JP2022142325A - Electric charge particle beam device - Google Patents

Abstract

To provide an electric charge particle beam device capable of measuring its own defocus in a short time with a sufficient measurable range and correcting the defocus.SOLUTION: An electric charge particle beam device includes a light source that generates a beam of electric charge particles, an optical system, blur measurement means, focus correction means, and two-fold astigmatism correction means. and while increasing or decreasing the intensity of a signal input to the two-fold astigmatism correction means, the blur measurement means measures the blur of a beam at the height position of a surface to be irradiated with the beam along two directions perpendicular to each other in the surface to be irradiated or one direction in the irradiated surface, and defocus appearing at the height position is determined from the intensity of the signal in which the measured blur is minimized.SELECTED DRAWING: Figure 1

Description

The present invention relates to a charged particle beam device with imaging capability. Such devices include electron beam writers, electron beam three-dimensional printers, scanning electron microscopes (SEM), and transmission electron microscopes (TEM).

A charged particle beam device is a device that controls and utilizes a charged particle beam for some purpose. Charged particle beams are broadly classified into electron beams and ion beams.
Among charged particle beam devices, devices that have an imaging function and are intended for fine drawing, modeling, observation, etc. are required to use beams with small aberration and blur. Electron beam writers, electron beam 3D printers, scanning electron microscopes, and transmission electron microscopes are representative of such devices.

As one of the above apparatuses, a variable shaping electron beam writing apparatus (see Patent Document 1 and Non-Patent Document 1) will be described below.

A variable shaping electron beam lithography system is an electron beam lithography system developed for the purpose of achieving high lithography throughput while enabling fine lithography. A variable shaping electron beam writing apparatus is mainly used for writing photomasks for manufacturing semiconductor devices.
For the above purpose, the variable shaped electron beam writing apparatus makes the shape and dimensions of the cross section of the electron beam on the material to be drawn (material to be exposed) variable, and controls the shape and dimensions according to the pattern to be drawn. . Here, the cross-section of the electron beam on the material to be written determines the area exposed by the beam at one time.

FIG. 24 shows an example of an optical system used in a variable shaped electron beam writing apparatus. In FIG. 24, for convenience of explanation, it is assumed that the optical system components (lens, deflector, and apertures) are arranged along the Z-axis and the electron beam 1 flows in the positive direction of the Z-axis. The lenses and deflectors in this optical system may be of either the magnetic type or the electrostatic type, but hereinafter, unless otherwise specified, they are assumed to be the magnetic type and the electrostatic type, respectively.

In the optical system, first, as shown in FIG. 24, the electron beam 1 is converged by the irradiation lens 2 and the first shaping aperture plate 3 is irradiated with the electron beam 1 . Here, as the supply source of the electron beam 1, that is, the light source 4, generally a crossover formed in an electron gun (not shown) may be assumed.
As a result of the irradiation, an image 5 of the light source is formed under the first shaping aperture plate 3 . The position of the image 5 of the light source coincides with the position where the rays (dashed lines) having the starting point at the height position of the light source 4 intersect in FIG. 24 . The light source images (light source images 5, 16, and 19) generated by the optical system are all formed at the positions where these light rays intersect.

The optical system then projects the image of the aperture of the first shaping aperture plate 3 onto the second shaping aperture plate 7 through the shaping lens 6 . The position of the image coincides with the first intersection of the light beam (solid line) having a starting point at the height position of the first shaping aperture plate 3 in FIG. 24 . All the images (including the projection figure 11) of the aperture of the shaping aperture plate produced by the optical system are formed at the positions where these rays intersect.
The optical system uses a reducing lens 8 and an objective lens 9 to convert the image of the figure (logical product) generated by overlapping the image of the aperture of the first shaping aperture plate 3 and the aperture of the second shaping aperture plate 7. , onto the material 10 . That is, material 10 is exposed. A resist (photosensitive material) is applied to the material 10 .
As a result of the exposure, the resist on material 10 is exposed. That is, the projected figure 11 is transferred to the resist on the material 10 . Therefore, a desired pattern can be drawn on the material 10 by controlling the shape, size, position, and exposure time of the projection figure 11 .

The shaping deflector 12 controls the shape and size (two-dimensional) of the projected figure 11 . If the electron beam 1 is deflected by the shaping deflector 12, the shape and size of the figure produced by the above-described overlap are changed, and accordingly the shape and size of the projected figure 11 are also changed.

The position (two-dimensional) of the projection figure 11 is controlled by the combined use of the objective deflector 13 and the material stage 14 . Due to the limited deflectable area of the objective deflector 13, i.e. the deflection field (square or rectangular), the material 10 is first positioned in large steps by moving the material 10 by means of the material stage 14, and then the objective Deflection of the electron beam 1 by the deflector 13 provides positioning in small steps.
The deflection field limitation is due to the rated output of a power supply (not shown) driving the objective deflector 13 and the deflection sensitivity of the objective deflector 13 . Here, the deflection sensitivity refers to the amount of deflection per unit applied voltage at the height position of the surface of the material 10 .

The larger the deflection field of the objective deflector 13 is, the better from the viewpoint of the drawing throughput (drawing speed). This is because increasing the deflection field of the objective deflector 13 reduces the number of times the material stage 14 moves and stops.
Here, moving and stopping the material stage 14 produces dead time (time that does not directly contribute to writing on the material 10) during writing on the material 10. FIG. The number of movements and stops is inversely proportional to the ratio of the area of the deflection field of the objective deflector 13 to the area of the material 10 .
In other words, by increasing the deflection field of the objective deflector 13 and reducing the number of times the material stage 14 is moved and stopped, the waste time can be reduced, and as a result, the drawing throughput can be improved.

The objective deflector 13 is an 8-pole electrostatic deflector (multipole element). Since the purpose of the objective deflector 13 is two-dimensional deflection of the electron beam 1, the number of poles of the objective deflector 13 may be four (two poles each for X deflection and Y deflection). However, if the number of poles is four, a nonlinear potential component is newly generated with respect to the radius from the center of the objective deflector 13, and therefore the deflection aberration resulting from the potential component cannot be ignored. By setting the number of poles of the objective deflector 13 to 8, the potential component is reduced, and therefore the deflection aberration resulting from the potential component is reduced.

A blanker 15 controls the exposure time of the projection figure 11 . While the blanker 15 is not activated, the electron beam 1 is incident on the material 10 and the material 10 is exposed. Therefore, the exposure time is the time from when the blanker 15 is deactivated to when it is reactivated. Here, the electron beam 1 is blocked by the actuation of the blanker 15 because the electron beam 1 is deflected by the actuation and enters the non-aperture portion of the blanking aperture plate 17 . The blanker 15 interrupts the electron beam 1 in this way also during dead time as described above.

The important performance indicators of the variable shaped electron beam lithography system are writing throughput (writing speed) and writing accuracy. Demands for these performance indicators are increasing year by year.

Although it is possible to improve these performance indexes at the same time, if the drawing throughput is improved among these performance indexes, the drawing accuracy tends to decrease. That is, when trying to improve the drawing throughput, the need for improving the drawing precision increases. This can be explained as follows using FIG.
Among the above performance indicators, the current density of the electron beam 1 on the material 10 must be increased in order to improve the writing throughput. (The higher the current density of the electron beam 1, the shorter the exposure time of the projection figure 11 in inverse proportion.) On the other hand, in order to improve the drawing accuracy, the blurring of the electron beam 1 on the material 10 must be reduced. There is
Here, drawing accuracy is, more specifically, dimensional accuracy. The writing accuracy of the variable shaped electron beam writing apparatus is roughly divided into dimensional accuracy and positional accuracy. Of these drawing accuracies, the former is mainly concerned with the blurring of the electron beam 1, and the latter is mainly concerned with the displacement of the electron beam 1. FIG.
However, if the current density of the electron beam 1 on the material 10 is increased in order to increase the drawing throughput, the blurring of the electron beam 1 on the material 10 tends to increase. This involves aberrations produced by the objective lens 9 and the objective deflector 13 .

For this reason, in the above optical system (see FIG. 24), defocus and astigmatism that are not caused by the deflection of the electron beam 1 by the objective deflector 13 occur regularly as part of improving drawing accuracy. Measured and corrected. Both the defocus and astigmatism are measured and corrected by the knife-edge method (see, for example, US Pat. No. 6,300,003).
Here, the defocus and astigmatism not caused by the deflection of the electron beam 1 by the objective deflector 13 respectively refer to the defocus that appears on the material 10 under the condition that the electron beam 1 is not deflected by the objective deflector 13. and 2-fold astigmatism, ie defocus and 2-fold astigmatism at the center of the deflection field of the objective deflector 13, respectively.
Hereinafter, such defocus and astigmatism are referred to as axial defocus and axial 2-fold astigmatism, respectively. In this specification, axial defocus is treated as an aberration, like axial 2-fold astigmatism and the like.

More specifically, the above-mentioned axial defocus and axial two-fold astigmatism are caused by the light rays around each principal ray included in the electron beam 1 becoming each principal ray on the surface (paraxial image plane) of the material 10. Defocus and 2-fold astigmatism shown against. These aberrations do not depend on the coordinates within the projected figure 11 and appear throughout the projected figure 11 . (The same applies to deflection curvature of field aberration and deflection 2-fold astigmatism, which will be described later.)
Here, the number of principal rays contained in the electron beam 1 is as many as the number of image points (innumerable) in the projected figure 11, and all of these principal rays are centered on the respective images 5, 16, and 19 of the light source. pass. That is, the principal ray included in the electron beam 1 intersects with each of the light source images 5, 16, and 19 at one height position.
Supplementally, defocus and two-fold astigmatism that depend on the coordinates in the projected figure 11 also appear in the projected figure 11 . However, as used herein, such defocus and 2-fold astigmatism are referred to above as axial defocus and 2-fold astigmatism (as well as deflection field curvature aberration and deflection 2-fold astigmatism described below). aberration), and therefore can be ignored.

In the optical system (see FIG. 24), deflection aberration caused by the deflection of the electron beam 1 by the objective deflector 13 is periodically measured and corrected as part of improving the drawing accuracy. These aberrations are measured and corrected by the knife-edge method with the electron beam 1 deflected by the objective deflector 13 . These aberrations are measured and corrected at points within the deflection field of the objective deflector 13 . More specifically, these points are lattice points obtained by equally or unequally dividing the deflection field of the objective deflector 13 in the X and Y directions.
Here, the deflection aberration caused by the deflection of the electron beam 1 by the objective deflector 13 is specifically the deflection distortion aberration appearing at the height position of the material 10 as the electron beam 1 is deflected by the objective deflector 13. , deflection field curvature aberration (defocus caused by deflection of electron beam 1 by objective deflector 13), deflection astigmatism (astigmatism caused by deflection of electron beam 1 by objective deflector 13), deflection coma , and polarization chromatic aberration. Among these aberrations, the deflection distortion contributes to the displacement of the electron beam 1, and the others contribute to blurring of the electron beam 1. FIG.
However, among these deflection aberrations, deflection distortion aberration, deflection field curvature aberration, and deflection astigmatism are usually measured and corrected. Here, the deflection astigmatism is, more specifically, two-fold deflection astigmatism. Henceforth, description about deflection distortion aberration is omitted.

