yi2, and the Yi2 at which the current I becomes maximum is found from the graph of Yi2 and I. Ask.

Such processing is performed for 10 points of i = i0 to i9 shown in FIG.
i (1) is the maximum I (Xi2, Yi2) is 10
A set is obtained, and from these values, 20 constants C (1,1) to C (2,10) in the above equations (1) and (2), that is, C2 in the above equation (3) are determined. This completes the correction of the conversion matrix C2. At this time, the cross section of the electron beam EB at the position of the angle diaphragm plate 34 becomes a crossover image CE1 shown in FIG. 7 (A).

(71) Transformation matrix CS and CF
Is corrected as follows. Conversion matrix CM,
With CS and CF remaining at the previous proper values, the conversion matrix C2 is set to the proper value obtained as described above, and C3 =
FIG. 9 is a view in the plane of the stencil mask 24 with C4 = 0.
The pattern code C corresponding to the point of i = i0 shown in (A)
i0 is supplied to the pattern coordinate memory 50. Then, the processing shown in FIG. 3 is performed.

(710) The initial value 0 is substituted for the variable j indicating the number of times of loop processing below. (711) (Xi2, Yi2) is scanned to measure the crossover image of the electron beam EB at the position of the stencil mask 24. For example, the differentiator while continuously changing the constant terms C2 (1,1) and C2 (2,1) in the above equations (1) and (2) near the current values A0 and B0 as in a raster scan, respectively. By reading the output of 42, the cross-sectional shape of the electron beam EB is measured. That is,
For each value of C2 (2,1) that changes by ΔY within the range of B0−D ≦ C2 (2,1) ≦ B0 + D, C2 (1,
1) is continuously changed from A0-D to A0 + D, and {C2 when a pulse is output from the differentiator 42 at this time
(1,1), C2 (1,1)} is stored as the edge position of the crossover image CE1 shown in FIG. 8 (B). Then, as shown in FIG. 8A, C2 (1,1) corresponding to each of the two current differential pulses with respect to one X-direction scanning line c.
The difference between the values of is obtained as the width wy across the crossover image CE1 in the X-axis direction.

(712) The feature amount of the cross-sectional shape of the crossover image CE1 is extracted. The crossover image CE1 is generally an ellipse, and since the ellipse is determined by three parameters, this feature amount is also three. The feature amount of the crossover image CE1 is, for example, L1, L2, and L3 shown in FIG.
And Here, L1 is the maximum value of the length that crosses the crossover image CE1 in the X-axis direction. L2 is an interval in the Y direction between the pair of scanning lines a and b when the length across the crossover image CE1 in the X axis direction is the set value W. Also,
L3 is the Y-direction distance between the midpoint of the line segment of the scanning line a that crosses the crossover image CE1 and the midpoint of the line segment of the scanning line b that crosses the crossover image CE1.

FIG. 8C shows another method of extracting the feature amount of the crossover image CE1, where L1 is the major axis length of the crossover image CE1 and L2 is the short axis length of the crossover image CE1. It is the axial length, and L3 is the angle of the long axis of the crossover image CE1 with respect to the X axis. (713) When j = 0, the process proceeds to step 715, and when j ≠ 0, the process proceeds to step 714.

(714) Feature quantity L1 = L when j = 0
These change amounts ΔL1ij, ΔL2ij and ΔL3ij for 1i0, L2 = L2i0 and L3 = L3i0 are calculated and stored. (715) If i <3, the process proceeds to step 716, j =
If it is 3, the process proceeds to step 718.

(716) 1 is added to the variable j. (717) SXi, SYi, and Fi are set to the previous proper values SX.
The predetermined minute amounts ΔSXij, ΔSYij, and ΔFij are changed from i0, SYi0, and Fi0, respectively. Then, the process returns to step 711. (718) Here, the change amount of the feature amount (ΔL1i, ΔL2
i, ΔL3i) is the change amount (ΔSXi, ΔSYi, Δ) of the output of the astigmatism correction circuit 56 and the field curvature correction circuit 58.
Fi) is linearly approximated, and the inverse relational expression between the two is simply expressed as (ΔSXi, ΔSYi, ΔFi) = Cα · (ΔL1i, ΔL2i, ΔL3i) (9).

This conversion matrix Cα is given by (ΔL1i
j, ΔL2ij, ΔL3ij) and (ΔSXij, ΔSY
ij and ΔFij) determined from three sets of values for j = 1, 2, 3. Then, (SXi, SYi, Fi) for making the crossover image CE1 into a circular crossover image CE2 having a diameter d0 equal to that of the aperture 34a is obtained using the equation (9) as follows.

(SXi, SYi, Fi) = (SXi0, SYi0, Fi0) + (ΔSXi, ΔSYi, ΔFi) = (SXi0, SYi0, Fi0) + Cα · (D0-L1, (D02-W2) 1 / 2- L2, −ΔL3) (10) The above steps 710 to 717 are performed by, for example, FIG.
By performing 10 points of i = i0 to i9 shown in (A), (Xi1, Yi1) and (SXi, SYi, F
10 sets of i) are obtained, and the transformation matrices CS and CF of the above equations (6) and (7) are determined from these values. This completes the correction of the conversion matrices CS and CF.

In this embodiment, since the cross-sectional shape of the electron beam EB at the position of the aperture 34a is measured and the conversion matrices CS and CF are corrected based on this, the number of repetitive processings for correction is reduced. , This allows
The conversion matrices CS and CF can be corrected in a relatively short time as compared with a method in which trial and error is repeated many times based on only the current I without obtaining the cross-sectional shape of the electron beam EB.

Further, by using the relation of the above equation (9), the cross-sectional shape of the electron beam EB is set to the diameter d equal to the aperture 34a.
Since (SXi, SYi, Fi) for making a circle of 0 is obtained, it is not necessary to previously give a complicated empirical formula whose parameters are the ambient temperature and the drive signals to various deflectors and magnetic field lenses, and it is easy and Exactly (SXi, SYi, Fi)
The optimum value of can be obtained.

With this correction, a crossover image is generated by the influence of the field curvature corrector 38 and the astigmatism corrector 36.
Like CE2 shown in (B), it deviates from the position of the aperture 34a. (72) Then, in the same manner as in step 70, the conversion matrix C2 is corrected to obtain the crossover image as shown in FIG.
It is matched with the aperture 34a like CE3 shown in (C).

By the above steps 70, 71 and 72, the optimum values of C2, CS and CF which are related to each other can be efficiently, easily and promptly obtained. Conversely, if the optimum values of C2, CS and CF that are related to each other are to be obtained simultaneously, the processing becomes extremely complicated and the required time becomes long. (73) Next, the conversion matrix C4 is corrected in the same manner as in step 70 above.

That is, the conversion matrices CM, C3,
The conversion matrix C
2. The conversion matrices CS and CF are values corrected as described above, and the pattern code Ci0 corresponding to the point of i = i0 shown in FIG. 9A is supplied to the pattern coordinate memory 50. The constant term on the right side of the above equation (5) is changed little by little, for example, five times in the vicinity of the previous value, and for each Xi4 and Yi4,
The current I is detected by the current detector 40, Xi4 having the maximum current I is obtained from the graph of Xi4 and I, and Yi4 having the maximum current I is obtained from the graph of Yi4 and I.

Such processing is performed for 10 points of i = i0 to i9 shown in FIG.
i (1) is the maximum I (10) (Xi4, Yi4)
The set is obtained, and the 20 constants on the right side of the above equation (5) are determined from these values. Thereby, the correction of the conversion matrix C4 is completed. (74) Next, as shown in FIG. 4, the stencil mask 24 is attached to the XY stage 25, and the conversion matrix CM is corrected in the same manner as in step 70.