The axial defocus, the axial 2-fold astigmatism, the deflection field curvature aberration, and the deflection 2-fold astigmatism are the focus correction current, the 2-fold astigmatism correction voltage, the focus correction voltage, and the 2-fold astigmatism, respectively. It is measured and corrected based on the blur curve of the electron beam 1 against the point correction voltage.
That is, in order to measure and correct each aberration (axial defocus, axial two-fold astigmatism, deflection field curvature aberration, or deflection two-fold astigmatism), each correction signal (focus correction current, A curve of the blurring of the electron beam 1 versus the double stigmation voltage, focus correction voltage, or double stigmation voltage is obtained. Then, from each curve obtained for that purpose, the value (intensity) of each correction signal that minimizes the blurring of the electron beam 1 is determined, and each correction signal with the value thus determined is applied to each lens or corrector. (the objective lens 9, the following electrostatic two-fold astigmatism corrector, the following electrostatic focus corrector, or the following electrostatic two-fold astigmatism corrector).
Here, the focus correction current refers to a current added to the excitation current of the objective lens 9 (or the reduction lens 8) for correcting the axial defocus. The two-fold astigmatism correction voltage refers to a voltage input to an electrostatic two-fold astigmatism corrector (not shown) for correcting the axial two-fold astigmatism and deflection two-fold astigmatism. . The focus correction voltage refers to a voltage input to an electrostatic focus corrector (not shown) for correcting the deflection curvature of field. The electrostatic two-fold astigmatism corrector and the electrostatic focus corrector are provided, for example, at a height position approximately equal to the height position of the objective deflector 13 . Each of the focus correction current and the focus correction voltage consists of one component, but the two-fold astigmatism correction voltage consists of two mutually independent components.

Curves of the blurring of the electron beam 1 with respect to the focus correction current, the two-fold astigmatism correction voltage, and the focus correction voltage are obtained by the knife edge method while increasing and decreasing the focus correction current, the two-fold astigmatism correction voltage, and the focus correction voltage, respectively. It is obtained by measuring the blur of the electron beam 1 . These curves are, more particularly, fitted curves to the multiple measurements of blur thus obtained. These curves are obtained for each correction signal and for each component thereof.

Measuring and correcting each of the aberrations (the axial defocus, the axial two-fold astigmatism, the deflection curvature of field, or the deflection two-fold astigmatism) according to the above-described procedures is the It corresponds to measuring by the null method. That is, the measured value and correction value of each aberration are balanced with each other when the blurring of the electron beam 1 is minimized. Here, the measured value of each aberration is the aberration converted into each correction signal (focus correction current, two-fold astigmatism correction voltage, focus correction voltage, or two-fold astigmatism correction voltage). The correction value of each aberration is the value of each correction signal that minimizes the blurring of the electron beam 1 .
The measured and corrected values of the above aberrations have the same magnitude and opposite signs. This relationship holds for each correction signal and each component thereof.

Supplementally, in principle, the axial defocus can be measured and corrected together with the deflection curvature of field when measuring and correcting the deflection curvature of field. That is, the axial defocus can be regarded as part of the deflection field curvature aberration.
However, if this is actually done, the amount to be measured and corrected at one time (the sum of the axial defocus and the deflection curvature of field aberration) becomes large when measuring and correcting the deflection curvature of field aberration. Become. As a result, an unfavorable situation is likely to occur in which a blur curve of the electron beam 1 with respect to the focus correction voltage cannot be acquired.
Here, the situation in which the blur curve of the electron beam 1 against the focus correction voltage cannot be obtained is that if the focus correction voltage is increased or decreased by a width sufficient to confirm the position of the trough of the curve, the electrostatic focus corrector deviates from the range of focus correction voltages that can be input to the electrostatic focus corrector.
The same applies to the axial 2-fold astigmatism mentioned above. That is, the on-axis 2-fold astigmatism can be measured and corrected together with the 2-fold deflection astigmatism when measuring and correcting the 2-fold deflection astigmatism. If the axial 2-fold astigmatism is measured and corrected together with the deflection 2-fold astigmatism, a situation may occur in which the blur curve of the electron beam 1 versus the 2-fold astigmatism correction voltage cannot be obtained.
However, such undesired situations are more likely to occur when measuring and correcting the axial defocus with the deflection field curvature than when measuring and correcting the axial 2-fold astigmatism with the deflection 2-fold astigmatism. It is more likely to occur when correcting. This is related to the possible magnitudes of the axial defocus and the axial two-fold astigmatism, as well as the focus correction sensitivity of the electrostatic focus corrector and the electrostatic two-fold astigmatism corrector. 2-fold astigmatism correction sensitivity. Here, the focus correction sensitivity refers to the defocus correction amount per unit focus correction voltage. The two-fold astigmatism correction sensitivity refers to the correction amount of two-fold astigmatism per unit two-fold astigmatism correction voltage.

In the above description, the axial defocus is corrected by the objective lens 9 (magnetic field type lens), and the axial two-fold astigmatism, the deflection curvature of field aberration, and the two-fold deflection astigmatism are corrected respectively. Correction is performed by the electrostatic 2-fold astigmatism corrector, the electrostatic focus corrector, and the electrostatic 2-fold astigmatism corrector (all are electrostatic correctors). Whether these aberrations should be corrected by a magnetic lens (or a magnetic corrector) or by an electrostatic corrector depends on whether the correction of the aberrations to be corrected is static correction or dynamic correction. do. Importantly, electrostatic correctors generally operate faster than magnetic lenses and magnetic correctors. (By acting fast, we mean exhibiting a faster response to correction signals.)
The reason why the optical system uses the objective lens 9 to correct the axial defocus is that in the optical system, the axial defocus may be corrected by static correction. The same applies to the axial 2-fold astigmatism mentioned above. That is, the axial two-fold astigmatism may be corrected by a magnetic two-fold astigmatism corrector (not shown) instead of the electrostatic two-fold astigmatism corrector.
The reason why the optical system (see FIG. 24) corrects the deflection field curvature aberration and the deflection 2-fold astigmatism by the electrostatic focus corrector and the electrostatic 2-fold astigmatism corrector, respectively, is that the optical system This is because, in the system, the correction of these aberrations must be dynamic correction. In other words, the variable shaping electron beam writing apparatus needs to perform such deflection aberration correction at high speed for each deflection coordinate in order to increase its writing throughput. (The deflection coordinates refer to coordinates within the deflection field of the objective deflector 13.)
Hereinafter, for convenience of explanation, unless otherwise specified, the correction of the axial defocus is assumed to be made by the objective lens 9 as described above, while the axial two-fold astigmatism and deflection field curvature Aberrations and deflection two-fold astigmatism are all corrected by the objective deflector 13 .
Here, the correction of the axial two-fold astigmatism, the deflection curvature of field, and the two-fold deflection astigmatism by the objective deflector 13 means that the objective deflector 13 has the electrostatic two-fold astigmatism. An astigmatism correction voltage is superimposed twice on the deflection voltage so as to serve also as a corrector (see, for example, Patent Document 2). It means to superimpose a focus correction voltage (see, for example, Patent Document 2). (The deflection voltage is the voltage input to the objective deflector 13 for deflection of the electron beam 1 by the objective deflector 13.) If the focus corrector is also used as a stigmator, the special stigmator and focus corrector (that is, the electrostatic double stigmator and the electrostatic focus corrector) become unnecessary. , the power supply (not shown) to drive the correctors is also eliminated.

Supplementally, the objective deflector 13 alone can have only low focus correction sensitivity. In order to provide the objective deflector 13 with a focus correction sensitivity high enough to correct the deflection field curvature aberration, it is necessary to place the objective deflector 13 in the magnetic field of the magnetic lens. In the optical system (see FIG. 24), the magnetic field of the objective lens 9 corresponds to the magnetic field.
If the objective deflector 13 is placed in the magnetic field of the objective lens 9, the input of the focus correction voltage to the objective deflector 13 makes the potential in the objective deflector 13 higher (or lower) than the surrounding potential. As a result, the strength of the focusing action of the objective lens 9 changes. This is because the input of the focus correction voltage to the objective deflector 13 changes the kinetic energy of the electron beam 1 in the magnetic field of the objective lens 9 .
These statements regarding the objective deflector 13 also apply to the electrostatic focus corrector.

Japanese Patent Application Laid-Open No. 2007-67192 JP 2014-194982 A

K. Komagata, Y. Nakagawa, H. Takemura and N. Gotoh, Proc. SPIE, Vol.3096, pp.125-136, (1997)

There are two problems with measuring and correcting the axial defocus and deflection field curvature aberrations in the manner described above.

First, the measurement and correction of the on-axis defocus requires a long time to acquire the curve of the blurring of the electron beam 1 against the focus correction current. This is due to the slow response of the objective lens 9 to the focus correction current. The response involves the self-resonant frequency of the objective lens 9 .

Second, in the measurement and correction of the deflection curvature of field aberration, the range in which the objective deflector 13 can operate as an electrostatic focus corrector is narrow. narrow. If the measurable range is not wide enough, the above-mentioned undesirable situations may occur. That is, when measuring the deflection curvature of field, a situation may occur in which the blur curve of the electron beam 1 with respect to the focus correction voltage cannot be obtained.
Here, the range in which the objective deflector 13 can operate as an electrostatic focus corrector is the range in which the objective deflector 13 can change the defocus on the knife edge (not shown) in accordance with the focus correction voltage. be. (The defocus on the knife edge is an aberration resulting from a combination of the deflection field curvature aberration and the defocus due to the input of the focus correction voltage to the objective deflector 13.) In this range, electrostatic The focus correction sensitivity of the objective deflector 13 as a type focus corrector is involved. On the other hand, the measurable range of the deflection curvature of field aberration is, more specifically, the range in which the deflection curvature of field aberration can be measured by the null method as described above.

The first problem becomes a serious problem depending on the number (frequency) of measurement and correction of the axial defocus. As the number of measurements and corrections increases, the number of curves to be acquired increases proportionally.
The number of times can be increased, for example, when there is a large height difference within the material 10 and therefore the axial defocus varies greatly from region to region within the material 10 (see Example 9 of the present invention). In. In such a case, it becomes necessary to increase the number of measurement points of the axial defocus in the material 10 in order to reduce the correction residual of the axial defocus for each region in the material 10 . That is, the number of times is increased in proportion to the number of measurement points.

In the first problem, it should be noted that in acquiring the curve, the focus correction current is given a plurality of discrete values per curve, and for each of these values, the blurring of the electron beam 1 is to be measured. That is, the blurring of the electron beam 1 is measured a number of times equal to the number of these values, and each time it is necessary to wait for the defocus on the knife edge to settle with respect to changes in the focus correction current. In other words, in addition to the net time required for measuring the blur of the electron beam 1 for that number of times, a waiting time (waste time) for defocus stabilization on the knife edge is required for that number of times. Furthermore, these times are required for the number of curves.

The above first problem can be solved by measuring and correcting the axial defocus by using the objective deflector 13 as an electrostatic focus corrector.
However, at the cost of this, the measurable range of the axial defocus is narrowed, increasing the possibility that the desired axial defocus cannot be measured. This is because the range in which the objective deflector 13 can operate as an electrostatic focus corrector is narrow.
In other words, the first problem relates to the trade-off between the operable range and the response speed of the means for correcting defocus on the knife edge.

The second problem above becomes a serious problem depending on the size of the deflection field of the objective deflector 13 . If the deflection field is increased, the drawing throughput is improved as described above, but the width of the range in which the deflection field curvature aberration can be changed is increased. The narrowness of the range in which it can operate as a compensator tends to become a major problem. Here, the width of the range in which the magnitude of the deflection field curvature aberration can be changed refers to the difference between the maximum value and the minimum value of the deflection field curvature aberration within the deflection field of the objective deflector 13 . (The maximum and minimum values of the deflection curvature of field aberration correspond to the maximum and minimum values of the height position of the projected figure 11, respectively.)

Regarding the second problem, it should be noted that the width of the variable range of the deflection field curvature aberration is affected by machining and assembly errors of the optical system components in FIG. machining and assembly errors of the deflector 13 ) and the alignment of the electron beam 1 with respect to the objective lens 9 and the objective deflector 13 .
Here, if the deflection field of the objective deflector 13 is circular, the magnitude of the deflection field curvature aberration is distributed axially symmetrically. The center of the distribution deviates from the center of the deflection field of the objective deflector 13 depending on the above errors and alignment. As the deviation increases, the width of the range in which the deflection curvature of field can vary also increases.

Further attention should be paid to the above second problem, even if the width of the range in which the deflection field curvature aberration can be changed is the width of the range in which the objective deflector 13 can operate as an electrostatic focus corrector. Even if the following condition is satisfied, the situation described above (the situation in which the blur curve of the electron beam 1 against the focus correction voltage cannot be obtained) can occur. This is related to the size of these widths as well as the initial value of the focus correction voltage.

The second problem above is to first replace the acquisition of the blur curve of the electron beam 1 against the focus correction voltage with the acquisition of the blur curve of the electron beam 1 against the focus correction current, and then from those curves, the electron beam 1 This problem can be solved by determining the value of the focus correction current that minimizes blurring and converting the focus correction current of that value into a focus correction voltage.
However, at the cost of this, the time required for measuring and correcting the deflection curvature of field aberration becomes longer. It should be noted here that in the measurement and correction of the deflection curvature of field, the curve of blurring of the electron beam 1 with respect to the focus correction signal (focus correction voltage or focus correction current) is obtained for each of the aforementioned lattice points. Is Rukoto. That is, the number of curves to be acquired is proportional to the number of grid points described above.
That is, the second problem is also related to the trade-off between the operable range and the response speed of the means for correcting defocus on the knife edge, like the first problem.

The present invention has been made in view of the above first and second problems. More specifically, the present invention measures self-generated defocus (either or both of axial defocus and deflection field curvature aberration) in a short time with a sufficiently large measurable range. Another object of the present invention is to provide a charged particle beam apparatus capable of correcting the defocus.

A charged particle beam apparatus of the present invention comprises a light source for generating a charged particle beam, an optical system for the charged particle beam, and blur measurement means for the charged particle beam.

The optical system includes at least one stage of lens for transmitting the charged particle beam to the surface to be illuminated illuminated by the charged particle beam.
The optical system is an image of the light source, a transmission image of an aperture in an aperture plate arranged in the optical system and illuminated by the charged particle beam, or an image arranged in the optical system and the charged particle beam. A transmission image of the thin film irradiated by is formed on the irradiated surface by the at least one stage of the lens.
The optical system further includes focus correction means for correcting defocus appearing at a height position of the surface to be illuminated, and two-fold astigmatism correction for correcting two-fold astigmatism appearing at a height position on the surface to be illuminated. means.

The focus correction means receives a signal for correcting the defocus as a focus correction signal, and changes the defocus according to the signal.
The two-fold astigmatism correction means receives a signal for correcting the two-fold astigmatism as a two-fold astigmatism correction signal, and changes the two-fold astigmatism according to the signal.
The blur measuring means measures blur of the charged particle beam appearing at a height position of the surface to be irradiated. The blur changes its size depending on the defocus and 2-fold astigmatism.

The charged particle beam device of the present invention is further configured to acquire the defocus by any of the following three methods (first, second and third methods).
A first method is to measure the magnitude of the blur along each of two mutually orthogonal directions in the irradiated surface by the blur measuring means while increasing or decreasing the intensity of the double astigmatism correction signal. and the intensity of the double astigmatic correction signal that minimizes the magnitude of the blur in one of the two directions and the magnitude of the blur in the other of the two directions is minimized. The difference from the intensity of the double astigmatic signal is determined as the first difference, which difference represents the defocus.
A second method is to measure the magnitude of the blur along each of two mutually orthogonal directions in the irradiated surface by the blur measuring means while increasing or decreasing the intensity of the twice astigmatic correction signal. and the intensity of the double astigmatic correction signal that minimizes the magnitude of the blur in one of the two directions and the magnitude of the blur in the other of the two directions is minimized. determining the average of the intensity of the twice astigmatism-corrected signal as a first average, and determining the magnitude of the twice astigmatism-corrected signal that minimizes the magnitude of one of the two directions of the blur; The difference between the intensity and the first average is determined as the second difference, which difference represents the defocus.
A third method is to measure the magnitude of the blur along a single direction in the irradiated surface by the blur measuring means while increasing or decreasing the intensity of the twice astigmatic correction signal, and intensity of the twice astigmatically corrected signal that minimizes the magnitude of the blur in that single direction; and increasing or decreasing the magnitude of the twice astigmatically corrected signal in that single direction The difference from the intensity of the twice astigmatized signal before measuring along is determined as the third difference, which difference represents the defocus.

The charged particle beam apparatus of the present invention, as described above, is configured to acquire defocus appearing at the height position of the surface to be irradiated by the first, second, or third method. .
As a result, the defocus appearing at the height position of the surface to be illuminated can be measured in a short time with a sufficient measurable range.

1 is a diagram showing configurations of an optical system, a measurement system, and a control system of an apparatus (variably shaped electron beam writing apparatus) according to Example 1 of the present invention; FIG. 4 is a diagram showing electrode voltage distribution followed by deflection voltages, distortion correction voltages, focus correction voltages, and two-fold astigmatism correction voltages in Example 1. FIG. FIG. 2 is a diagram showing an example of current density distribution of the projected figure in FIG. 1; FIG. 10 is a diagram showing an example of X-direction and Y-direction blur curves of the electron beam 1 with respect to the A two-fold astigmatism correction voltage V _2A ; FIG. 10 is a diagram showing an example of X-direction and Y-direction blur curves of the electron beam 1 with respect to the B two-fold astigmatism correction voltage V _2B ; The X direction of the electron beam _{1 with respect to V 2A (or V 2B ) and} FIG. 10 is a diagram showing an example of a blur curve in the Y direction; FIG. 5 is a diagram showing an example of the relationship between the XY difference _DA and the focus correction current _I0A , and the relationship between the XY difference _DB and the focus correction current _I0A ; 4 is a diagram showing a routine for measuring and correcting axial defocus produced by the apparatus of Example 1. FIG. 7 is a diagram showing the contents of a subroutine (determination of XY difference D _An ) in FIG. 6; FIG. FIG. 7 is a diagram showing the contents of another subroutine (determination of I _0A0 ^(m)) in FIG. 6; FIG. 4 shows another measurement and correction routine for the axial defocus produced by the apparatus of Example 1; 4 is a diagram showing routines for measuring and correcting deflection curvature of field produced by the apparatus of Example 1. FIG. FIG. 11 is a diagram showing the contents of a subroutine (determination of V _0A0 ^(m) (x, y)) in FIG. 10; FIG. 12 is a diagram showing the contents of a subroutine (determination of V _0A0 ^(m)) in FIG. 11; It is a figure which shows the example of _defocus figure SF. It is a figure which shows the example of ₂ times astigmatism figure S2. It is a figure which shows the example of the aberration figure _{SF +} S2 which _{V2SA produces} when ASF<>0 and _V2SB =0 are materialized. It is a figure which shows the example of the aberration figure _{SF +} S2 which _{V2SB produces when BSF<>0 and V2SA = 0} are materialized. Fig. ₂ shows how an ellipse formed by an aberration diagram SF + S2 _{created by V2SA when ASF} ≠ 0 and _V2SB = 0 has a major axis (or minor axis) parallel to the X-axis or the Y-axis. It is a diagram. The ellipse formed by the aberration diagram S _F +S ₂ created by V _2SB when _B SF ≠0 and V _2SA =0 holds, is rotated by π/4 (45°) with respect to the X and Y axes. FIG. 4 is a diagram showing how it has a major axis (or a minor axis); FIG. 4 shows that x _d and y _d are the widths of the aberration figure _SF +S ₂ in the X and Y directions, respectively. _{Subroutines (I 0A0 (m), V 2A0 (} ^{m )} _, ^and V _2B0 ^(m) decision). FIG. 19 is a diagram showing the contents of a subroutine (determination of XY difference D _A0 ^(m) , XY average M _A0 ^(m) , and XY average M _B0 ^(m)) in FIG. 18; Subroutine (Determination of V _0A0 (m ⁾ , V _2A0 ^(m) , and V _2B0 ^(m) ) executed by the apparatus of Example 2 instead of the subroutine (Determination of V _0A0 ( ^m )) shown in FIG. It is a figure which shows the content. FIG. 12 is a diagram showing the configuration of the measurement system of the device (variably shaped electron beam writing device) of Example 9; FIG. 12 is a diagram showing the configuration of the optical system of the device (spot electron beam drawing device) of Example 10; FIG. 12 is a diagram showing the configuration of the optical system of the apparatus (transmission electron microscope) of Example 12; FIG. 2 is a diagram showing an example of an optical system used in a variable shaping electron beam writing apparatus;

Outline of the present invention>
SUMMARY OF THE INVENTION It is an object of the present invention to provide a charged particle beam apparatus capable of measuring its own defocus in a short time with a sufficient measurable range, and furthermore capable of correcting the defocus. and
A charged particle beam apparatus according to the present invention is configured as follows for the above purpose.

A charged particle beam apparatus of the present invention comprises a light source for generating a charged particle beam, an optical system for the charged particle beam, and blur measurement means.

The optical system includes at least one stage of lens for transmitting the charged particle beam to a surface to be irradiated with the charged particle beam.
The optical system may be an image of the light source, a transmission image of an aperture in an aperture plate arranged in the optical system and illuminated by the charged particle beam, or an image of the aperture plate arranged in the optical system and illuminated by the charged particle beam. A transmission image of the thin film irradiated by is formed on the irradiated surface by the at least one stage of the lens.
The optical system further includes focus correction means for correcting defocus appearing at a height position of the surface to be illuminated, and two-fold astigmatism correction for correcting two-fold astigmatism appearing at a height position on the surface to be illuminated. means.

The focus correction means receives a signal for correcting the defocus as a focus correction signal, and changes the defocus according to the signal.
The two-fold astigmatism correction means receives a signal for correcting the two-fold astigmatism as a two-fold astigmatism correction signal, and changes the two-fold astigmatism according to the signal.
The blur measuring means measures blur of the charged particle beam appearing at a height position of the surface to be irradiated. The blur changes its size depending on the defocus and 2-fold astigmatism.

The charged particle beam device of the present invention is further configured to acquire the defocus by any of the following three methods (first, second and third methods).
A first method is to measure the magnitude of the blur along each of two mutually orthogonal directions in the irradiated surface by the blur measuring means while increasing or decreasing the intensity of the double astigmatism correction signal. and the intensity of the double astigmatic correction signal that minimizes the magnitude of the blur in one of the two directions and the magnitude of the blur in the other of the two directions is minimized. The difference from the intensity of the double astigmatism signal is determined as the first difference, which difference represents the defocus.
A second method is to measure the magnitude of the blur along each of two mutually orthogonal directions in the irradiated surface by the blur measuring means while increasing or decreasing the intensity of the twice astigmatic correction signal. and the intensity of the double astigmatic correction signal that minimizes the magnitude of the blur in one of the two directions and the magnitude of the blur in the other of the two directions is minimized. determining the average of the intensity of the twice astigmatism-corrected signal as a first average, and determining the magnitude of the twice astigmatism-corrected signal that minimizes the magnitude of one of the two directions of the blur; The difference between the intensity and the first average is determined as the second difference, which difference represents the defocus.
A third method is to measure the magnitude of the blur along a single direction in the irradiated surface by the blur measuring means while increasing or decreasing the intensity of the twice astigmatic correction signal, and intensity of the twice astigmatically corrected signal that minimizes the magnitude of the blur in that single direction; and increasing or decreasing the magnitude of the twice astigmatically corrected signal in that single direction The difference between the intensity of the twice astigmatized signal before being measured along is determined as the third difference, which difference represents the defocus.