That is, the conversion matrices CM and C3 are left at the previous proper values, and the conversion matrices C2 and C3.
4. The conversion matrixes CS and CF are set to the values corrected as described above, the XY stage 25 is driven, and the center of the area E33 shown in FIG. 9B is moved to the XY coordinate origin of the exposure apparatus to form a rectangle. Pattern code C corresponding to passage hole p0
i is supplied to the pattern coordinate memory 50. The constant term on the right side of the above equation (8) is gradually changed, for example, five times in the vicinity of the previous value, the current I is detected by the current detector 40 for each Xi1 and Yi1, and the current I is calculated from the graph of Xi1 and I. The maximum Xi1 is calculated, and the maximum current Ii1 is calculated from the graph of Yi1 and I.

Such processing is performed by the stencil mask 24.
For example, 1 of the rectangular passage holes p0 to p9 shown in FIG.
This is performed for 0 points, 10 sets of (Xi1, Yi1) that maximize I with respect to (Xi, Yi) are obtained, and 20 constants on the right side of the above equation (8) are determined from these values. This completes the correction of the conversion matrix CM. (75) With the above correction, the electron beam EB passes through the aperture 34a as shown in FIG. 10, but the exposure position is generally displaced. Therefore, this shift is detected by a known method based on the detection value of the secondary electron detector 46, and the conversion matrix C3 is corrected so that the shift becomes zero.

That is, the conversion matrices CM, C2,
Assuming that C4, CS and CF are values corrected as described above, FIG.
The pattern code C corresponding to the point of i = i0 shown in (A)
i0 is supplied to the pattern coordinate memory 50. Equation (4)
The constant term on the right side of is gradually incremented three times near the previous appropriate value (j =
1 to 3), and each deviation (ΔXi3j, Δ) from the previous appropriate value (Xi30, Yi30) of (Xi3, Yi3) is changed.
The deviation (Δxij, Δyij) of the exposure position is detected for Yi3j) and j = 1 to 3.

Here, (Δxi, Δyi) is changed to (ΔXi
3, ΔYi3) is linearly approximated, and the relational expression between them is simply expressed as (Δxi, Δyi) = Cβ · (ΔXi3, ΔYi3) (11). This conversion matrix Cβ is given by (ΔXi3
j, ΔYi3j) and (Δxij, Δyij) j = 1,
Determine from three sets of values for a few. And (Δ
xi, Δyi) = (0,0) holds (ΔXi3, Δ
Yi3) is calculated.

Such processing is performed for 10 points of i = i0 to i9 shown in FIG.
i1) and (Xi30 + ΔXi3, Yi30 + ΔYi3)
10 sets are obtained, and from these values, 20 on the right side of the above equation (4)
Determine the number of constants. This gives the transformation matrix C
The correction of 3 is completed. For the correction of the conversion matrix C3, the relationship between the conversion matrix C3 and the conversion matrix C4 is maintained, that is, the same (Xi3, Y
In order not to change (Xi4, Yi4) for i3),
Change the transformation matrix C4. As a result, the midpoint of the crossover image of the electron beam EB coincides with the midpoint of the aperture 34a.

[0045]

As described above, in the electron beam cross-section shaping method at the position of the aperture for the aperture stop of the electron beam exposure apparatus according to the present invention, the cross-sectional shape of the charged particle beam at the position of the aperture is measured, Based on this, the first drive signal for making the cross-section circular and making the cross-sectional diameter substantially coincide with the diameter of the aperture is obtained, so that the current flowing through the diaphragm or sample without obtaining the cross-sectional shape of the charged particle beam. The number of iterations during correction is comparatively smaller than the method of repeating trial and error many times based on only the result, which makes the cross section of the charged particle beam at the position of the aperture round and the set diameter in a relatively short time. The excellent effect of being able to be achieved is largely contributed to the improvement of the throughput of the electron beam exposure apparatus.

In the first aspect of the present invention, the first drive signal is changed to obtain the relationship between the change amount of the first drive signal and the change amount of the feature amount of the cross-sectional shape. Also, it is possible to easily and accurately obtain the optimum value of the first drive signal without the need to previously give a complicated empirical formula using the drive signal to the magnetic field lens as a parameter.