In the above configuration, the defocus is acquired by the first, second, or third method, as described above. As a result, the defocus can be measured in a short time with a sufficient measurable range.
This is because in the above configuration, the defocus is measured based on the blur measurement result for the two-fold astigmatism correction signal, not the blur measurement result for the focus correction signal. For the measurement, the speed of response of the two-time astigmatism correction means to the two-time astigmatism correction signal, The size of the range in which the aberration can be changed is utilized.
Here, the measurement result of the blur for the two-fold astigmatism-corrected signal is the measured value of the magnitude of the blur in each of the directions for a plurality of intensity values of the two-fold astigmatism-corrected signal. These measurements can be represented by a fitted curve for these measurements. In this way, the intensity of the double astigmatic correction signal that minimizes the magnitude of the blur in each direction can be determined from the position of the trough of each approximation curve.

The charged particle beam apparatus of the present invention further comprises a knife-edge blur measurement medium for measuring the blur at a height position of the surface to be irradiated, and a blur measurement medium between the light source and the surface to be irradiated. and scanning means for causing the charged particle beam to scan the blur measuring medium by deflecting the charged particle beam, wherein the blur measuring medium and the scanning means are used as the blur measuring means. , configurable.
More specifically, the blur measurement by the blur measuring means is performed by scanning the blur measuring medium with the scanning means while the charged particle beam is not interrupted by the blur measuring medium. , detecting the current in the portion blocked by the blur measuring medium, and evaluating the blur based on the blunting of the waveform of the detected current.

In the charged particle beam apparatus of the present invention, the two-time astigmatism correction signal is composed of two components independent of each other, and the two components constituting the two-time astigmatism correction signal are independent of each other. By being applied to the two-fold astigmatism correction means, the two-fold astigmatism can be provided with a change consisting of two linearly independent components.
In addition, the charged particle beam apparatus of the present invention further provides two types of the first, second, or third differences obtained by increasing or decreasing each of the two mutually independent components constituting the two-time astigmatism correction signal. The defocus can be represented by any one of the first, second, or third differences having a non-zero magnitude.
With the above configuration, it is possible to reliably prevent an undesirable situation in which the defocus is erroneously determined to be zero even though the defocus has a non-zero magnitude. Without the above configuration, such a situation would occur depending on the rotation angle of the two-fold astigmatism generated on the surface to be illuminated due to the two-fold astigmatism correction signal, and the two-way or single direction. Depending on one direction, it can occur. (Here, the 2-fold astigmatism generated on the surface to be illuminated due to the 2-fold astigmatism correction signal is a change consisting of two linearly independent components given to the 2-fold astigmatism. ,Equivalent to.)

In the charged particle beam apparatus of the present invention, any one of the first, second, or third differences having the non-zero magnitude may be any one of the two of the first, second, or Among the third differences, one of the first, second, or third differences having a non-smaller absolute value, and one of the first having a non-smaller absolute value , the second, or the third difference to represent the defocus.
This configuration prevents the defocus measurement accuracy from deteriorating.

In the charged particle beam apparatus of the present invention, a component exhibiting one of the first, second, or third differences for expressing the defocus is called a first component, and the first determining the first difference exhibited by the first component by the method of determining the second difference exhibited by the first component by the second method; or determining the second difference exhibited by the first component by the method of and wherein the defocus is represented by the first, second, or third difference exhibited by the first component by the method of , configurable.
The charged particle beam apparatus of the present invention further provides a predetermined amount of change in the intensity of the focus correction signal, and the first, second, or third difference exhibited by the first component, the focus From the first, second, or third difference before and after the predetermined amount of change is given to the intensity of the correction signal, and the predetermined amount of change, the first, second or determining the partial derivative of the third difference with respect to the intensity of the focus correction signal as the first partial derivative, and depending on the first, second, or third difference exhibited by the first component It can be configured such that the expressed defocus is converted into an amount of change in intensity of the focus correction signal by the first partial differential coefficient.
With the above configuration, the defocus is evaluated as the amount of change in intensity of the focus correction signal. Therefore, after that, the defocus can be corrected based on the amount of change.

After determining the first partial differential coefficient, the charged particle beam apparatus of the present invention further stores the determined value of the first partial differential coefficient, and thereafter the first component exhibits converting the defocus represented by the first, second, or third difference into a change in intensity of the focus correction signal, the first partial derivative of the stored value, It can be configured for repeated use.
This eliminates the need to recalculate the first partial differential coefficient each time the defocus is measured (each time the conversion is performed).

In the charged particle beam apparatus of the present invention, after converting the defocus into the amount of change in the intensity of the focus correction signal by the first partial differential coefficient, the amount of change in the intensity of the focus correction signal is converted into The converted defocus is subtracted from the intensity of the focus correction signal before the predetermined amount of change is given to the intensity of the focus correction signal, and the focus correction signal having an intensity equal to the result of the subtraction is It can be configured to provide to the focus correction means.
This corrects the defocus.

The charged particle beam device of the present invention is further characterized in that, when the defocus is represented by the second difference exhibited by the first component, the first component of intensity equal to the first average is converted to the When the defocus is given to the two-fold astigmatism correction means and the defocus is represented by the first difference exhibited by the first component, the magnitude of the blur in one of the two directions is minimized. determining, as the first average, the average of the intensity of the first component that reduces the magnitude of the blur and the intensity of the first component that minimizes the magnitude of the other of the two directions of the blur; It can be arranged to apply said first component of intensity equal to its average to said two-fold astigmatism correction means.
As a result, when the defocus is acquired by the first or second method, the defocus is corrected, and the component of the two-fold astigmatism that can be corrected by the first component is corrected.
In the arrangement, the blur measurement for the first component is used not only to determine the first or second difference, but also to determine the first average. Here, the measurement result of the blur for the first component is a measurement value of the magnitude of the blur in each direction for the plurality of intensity values of the first component.
Therefore, in the above configuration, there is no need to obtain these measured values again for the measurement and correction of the two-fold astigmatism. That is, in the above configuration, the time required for measuring and correcting the two-fold astigmatism is shortened accordingly.

The charged particle beam device of the present invention further increases or decreases the intensity of the first component to obtain the defocus when the defocus is represented by the third difference exhibited by the first component. before measuring the magnitude of the blur along a single direction in the illuminated surface while measuring the intensity of the first component of the blur that minimizes the magnitude of the single direction; , the average of the intensity of the first component of the blur in the direction orthogonal to the single direction in the illuminated surface is determined as a second average, and the average is The first component of equal intensity can be provided to the two-fold astigmatism correction means.
Thereby, when the defocus is obtained by the third method, the component of the two-fold astigmatism that can be corrected by the first component is corrected before the defocus is measured. be.
As a result, when the defocus is obtained by the third method, the defocus is measured with higher accuracy, and thus the defocus is corrected with higher accuracy.

In the charged particle beam apparatus of the present invention, further, of the two mutually independent components constituting the two-time astigmatism correction signal, one component other than the first component is referred to as a second component, determining the first, second, or third difference exhibited by the second component, and applying the predetermined amount of change, or a predetermined amount other than the predetermined amount of change, to the intensity of the focus correction signal; and the first, second, or third difference exhibited by the second component, wherein the predetermined amount of change or another predetermined amount of change is applied to the intensity of the focus correction signal. The first, second, or third difference exhibited by the second component is calculated from the difference between the first, second, or third before and after the change and the predetermined amount of change or another predetermined amount of change. determining a partial derivative of the difference with respect to the intensity of the focus correction signal as a second partial derivative; , or after representing by a third difference, and after converting the defocus thus represented into an amount of change in intensity of the focus correction signal by the first partial differential coefficient, then the above The defocus converted into the amount of change in the intensity of the focus correction signal is further converted into the first, second, or third difference exhibited by the second component by the second partial differential coefficient. can be configured as
However, when the defocus is represented by the first difference exhibited by the first component, the first difference exhibited by the second component is determined by the first method and , the second partial derivative is determined from the first difference exhibited by the second component. When the defocus is represented by the second difference exhibited by the first component, the second difference exhibited by the second component is determined by the second method, and the A second partial derivative is determined from the second difference exhibited by the second component. When the defocus is represented by the third difference exhibited by the first component, the third difference exhibited by the second component is determined by the third method, and A second partial derivative is determined from the third difference exhibited by the second component.
The charged particle beam device of the present invention further comprises: the intensity of the second component that minimizes the magnitude of one of the two directions of the blur; determining the average with the intensity of the second component that minimizes the magnitude as another first average; The average of the intensity of the two components and the intensity of the second component that minimizes the magnitude of the blur in the direction orthogonal to the single direction in the irradiated surface is calculated as another second It can be determined as an average or arranged to provide said second component of intensity equal to said determined another first or another second average to said two-fold stigmator means.
However, the alternative first or alternative second average is determined by any of the following three methods (fourth, fifth, and sixth methods).
A fourth method is the intensity of the second component that minimizes the magnitude of one of the two directions of the blur and the first difference exhibited by the second component. Defocusing and defocusing determine said another first average.
The fifth method is the intensity of the second component that minimizes the magnitude of one of the two directions of the blur and the second difference exhibited by the second component. Defocusing and defocusing determine said another first average.
A sixth method is to minimize the magnitude of the blur in the single direction, or to minimize the magnitude of the blur in the direction perpendicular to the single direction within the illuminated surface. The other second average is determined from the intensity of the second component and the defocus converted to the third difference exhibited by the second component.
The fourth method is used when the defocus is represented by the first difference exhibited by the first component. The fifth method is used when the defocus is represented by the second difference exhibited by the first component. The sixth method is used when the defocus is represented by the third difference exhibited by the first component.
With the above configuration, the defocus is corrected, the component of the two-fold astigmatism that can be corrected by the first component is corrected, and the second component of the two-fold astigmatism is corrected. The correctable component is corrected by the component of .
In the above configuration, when the another first or another second average is determined, the magnitude of the blur is not in the two directions, but only in one of the two directions. Measured along only a single direction or along only directions orthogonal to said single direction within said illuminated surface. That is, in the above configuration, the time required for measuring and correcting the two-fold astigmatism is shortened accordingly.

The charged particle beam apparatus of the present invention further stores the value of the determined second partial differential coefficient after determining the second partial differential coefficient, and thereafter stores the value of the determined second partial differential coefficient. to the first, second, or third difference exhibited by the second component of the defocus converted into the amount of change in the intensity of the focus correction signal by The second partial derivative can be configured to be used repeatedly.
This eliminates the need to recalculate the second partial differential coefficient each time the two-fold astigmatism is measured and corrected.

The charged particle beam device of the present invention can be configured as various electron beam devices having an imaging function. Such devices include electron beam writers, electron beam three-dimensional printers, scanning electron microscopes, and transmission electron microscopes.
The charged particle beam device of the present invention can also be configured as various ion beam devices with imaging functions.

<2. Example>
Specific examples of the charged particle beam apparatus of the present invention are described below.