According to the second aspect of the present invention, the optimum values of the first drive signal and the second drive signal which are related to each other can be efficiently and easily obtained quickly.

[Brief description of drawings]

FIG. 1 is a flowchart showing a principle configuration of the present invention.

FIG. 2 is a schematic flowchart showing a deflection conversion type correction procedure.

FIG. 3 is a flowchart showing details of a main part of step 71 of FIG.

FIG. 4 is a configuration diagram of a main part of an electron beam exposure apparatus.

FIG. 5 is a perspective view showing a principle configuration of an astigmatism corrector and a field curvature corrector.

FIG. 6 is a block diagram of a deflector drive signal generation circuit.

7 is a diagram showing a crossover image of an electron beam after the processes of steps 70, 71 and 72 of FIG. 2;

FIG. 8 is an explanatory diagram of a stencil mask passing position of an electron beam, which is selected at the time of deflection conversion type correction.

FIG. 9 is a diagram illustrating a feature amount of a crossover image of an electron beam.

FIG. 10 is an explanatory diagram of exposure position shift.

[Explanation of symbols]

 10 semiconductor wafer 12 XY stage 14 to 18, 20, 22 magnetic field lens 24 stencil mask 26, 28, 30, 32 deflector 34 angle diaphragm plate 34a aperture 36 astigmatism corrector 38 field curvature corrector 40 current detection 42 Differentiator 44 Signal processing circuit 46 Secondary electron detector C Optical axis EB Electron beam CE1 to CE3 Crossover image

Claims (3)

[Claims] 1. A charged particle beam cross-section shaping method at the position of an aperture (1a) for angular aperture of a charged particle beam exposure apparatus, the charged particle beam exposure apparatus comprising a selected passage on a stencil mask (2). A charged particle beam (EB) is passed through a hole pattern to form a cross section of the charged particle beam, and the aperture formed in the diaphragm plate (1) is focused and irradiated onto the sample (3) through the charged particle beam. , And a deflector (4) for deflecting the charged particle beam above the aperture, and the charged particle beam cross section at the position of the aperture is shaped by an electric field or a magnetic field in response to a first drive signal (S1). And a compensator (5) for the second driving signal (S2) to be supplied to the deflector.
By scanning the charged particle beam on the diaphragm plate while changing the value of, to detect the current (I) of the charged particle beam that is blocked by the diaphragm plate or passes through the aperture,
The cross-sectional shape of the charged particle beam is measured (A2), and based on the cross-sectional shape, the first drive signal for making the cross-section circular and making the cross-sectional diameter substantially coincide with the diameter of the aperture is obtained (A5). And supplying the obtained first drive signal to the corrector (A6). An electron beam cross-section shaping method at the position of an aperture for an angle diaphragm of an electron beam exposure apparatus, comprising: 2. The three feature quantities (L1, L2, L3) of the shape are extracted from the sectional shape (A3), and the first drive signal (S1) is changed to change the first drive signal. (ΔS1) and the amount of change in the feature amount (ΔL1, ΔL
2, ΔL3) is obtained (A4), the cross section is made circular, and the cross section diameter is set to the aperture (1a).
2. The method according to claim 1, wherein the first drive signal for making the diameter substantially match the diameter of the first drive signal is obtained from the relationship (A5). 3. The deflector (4) is arranged above the stencil mask (2), and the charged particle beam (E) is provided.
B) is to pass the stencil mask substantially perpendicular to the plane of the stencil mask and to pass through the selected through hole pattern, and the deflection is performed so that the amount of the charged particle beam passing through the aperture (1a) becomes a maximum value. Adjusting the second drive signal (S2) supplied to the chamber (A1), and then performing the process of claim 1 or 2, and the deflection so that the amount of the charged particle beam passing through the aperture becomes a maximum value. The second drive signal supplied to the apparatus is readjusted (A7), The electron beam cross-section shaping method at the position of the aperture for the angle diaphragm of the electron beam exposure apparatus.