(Example 1)
A basic embodiment of the charged particle beam device of the present invention will be described below as a first embodiment.

In this embodiment, the charged particle beam apparatus of the present invention is configured as a variable shaping electron beam writing apparatus. FIG. 1 shows the configuration of the apparatus (variably shaped electron beam writing apparatus) of this embodiment. FIG. 1 shows the configuration of the optical system, measurement system, and control system of the apparatus of this embodiment.

The optical system of the apparatus of this embodiment, as shown in FIG. 1, has the same configuration as that of the optical system of FIG. (In the optical systems of FIGS. 1 and 24, the same components are denoted by the same reference numerals.) However, in FIG. Components omitted.
The lenses and deflectors in the optical system of the apparatus of this embodiment may be either of the magnetic type or the electrostatic type.
Of these lenses and deflectors, the objective deflector 13 is important in this embodiment. The objective deflector 13 is an 8-pole electrostatic deflector. The objective deflector 13 consists of a conductive and non-magnetic hollow cylinder equally divided into eight at π/4 (45°) intervals. The number of poles of the objective deflector 13 may be other than 8 poles (for example, 12 poles), but hereinafter it is assumed to be 8 poles.

The measuring system of the apparatus of this embodiment comprises a knife edge 20 and a Faraday cup 21, as shown in FIG. Knife edge 20 and Faraday cup 21 are provided on material stage 14 in a separate area from material 10, as shown in FIG. The knife edge 20 and the material 10 have the same surface height position.
Although knife edge 20 is depicted as a single knife edge in FIG. 1 for convenience, knife edge 20 consists of two knife edges, knife edges 20X and 20Y (neither shown). Knife edges 20X and 20Y have edges parallel to the Y and X axes, respectively, and are scanned in the X and Y directions, respectively. Therefore, the knife edge 20 provides the X- and Y-direction components of the position, size, and edge dullness of the projected figure 11 . The position, size, and dullness of edges of the projected figure 11 will be described later.

Supplementally, knife edges 20X and 20Y can each be replaced with electron reflectors. (This also applies to Examples 2 to 11.) Here, the electron reflector is, for example, a fine wire made of heavy metal.
However, when the knife edges 20X and 20Y are replaced with electron reflectors, the electron reflectors are scanned by deflecting the electron beam 1 by the objective deflector 13, and the electrons reflected by the electron reflectors are , is received by a backscattered electron detector (not shown), and the blur of the electron beam 1 is measured based on the output signal of the backscattered electron detector. The backscattered electron detector is provided directly below the objective lens 9, for example.

As shown in FIG. 1, the control system of the apparatus of this embodiment comprises a lens control section 24, an objective deflector control section 25, a material stage control section 26, a Faraday cup absorption current signal processing section 27, a central control section 28, and It consists of a storage unit 29 . The lens control section 24 , objective deflector control section 25 , material stage control section 26 , Faraday cup absorption current signal processing section 27 and storage section 29 are all connected to the central control section 28 .

The apparatus of this embodiment uses its own optical system (in particular, the objective deflector 13) and measurement system to measure its own axial defocus, axial two-fold astigmatism, deflection field curvature aberration, and deflection Measure and correct for astigmatism twice. The blurring of the electron beam 1 is measured by the knife edge method. These things themselves also apply to the optical system of FIG.

However, the apparatus of this embodiment bases the measurement and correction of the axial defocus on a curve different from the curve of the blurring of the electron beam 1 against the focus correction current, and furthermore measures the deflection curvature of field. and the correction is based on a curve different from the electron beam 1 blur curve with respect to the focus correction voltage.
Specifically, the apparatus of this embodiment bases the measurement and correction of the axial defocus on the blur curve of the electron beam 1 with respect to the two-fold astigmatism correction voltage, and measures the deflection curvature of field. And the correction is also based on the blur curve of the electron beam 1 against the twice astigmatism correction voltage. That is, the apparatus of this embodiment increases and decreases the astigmatism correction voltage twice, not the focus correction current or the focus correction voltage, when measuring these aberrations.

The apparatus of this embodiment controls its own optical system and measurement system, including the control therefor, by the control system described above. These controls are performed by the central control unit 28 integrally controlling control system components other than the central control unit 28 .

That is, the central controller 28 controls the lens controller 24 , the objective deflector controller 25 , the material stage controller 26 , and the Faraday cup absorption current signal processor 27 .
The central control unit 28 further provides data necessary for measurement and correction using these control system components, specifically, the above-mentioned axial defocus, axial two-fold astigmatism, deflection curvature of field aberration, and The data necessary for measuring and correcting the deflection 2-fold astigmatism are read from the storage unit 29 . The central control unit 28 performs necessary calculations and controls based on these data, and writes the resulting data into the storage unit 29 .

The lens controller 24, the objective deflector controller 25, the material stage controller 26, and the Faraday cup absorption current signal processor 27 are responsible for the following controls.
A lens control unit 24 controls defocusing on the knife edge 20 (or electron reflector) via the objective lens 9 . The objective deflector control unit 25 controls the deflection of the electron beam 1 via the objective deflector 13 and also controls defocus and two-fold astigmatism on the knife edge 20 . Material stage controller 26 controls the positions of material 10 and knife edge 20 via material stage 14 . The Faraday cup absorption current signal processing unit 27 analog-to-digital converts the Faraday cup absorption current signal and inputs the resulting signal to the central control unit 28 . The Faraday cup absorption current signal will be described later.
Here, the defocus on the knife edge 20 means either or both of the axial defocus and deflection curvature of field aberration, input of the focus correction current to the objective lens 9 and focus correction to the objective deflector 13. It means aberration combined with defocus caused by either or both of the voltage inputs. (Therefore, the defocus on the knife edge 20 changes its magnitude as either or both of the focus correction current and the focus correction voltage are increased or decreased.) The two-fold astigmatism on the knife edge 20 is the above Aberration resulting from synthesis of either or both of axial two-fold astigmatism and deflection two-fold astigmatism, and two-fold astigmatism caused by inputting a two-fold astigmatism correction voltage to the objective deflector 13 do. (Thus, the two-fold astigmatism on the knife edge 20 changes its magnitude as the two-fold astigmatism voltage is increased or decreased.)

Of the lens controller 24, objective deflector controller 25, material stage controller 26, and Faraday cup absorption current signal processor 27, the objective deflector controller 25 is important in this embodiment.
The objective deflector controller 25 applies electrode voltages based on a predetermined electrode voltage distribution to the electrodes of the objective deflector 13 . Here, the predetermined electrode voltage distribution is specifically the electrode voltage distribution followed by the deflection voltage, the distortion correction voltage, the focus correction voltage, and the two-fold astigmatism correction voltage.

FIG. 2 shows the electrode voltage distribution according to the deflection voltage, distortion correction voltage, focus correction voltage, and two-fold astigmatism correction voltage in this embodiment (see Patent Document 2). Here, the distortion correction voltage refers to a voltage added to the deflection voltage for correcting the deflection distortion aberration produced by the apparatus of this embodiment. Among these voltages, however, the deflection voltage and the distortion correction voltage have the same electrode voltage distribution. Hereinafter, description of the distortion correction voltage will be omitted.
In FIG. 2, the eight arcs represent the eight electrodes that make up the objective deflector 13, and the symbols inside the eight arcs are deflection voltages, focus correction voltages, and two-fold astigmatism correction voltages. represents the distribution function describing the electrode voltage distribution of In FIG. 2, the positive direction of the Z-axis is toward the front of the page. That is, the electron beam 1 flows in that direction.

More specifically, in FIG. 2 f _1A and f _1B represent distribution functions describing the electrode voltage distribution of the two orthogonal components of the deflection voltage. f _0A represents the distribution function describing the electrode voltage distribution of the focus correction voltage. _f2A and _f2B represent distribution functions that describe the electrode voltage distribution of the two orthogonal components of the two-fold astigmatic correction voltage. All of these distribution functions are functions of angles around the central axis of the objective deflector 13 . Henceforth, let the angle be (psi).
The angle ψ has its own reference (that is, zero) at the center of the rightmost electrode of the eight electrodes on the paper surface, and the counterclockwise rotation in FIG. 2 is its own positive direction. The angular positions of the electrodes are such that one and the other of the two orthogonal components of the deflection voltage respectively change the position of the electron beam 1 at the height of the material 10 (and the knife edge 20) in the X and Y directions. It is determined. This determines not only the rotation angles of f _1A and f _1B , but also the rotation angles of the remaining distribution functions. (Here, one and the other of the two orthogonal components of the deflection voltage are specifically _V1A and _V1B , respectively. _V1A and _V1B will be described later.)
The _{numbers attached} to _{each arc} in _FIG . (angular position of the electrode center). ψ is equal to an integer multiple of π/4 (=2π/8) at each electrode position. Therefore, at each electrode position, f _1A (ψ), f _0A (ψ), and f _2A (ψ) correspond to cos ψ, 1 (=cos 0), and cos 2ψ, respectively, and f _1B (ψ) and f _2B (ψ) correspond to sin ψ and sin2ψ, respectively. That is, among these distribution functions, f _1A (ψ) and f _1B (ψ) are linearly independent of each other and orthogonal to each other, and f _2A (ψ) and f _2B (ψ) are also linearly independent of each other. and are orthogonal to each other. Here, the fact that these distribution functions are orthogonal to each other means that the phases of these distribution functions are different from each other by π/2.

The components of the electrode voltage of the objective deflector 13 can be represented by the distribution function, and the electrode voltage of the objective deflector 13 can be represented by these components. Here, the electrode voltage components of the objective deflector 13 are specifically electrode voltage components resulting from the deflection voltage, the focus correction voltage, and the two-fold astigmatism correction voltage. The electrode voltage components resulting from the deflection voltage, the focus correction voltage, and the two-fold astigmatism correction voltage are V ₁ , V ₀ , and V ₂ (all real numbers), respectively, and the electrode voltage of the objective deflector 13 is V (real number ), V ₁ , V ₀ , V ₂ , and V can be expressed by equations (1) to (4) as functions of ψ.
V ₁ (ψ)=V _1A f _1A (ψ)+V _1B f _1B (ψ) (1)
V ₀ (ψ)=V _0A f _0A (ψ) (2)
V ₂ (ψ)=V _2A f _2A (ψ)+V _2B f _2B (ψ) (3)
V(ψ)=V ₁ (ψ)+V ₀ (ψ)+V ₂ (ψ) (4)
In equations (1) to (4), V _1A and V _1B (both real numbers) represent two orthogonal components of the deflection voltage, V _0A (real number) represents the focus correction voltage, and V _2A and V _2B ( (both are real numbers) represent two orthogonal components of the two-fold astigmatism correction voltage. V _1A and V _1B , V _0A , and V _2A and V _2B all correspond to the voltages of the electrodes to which the voltage of +1V is applied in FIG. All of these are independent variables and can be controlled independently. Hereinafter, V _1A and V _1B , V _0A , and V _2A and V _2B are simply referred to as deflection voltage, focus correction voltage, and two-fold astigmatism correction voltage, respectively, unless otherwise required.

In the following, the procedure for measuring and correcting axial defocus, axial two-fold astigmatism, deflection field curvature aberration, and deflection two-fold astigmatism produced by the apparatus of this embodiment will be described. .

Of the above four aberrations, the apparatus of this embodiment first measures and corrects the axial defocus produced by itself. The device of this embodiment expresses its axial defocus by the XY difference. The XY difference will be described later.

In the apparatus of this embodiment, first, the material stage 14 moves the knife edge 20 (and the Faraday cup 21) directly below the electron beam 1 in order to measure and correct the axial defocus. The apparatus of this embodiment then measures the blurring of the electron beam 1 with the knife edge 20 while increasing and decreasing the astigmatism correction voltage _V2A twice. In the meantime, giving discrete changes to _V2A and scanning the knife edge 20 with the electron beam 1 are alternately repeated. The apparatus of this embodiment also measures the blurring of the electron beam 1 by the knife edge 20 while increasing or decreasing _V2B . In the meantime, giving discrete changes to _V2B and scanning the knife edge 20 with the electron beam 1 are alternately repeated.
Here, the knife edge 20 is finely scanned in a direction perpendicular to the direction of the edge (Y direction and X direction). That is, each of knife edges 20X and 20Y is finely scanned. The order of fine scanning of knife edges 20X and 20Y does not matter.
The apparatus of this embodiment slightly changes each of the deflection voltages _V1A and _V1B for fine scanning of these knife edges. As a result, the electron beam 1 is slightly deflected along each of the X and Y directions by the objective deflector 13 .

While the knife edge 20 is micro-scanned as described above, the apparatus of this embodiment receives at the Faraday cup 21 the current of the electron beam 1 that changes due to the electron beam 1 being intercepted by the knife edge 20 . By doing so, the apparatus of this embodiment acquires the desired signal, ie, the Faraday cup absorption current signal.
Supplementally, the Faraday cup absorption current signal represents the current in the portion of the electron beam 1 that was not intercepted by the knife edge 20 . If possible, the current interrupted by the knife edge 20 may be detected instead of that current. Even so, a signal similar to the Faraday cup absorption current signal is obtained. Here, the portion of the current intercepted by the knife edge 20 is the portion of the current that is absorbed (or reflected) by the knife edge 20 .

After the Faraday cup absorption current signal is acquired in the manner described above, it is input to the Faraday cup absorption current signal processing unit 27, analog-to-digital converted by the Faraday cup absorption current signal processing unit 27, and sent to the central control unit 28. is entered.
The Faraday cup absorption current signal is then analyzed by central controller 28 . Here, the central control unit 28 differentiates the Faraday cup absorption current signal with respect to position X or Y (or time) and evaluates the resulting signal waveform (differential waveform).
The above signal waveform represents the current density distribution of the projected figure 11 . Therefore, the position, size, and edge bluntness of the projection figure 11 can be obtained from the signal waveform. More specifically, the position of the projected figure 11 can be obtained from the positions of the rising and falling edges of the signal waveform, and the size of the projected figure 11 can also be obtained. edge dullness is obtained. Of these quantities, the dullness of the edge of the projected figure 11 is important in measuring and correcting the axial defocus.

An example of the signal waveform is shown in FIG. In FIG. 3, the solid line and broken line respectively illustrate the above signal waveforms when the electron beam 1 has a small blur and a large blur. That is, the solid line and dashed line respectively illustrate the current density distribution of the projection figure 11 when the blur of the electron beam 1 is small and large.
As shown in FIG. 3, the greater the blurring of the electron beam 1, the greater the bluntness of the edge of the projection figure 11. FIG. (Here, the blunting of the edge of the projected figure 11 is increased means that the gradient of the current density distribution at the edge of the projected figure 11 is gradual.) Accordingly, the blunting of the edge of the projected figure 11 is 1 is an index of blur size.

The above about the signal waveform (differential waveform) is not only for the measurement and correction of the axial defocus, but also for the remaining three aberrations (the axial 2-fold astigmatism, deflection field curvature, and This also applies to the measurement and correction of deflection 2-fold astigmatism. Hereinafter, the blurring of the electron beam 1 is measured according to the above-described procedure (fine scanning of the knife edge 20, acquisition of the Faraday cup absorption current, and analysis of the signal waveform) regardless of the difference in aberration.

During the above process (measurement of the blur of the electron beam 1 while increasing and decreasing the astigmatism correction voltage _V2A or _V2B twice), the apparatus of this embodiment changes the values of the focus correction current and the focus correction voltage to is also fixed to zero. In the apparatus of this embodiment, the focus correction condition and the focus correction current during that period are set to the 0th focus correction condition and the 0th focus correction current, respectively. Hereinafter, the focus correction current is represented by _I0A , and the n-th (n=0 or 1) focus correction current is represented by _I0An .
However, the reason why the apparatus of this embodiment fixes the values of both the focus correction current and the focus correction voltage to zero during the above process is that the on-axis defocus and deflection are performed immediately after the apparatus of this embodiment is assembled. This is the case when none of the field curvature aberrations have yet been corrected, so that both the values of the focus correction current and the focus correction voltage are zero. Otherwise, i.e., if either or both of the focus correction current and the focus correction voltage have non-zero values, during the above process, the respective values of the focus correction current and the focus correction voltage are set to the above axis It suffices to fix them to their respective values immediately before the upper defocus is measured. That is, the 0th focus correction current can have a non-zero value. In the following, regardless of this, the values of the focus correction current and the focus correction voltage are both fixed at zero during the above process. Given this, _I0A0 is zero.

In the above process, the apparatus of this embodiment measures the blur of the electron beam 1 as described above to obtain curves of the blur in the X direction and the Y direction of the electron beam 1 with respect to the two-time astigmatism correction voltage V _2A , Curves of blurring in the X and Y directions of the electron beam 1 with respect to the double astigmatism correction voltage _V2B are obtained. Each of these curves is an approximation curve to the blur measurements obtained for multiple values of the twice astigmatic voltage in the above process.
The apparatus of this embodiment then determines the location of the valleys of these curves. That is, from these curves, V _{2A for minimum blur in the X direction, V 2A for} minimum blur in the Y direction, V 2B for minimum blur in the X direction, and V _2B for minimum blur in the Y direction. _2B is determined. If the approximate curves are analytic functions (for example, hyperbolas), it is possible to analytically determine the positions of the troughs of these curves.

In the above process, in order to determine the positions of the troughs of these curves, it is necessary to provide a range of increase and decrease of the two-time astigmatism correction voltage with sufficient width to confirm the troughs of each of these curves.
The width tends to increase immediately after assembly of the apparatus of this embodiment, that is, when the excitation current is first applied to the objective lens 9 and the reduction lens 8 . This is because the appropriate value of the excitation current for the objective lens 9 has not been determined at that time.
The same thing can happen if the excitation currents for these lenses are turned off and then turned back on. This involves the magnetic hysteresis of the magnetic material of these lenses.

The apparatus of this embodiment defines each of the two-time astigmatism correction voltages _V2A and _V2B before increasing or decreasing the two-time astigmatism correction voltage as described above as the 0th two-time astigmatism generation voltage, Each of _V2A and V2B that minimizes blur in the X direction is defined as the first two-fold _astigmatic voltage, and each of _V2A and _V2B that minimizes blur in the Y direction is defined as the second voltage. It is defined as a two-fold astigmatism voltage. The reason why the two-fold astigmatism correction voltage is referred to as the two-fold astigmatism generation voltage in this embodiment is that the two-fold astigmatism is positively measured in order to measure the defocus on the knife edge 20 in the present invention. by spontaneously generating

The apparatus of this embodiment further defines the difference between the first two-fold astigmatic voltage and the second two-fold astigmatic voltage as the XY difference.
The XY difference is an index of the degree of defocus on the knife edge 20. FIG. That is, if the XY difference is zero, the defocus on the knife edge 20 is also zero, but unless the XY difference is zero, the defocus is not zero. to reduce

The apparatus of this embodiment defines the XY difference for each focus correction condition. Hereinafter, the XY difference under the n-th (n=0 or 1) focus correction condition is referred to as the n-th XY difference. The apparatus of this embodiment also defines an XY difference for each component of the twice astigmatic voltage.
More specifically, the apparatus of this embodiment acquires curves of blurring in the X direction and Y direction of the electron beam 1 with respect to V _2A under the n-th (n=0 or 1) focus correction condition, and these curves , and the difference between these voltages (XY difference) is defined as the A component of the nth XY difference. The apparatus of this embodiment further obtains curves of blurring in the X direction and Y direction of the electron beam 1 with respect to V _2B under the same focus correction conditions, and from these curves, the first and second double astigmatism Determine the voltages and define the difference between these voltages (XY difference) as the B component of the nth XY difference.

To illustrate the XY difference, an example of the X and Y blur curves of electron beam 1 _versus V _2A is shown in FIG. An example is shown in FIG. 4B. Also shown in FIGS. 4A and 4B are the zeroth two-fold astigmatic voltage 50, the first two-fold astigmatic voltage 51, and the second two-fold astigmatic voltage 52. FIG.
4A and 4B both assume that the two-fold astigmatism on the knife edge 20 is zero when both _V2A and _V2B are zero. Figures 4A and 4B also assume that _V2B and _V2A are zero, respectively. Thus, in FIG. 4A, the two-fold astigmatism on knife edge 20 is zero when _V2A is zero, and in FIG. 4B, two-fold astigmatism on knife edge 20 is zero when _V2B is zero. Aberration becomes zero. FIGS. 4A and 4B further assume that the 0th two-fold astigmatic voltage 50 is zero.

4A and 4B, the XY difference is the position of the valley (first double astigmatic voltage 51) of the curve (solid line) representing the blur in the X direction and the position of the curve (broken line) representing the blur in the Y direction. It corresponds to the difference from the valley position (second double astigmatic voltage 52). If the XY difference is zero, the locations of these valleys coincide with each other. The A and B components of the XY difference are obtained from FIGS. 4A and 4B, respectively.

Hereinafter, the A component and the B component of the XY difference are represented as D _A and D _B (both are real numbers), and the n-th (n=0 or 1) XY difference D _A and D _B are represented as D _An and D _Bn . For example, the A and B components of the 0th XY difference are denoted D _A0 and D _B0 , respectively. Hereinafter, D _A , D _B , and I _0A are also collectively referred to as D _An , D _Bn , and I _0An (where n=0 or 1), respectively.
For convenience of explanation, the signs of D _A and D _An are positive when the second two-fold astigmatic voltage 52 is greater than the first two-fold astigmatic voltage 51, and D _B and D _Bn are positive. The sign is positive when the first two-fold astigmatic voltage 51 is larger than the second two-fold astigmatic voltage 52 . Following this rule, FIGS. 4A and 4B yield D _A and D _B , respectively, with a positive sign.

4A and 4B, the first two-time astigmatism voltage 51 and the second two-time astigmatism voltage 52 have the same absolute value and opposite signs. This is because both FIGS. 4A and 4B assume that the two-fold astigmatism on the knife edge 20 is zero when both _V2A and _V2B are zero, as described above. .

The above assumption does not always hold. That is, there may be cases where the two-fold astigmatism on the knife edge 20 when both _V2A and _V2B are zero is not zero. Here, the 2-fold astigmatism on the knife edge 20 when both _V2A and _V2B are zero is nothing but the axial 2-fold astigmatism produced by the apparatus of this embodiment.
Examples of X- and Y-blur curves of the electron beam 1 versus V _2A (or V _2B ) obtained when the on-axis two-fold astigmatism is non-zero are shown in FIGS. 4C and 4D. FIG. 4C shows the relationship when the value of the 0th two-fold astigmatic voltage 50 is zero. FIG. 4D shows the relationship when the value of the 0th two-fold astigmatism voltage 50 is not zero. Here, the fact that the value of the 0th two-fold astigmatism generation voltage 50 is not zero corresponds to the fact that the axial two-fold astigmatism has been corrected in advance by _V2A and _V2B .

If the axial two-fold astigmatism has a non-zero value, the first two-fold astigmatism voltage 51 and the second two-fold astigmatism voltage 52 shift in the horizontal direction. Here, the first two-time astigmatic voltage 51 and the second two-time astigmatic voltage 52 shift in the same direction and by the same magnitude. This is the effect of the above-mentioned on-axis two-fold astigmatism. As a result of this action, the first two-time astigmatic voltage 51 and the second two-time astigmatic voltage 52 have different absolute values, as shown in FIGS. 4C and 4D.

Even if the first two-fold astigmatism voltage 51 and the second two-fold astigmatism voltage 52 are shifted in this way, the XY difference (the first two-fold astigmatism The difference between the astigmatism-generated voltage 51 and the second astigmatism-generated voltage 52) is not changed. This is the result of these voltage shifts being in the same direction and magnitude as each other.
This is independent of the magnitude of the shift. This is also independent of the value of the 0th two-fold astigmatism voltage 50 . That is, this is true even if the axial two-fold astigmatism was previously corrected by V _2A and V _2B , as shown in FIG. 4D.

That is, the axial two-fold astigmatism can cause determination errors of the first two-fold astigmatism voltage 51 and the second two-fold astigmatism voltage 52, but these determination errors are associated with the shift. Regardless of the magnitude or value of the 0th two-fold astigmatism voltage 50, the XY difference is canceled at the time it is determined. This is the reason why the apparatus of the present embodiment obtains the difference between the first two-fold astigmatic voltage 51 and the second two-fold astigmatic voltage 52 rather than the difference between them (XY difference).

The apparatus of this embodiment first measures the 0th XY difference D _A0 and D _B0 under the 0th focus correction condition, and converts the measured D _A0 and D _B0 to the 0th focus correction current. Store with _I0A0 . Here, D _A0 and D _B0 each represent the defocus on the knife edge 20 under the 0th focus correction condition. The defocus is equal to the on-axis defocus produced by the apparatus of this embodiment. This is because _I0A0 is zero in this embodiment as described above. That is, D _A0 and D _B0 each represent the axial defocus produced by the apparatus of this embodiment.
Next, the apparatus of this embodiment changes the defocus on the knife edge 20 by giving a change amount to the focus correction current _I0A , and the focus correction condition and the focus correction current _I0A at that time are changed to Assume that the first focus correction condition and the first focus correction current are _I0A1 . The apparatus of this embodiment then measures the first XY differences D _A1 and D _B1 under the first focus correction conditions, and converts the thus measured D _A1 and D _B1 to the first focus correction current Store with _I0A1 .

Next, the apparatus of this embodiment calculates the partial differential coefficients of D _A and D _B with respect to I ₀ A from the amount of change in the focus correction current I ₀ A and the amount of change in the XY differences D _A and D _B in the above process. decide. The apparatus of this embodiment then stores the values of these partial derivatives.
These partial differential coefficients represent the XY difference produced by the unit focus correction current. Assuming that these partial differential coefficients are d _IA and d _IB (both are real numbers), d _IA and d _IB are given by equations (5) and (6), respectively.

（５）および（６）式において、ΔＩ_０Ａ（＝Ｉ_０Ａ１－Ｉ_０Ａ０）は、Ｉ_０Ａの変化量を表し、ΔＤ_Ａ（＝Ｄ_Ａ１－Ｄ_Ａ０）およびΔＤ_Ｂ（＝Ｄ_Ｂ１－Ｄ_Ｂ０）は、それぞれＤ_ＡおよびＤ_Ｂの変化量を表す。
補足すれば、（５）式におけるＩ_０Ａ１、Ｉ_０Ａ０、およびΔＩ_０Ａは、それぞれ、（６）式におけるＩ_０Ａ１、Ｉ_０Ａ０、およびΔＩ_０Ａと同じであってもよいし、異なっていてもよい。

In equations (5) and (6), ΔI _0A (=I _0A1 −I _0A0 ) represents the amount of change in I _0A , and ΔD _A (=D _A1 −D _A0 ) and ΔD _B (=D _B1 −D _B0 ) represent the amount of change in D _A and D _B , respectively.
Supplementally, I _{0A1 , I 0A0 , and ΔI 0A in formula (5) may be the same as or different from I 0A1 , I 0A0 , and ΔI 0A in formula (} 6), respectively. .

Using the coefficients _{d_IA} and _{d_IB} , the XY differences D _A and D _B can be expressed by equations (7) and (8), respectively. In equations (7) and (8), _DOA and _DOB are both constants.
D _A =D _OA +d _IA I _0A (7)
D _B =D _OB +d _IB I _0A (8)
If the equations (7) and (8) are illustrated, the equations (7) and (8) are as shown in FIG. 5, for example. Two straight lines (dashed lines) in FIG. 5 are a straight line representing the relationship between D _A and I _0A and a straight line representing the relationship between D _B and I _0A . The former and latter of these straight lines have slopes equal to the coefficients _{d_IA} and _{d_IB} , respectively.

The apparatus of this embodiment then updates I _0A0 according to equation (9) to make D _A0 zero, or updates I _0A0 according to equation (10) to make D _B0 zero. In equations (9) and (10), I _0A0 ′ represents I _0A0 updated to zero D _A0 or D _B0 .
I _0A0 ′=I _0A0 −D _A0 /d _IA =I _0A0 −I _D0 (9)
I _0A0 ′=I _0A0 −D _B0 /d _IB =I _0A0 −I _D0 (10)

Equation (9) can be derived from equations (11) and (12). Equation (10) can be derived from equations (13) and (14).
D _A0 =D _OA +d _IA I _0A0 (11)
D _A0 ′=0=D _OA +d _IA I _0A0 ′ (12)
D _B0 =D _OB +d _{IBI 0A0 (} 13)
D _B0 ′=0=D _OB +d _{IBI 0A0 ′} (14)
Formula (11) is a formula obtained by replacing D _A and I _0A in formula (7) with D _A0 and I _0A0 , respectively. Formula (12) is a formula obtained by replacing D _A0 and I _0A0 in formula (11) with D _A0 ′ (=0) and I _0A0 ′, respectively. D _A0 ′ represents D _A0 obtained by inputting I _0A0 ′ to the objective lens 9 . Formula (13) is a formula obtained by replacing D _B and I _0A in formula (8) with D _B0 and I _0A0 , respectively. Formula (14) is a formula obtained by replacing D _B0 and I _0A0 in Formula (13) with D _B0 ′ (=0) and I _0A0 ′, respectively. D _B0 ′ represents D _B0 obtained by inputting I _0A0 ′ to the objective lens 9 .

_ID0 in equations (9) and (10) are given by equations (15) and (16), respectively. I _D0 in equations (9) and (15) represents the XY difference D _A0 converted into the amount of change in the focus correction current. I _D0 in equations (10) and (16) represents the XY difference D _B0 converted to the amount of change in the focus correction current. Equations (15) and (16) indicate that the apparatus of this embodiment converts the XY difference into the amount of change in the focus correction current. The conversion is by factor _dIA or _dIB .
I _D0 =D _A0 /d _IA (15)
I _D0 =D _B0 /d _IB (16)
Here, _ID0 also represents the axial defocus converted into the amount of change in the focus correction current. This is because the XY differences D _A0 and D _B0 each represent the axial defocus as described above.
Therefore, equations (9) and (10) show that the device of this embodiment subtracts the above-described axial defocus converted to the amount of change in the focus correction current from _I0A0 , and thereby updates _I0A0 ( That is, it indicates that I _0A0 ′ is determined).

The measured value of the axial defocus may be D _A0 or D _B0 , but may be I _D0 (=D _A0 /d _IA =D _B0 /d _IB ). Hereinafter, for convenience of explanation, the measured value of the axial defocus is assumed to be _ID0 .

If the measured value of the axial defocus is I _D0 , the correction value of the axial defocus is −I _D0 (=I _0A0 ′−I _0A0 =−D _A0 /d _IA =−D _B0 /d _IB )be equivalent to. That is, the measured value and the corrected value of the axial defocus have the same magnitude and opposite signs.
Supplementally, the measured value I _D0 and the correction value -I _D0 of the axial defocus are respectively the measured value and correction of the axial defocus obtained by measuring the axial defocus by the null method. equal to the value Here, measuring the axial defocus by the null method means measuring and correcting the axial defocus based on a curve of blurring of the electron beam 1 with respect to the focus correction current. The manner of measurement and correction thereof is the same as the manner in which the same aberrations are measured and corrected in the optical system of FIG.

I _0A0 ′ may be determined from either one of equations (9) and (10). More specifically, I _0A0 ′ may be determined from any one of Eqs. (9) and (10) including non-zero D _A0 or D _B0 . That is, the apparatus of this embodiment represents the defocus on the knife edge 20 by the non-zero one of D _A0 and D _B0 .
In this embodiment, as will be described later, when the defocus on the knife edge 20 is not zero, either one of D _A and D _B may be zero, but both D _A and D _B are zero. will not be. Also in this embodiment, either _{d_IA} or _{d_IB} can be zero, but both _{d_IA} and _{d_IB} cannot be zero.
Note that if either of the D _A0 and D _B0 values are evaluated without regard to whether they are zero or not, then the defocus on the knife edge 20 is not zero. is erroneously determined to be zero. This situation will be discussed later.

More preferably, I _0A0 ' can be determined from formula (9) when formula (17) holds, and can be determined from formula (10) when formula (18) holds. good. This is from the point of view of measurement accuracy.
|D _A0 |≧|D _B0 | (17)
|D _A0 |<|D _B0 | (18)
That is, in the apparatus of this embodiment, the defocus on the knife edge 20 is represented by either one of D _A and D _B whose absolute value is not small from the viewpoint of measurement accuracy.
Supplementally, the reason that equation (17) includes the case where |D _A0 |=|D _B0 | holds, while equation (18) does not include that case is for convenience of explanation. If |D _A0 |=|D _B0 | holds, I _0A0 ′ may be determined from either equation (9) or (10).

If the formula (17) holds, then the formula (19) holds. On the other hand, if the formula (18) holds, the formula (20) holds. That is, under the condition that the defocus on the knife edge 20 is not zero, either the equations (17) and (19) hold simultaneously, or the equations (18) and (20) hold simultaneously. This is because D _A0 , D _B0 , d _IA , and d _IB satisfy equation (21). Equation (21) is derived from equations (9) and (10).
|d _IA |≧|d _IB | (19)
|d _IA |<|d _IB | (20)
D _A0 /D _B0 =d _IA /d _IB (21)
Therefore, I _0A0 ′ can be determined from the formula (9) when the formula (19) holds, and can be determined from the formula (10) when the formula (20) holds.

Which of the formulas (9) and (10) should be used to determine I _0A0 ′ is thus determined by the magnitude relationship between |d _IA | and |d _IB |. That is, which of the formulas (9) and (10) should be used to determine I _0A0 ′ _depends on _{whether D An ,} and D _Bn substituted in equation (6) is determined).
This means that either _dIA or _dIB , rather than both _dIA and _dIB , need to be stored in this embodiment. Here, d _IA and d _IB do not depend on any of I _0A , V _0A , V _2A and V _2B and are therefore invariant.

This also _translates into the X direction and _Y It is not necessary to obtain both the directional blur curves (2 total) and the X and Y blur curves (2 total) of the electron beam 1 for _V2B (2×2=4 total). means not.
That is, the curves required for determining I _0A0 ′ are either the former or the latter (two curves in total). This reduces the time required to determine I _0A0 ′ and thus the time required to measure and correct for the axial defocus.
However, it is not determined which of equations (19) and (20) holds until _{d_IA} and _{d_IB} are determined first. Therefore, when first determining _dIA and _dIB , the X and Y blur curves of electron beam 1 for V _2A and the X and Y blur curves of electron beam 1 for V _2B need to get both.
Supplementally, the number of curves required to determine I _0A0 ′ (i.e., update I _0A0 ) is are also two in total. Here, the case where the measurement and correction of the axial defocus is based on the curve of the blurring of the electron beam 1 with respect to the focus correction current, that is, in the optical system of FIG. Equal if measured and corrected. Specifically, the total of two curves is one curve of blurring in the X direction of the electron beam 1 with respect to _I0A0 and one curve of blurring in the Y direction of the electron beam 1 with respect to _I0A0 .
This is because, unless the on-axis two-fold astigmatism on the knife edge is zero, the positions of the _troughs of these curves do not coincide with each other. to the average of the trough locations of these curves, rather than to either one of them. Here, the positions of the troughs of these curves do not coincide with each other because the axial two-fold astigmatism on the knife edge presents two focal lines with different heights. The directions of these focal lines are orthogonal to each other.

The apparatus of this embodiment determines I _0A0 ′ from equation (9) or (10) and inputs I _0A0 ′ thus determined to the objective lens 9 . This causes D _A0 and D _B0 to be zero. At this time, the defocus on the knife edge 20 becomes zero. That is, the axial defocus is corrected.
It should be noted that inputting I _0A0 ′ determined from equations (9) or (10) into the objective lens 9 makes it zero, not either D _A0 or D _B0 , but D To be both _A0 and D _B0 . This will be discussed later.

However, even if I _0A0 ′ obtained from equation (9) or (10) is input to the objective lens 9, the XY differences D _A0 and D _B0 do not become completely zero. Defocus does not become completely zero either. That is, the correction residual of the axial defocus remains. This involves errors included in D _A0 and d _IA in equation (9) or errors included in D _B0 and d _IB in equation (10), as well as defocusing caused by objective lens 9. error is involved.

If the correction residual causes a problem, the evaluation and reduction of the correction residual may be repeated. Specifically, the evaluation and reduction may be performed as follows.
First, the X-direction and Y-direction blur curves of the electron beam 1 with respect to V _2A are obtained again, and D _A0 ' is determined as the corrected residual from the obtained curves. Next, D _A0 ' thus determined is _regarded as a new D _A0 , and is _substituted into the equation (9). (9) Substitute into the formula. The new I _0A0 ′ determined in doing so is then input to the objective lens 9 .
Alternatively, first, the X-direction and Y-direction blur curves of the electron beam 1 with respect to V _2B are obtained again, and from the curves thus obtained, D _B0 ' is determined as the correction residual. Next, D _B0 ′ thus determined is _regarded as a new D _B0 , and is _substituted into the equation (10). (10) Substitute into the formula. The new I _0A0 ′ determined in doing so is then input to the objective lens 9 .
From the acquisition of the curves (E-beam 1 X and Y blur curves for V _2A or E-beam 1 X and Y blur curves for V _2B ), the new I _0A0 ′ objective By repeating the sequence of steps up to the input to lens 9, D _A0 and D _B0 approach zero, and thus the defocus on knife edge 20 approaches zero.
Supplementally, such evaluation and reduction of correction residuals is also necessary if the measurement and correction of the axial defocus is based on the blur curve of the electron beam 1 versus the focus correction current. Here, the correction residual error for the axial defocus is an excess or deficiency of the correction value for the axial defocus. The excess or deficiency increases as the correction value for the axial defocus increases. (The same applies to the deflection curvature of field mentioned above.)

Determination of new I _0A0 ′ (that is, update of I _0A0 ) in the above series of steps is according to equation (9a) or (10a).
I _0A0 ^(m) =I _0A0 ^(m-1) -D _A0 ^(m-1) /d _IA =I _0A0 ^(m-1) -I _D0 ^(m-1) (9a)
I _0A0 ^(m) =I _0A0 ^(m-1) -D _B0 ^(m-1) /d _IB =I _0A0 ^(m-1) -I _D0 ^(m-1) (10a)
In equations (9a) and (10a), I _0A0 ^(m) represents I _0A0 updated for the m (≧1)th time. D _A0 ^(m−1) and D _B0 ^(m−1) represent D _A0 and D _B0 obtained by inputting I _0A0 ^(m−1) into the objective lens 9, respectively.
In equations (9a) and (10a), I _0A0 ^(m) represents I _0A0 ′ and I _0A0 ^(m−1) represents I _0A0 when m=1. That is, equations (9a) and (10a) generalize equations (9) and (10), respectively. Therefore, hereinafter, unless otherwise specified, formulas (9a) and (10a) also serve as formulas (9) and (10), respectively.

I _D0 ^(m−1) in equations (9a) and (10a) are given by equations (15a) and (16a), respectively. I _D0 ^(m−1) in equations (9a) and (15a) represents the XY difference D _A0 ^(m−1) converted into the amount of change in the focus correction current. I _D0 ^(m−1) in equations (10a) and (16a) represents the XY difference D _B0 ^(m−1) converted into the amount of change in the focus correction current.
I _D0 ^(m−1) = D _A0 ^(m−1) /d _IA (15a)
I _D0 ^(m−1) = D _B0 ^(m−1) /d _IB (16a)
Here, I _D0 ^(m−1) also represents the defocus on the knife edge 20 (however, the defocus converted to the amount of change in the focus correction current). This is because both the XY differences D _A0 ^(m−1) and D _B0 ^(m−1) represent defocus on the knife edge 20 . (I _D0 ^(m−1) is also a measure of defocus on the knife edge 20.)

While the above series of steps are repeated, the coefficient _{d_IA} or _{d_IB} is repeatedly used without being updated. More specifically, the _dIA determined and stored in the previous process is used repeatedly each time the _dIA converts D _A0 to I _D0 (see equations (15) and (15a)). , or the _dIB determined and stored in the previous process is used repeatedly for each conversion of _DBO to _{ID0 by dIB (} see equations (16) and (16a)). Such repeated use of coefficients _{d_IA} or _{d_IB} assumes that these coefficients are unchanged. As previously mentioned, these coefficients are independent of the focus correction current _I0A and are therefore unchanged.
Strictly speaking, the dependence of these coefficients on the focus correction current _IOA is not zero, so due to that dependence the coefficients _{d_IA} and _{d_IB} may have errors. However, the error does not matter. Even if the error is not zero, repeating the above series of steps brings D _A0 and D _B0 closer to zero, and therefore the defocus on the knife edge 20 approaches zero, as described above. (The same applies to the coefficients _dVA and _dVB , which will be discussed later.)

The above series of steps may be repeated until formula (22) is established when _I0A0 is updated by formula (9a), and when _I0A0 is updated by formula (10a), (23 ) is repeated until the formula holds.
|D _A0 ^(m) |<ε _D (22)
|D _B0 ^(m) |<ε _D (23)
In equations (22) and (23), ε _D (>0) represents the tolerance for |D _A0 ^(m) | and |D _B0 ^(m) |. That is, |D _A0 ^(m) | and |D _B0 ^(m) | correspond to the correction residual error of the axial defocus.
Accordingly, the establishment of formula (22) or (23) means sufficient reduction of the correction residual of the axial defocus. If D _A0 ^(m) satisfies equation (22) or D _B0 ^(m) satisfies equation (23), then update I _0A0 according to equation (9a) or (10a) (that is, I _0A0 ^(m+1) determination) and the input of _I0A0 thus updated to the objective lens 9 becomes unnecessary. (The conversion of D _A0 ^(m) or D _B0 ^(m) to I _D0 ^(m) is no longer necessary.) The same applies to the remaining three aberrations (above-mentioned axial two-fold astigmatism, deflection image This is also true for correction residuals for surface curvature aberrations, and deflection 2-fold astigmatism). Correction residuals of these aberrations will be described later.

Supplementally, depending on the magnitudes of D _A0 ^(m) and D _B0 ^(m) , one of equations (22) and (23) may hold but the other may not hold.
This means that evaluation of the smaller absolute value of D _A0 and D _B0 should be avoided in this example. If, in this example, whichever of D _A0 and D _B0 has the smaller absolute value is evaluated, (22) even though the defocus on the knife edge 20 is not sufficiently reduced in practice. and (23), it may be erroneously determined that the defocus has been sufficiently reduced.
In principle, such erroneous determination can be prevented by reducing the allowable value _εD . However, in practice, doing so makes the establishment (or non-establishment) of equation (22) or (23) more strongly influenced by the measurement error of D _A0 or D _B0 . Therefore, it is not a good idea to evaluate one of D _A0 and D _B0 , whichever has the smaller absolute value, in this embodiment.

According to the above procedure, the axial defocus is measured and corrected based on the blur curve of the electron beam 1 with respect to the two-fold astigmatism correction voltage, not the curve of the blur of the electron beam 1 with respect to the focus correction current. be able to.
If the measurement and correction of the axial defocus are based on the curves of the blurring of the electron beam 1 with respect to the two-fold astigmatism correction voltage, the measurement and correction of the axial defocus can be expected from the short time required to acquire these curves. The time required for correction is shortened. The short time required to acquire these curves is due to the quick response of the objective deflector 13 as an electrostatic double stigmator to the double stigmator voltage.

The above is the description of the measurement and correction of the axial defocus produced by the apparatus of this embodiment. The measurement and correction routine is shown in FIG. 6 as main routine 1 .
However, both the main routine 1 and the subroutines therein (subroutine 1 and subroutine 2) are premised on the establishment of the formula (19). That is, in these main routines and subroutines, only _dIA of the coefficients _dIA and _dIB is determined, and only _DAn is determined of the XY differences _DAn and _DBn .

(Main routine 1)
As shown in FIG. 6, the main routine 1 consists of the following steps.
First, in step S1, n is set to zero (n=0).
Next, in step S2, I _0An (I _0A0 ) is read and input to the objective lens 9 . (Here, _I0A0 is equal to the value of the focus correction current _I0A immediately before the start of main routine 1. Therefore, inputting _I0A0 to the objective lens 9 in step S2 will eventually change the value of _I0A to I0A0. Equivalent to keeping it at _0A0 .)
Next, in step S3, subroutine 1 (see FIG. 7) is executed to determine the XY difference D _An (D _A0 ).
Next, in step S4, n is set to 1 (n=1).
Next, in step S5, I _0An (I _0A1 ) is read and input to the objective lens 9 .
Next, in step S6, subroutine 1 (see FIG. 7) is executed to determine the XY difference D _An (D _A1 ).
Next, in step S7, the coefficient _{d_IA} is determined.
Next, in step S8, subroutine 2 (see FIG. 8) is executed to determine I _0A0 ^(m) . Then, the main routine 1 ends.

(Subroutine 1)
Subroutine 1 consists of the following steps, as shown in FIG.
First, in step S11, the astigmatism correction voltage _V2A is increased and decreased twice to measure the blurring of the electron beam 1 in the X and Y directions. By doing so, curves of X and Y blurring of the electron beam 1 against V _2A are obtained.
Next, in step S12, the XY difference D _An (in step S23, D _A0 ^(m) ) is determined. Then, the subroutine 1 is terminated and the main routine 1 (see FIG. 6) is returned to. (Alternatively, exit subroutine 1 and return to any of subroutine 2, main routine 2, and subroutine 4. These routines and subroutines will be described later.)

(Subroutine 2)
Subroutine 2 consists of the following steps, as shown in FIG.
First, in step S21, n is set to 0 (n=0) and m is set to 0 (m=0).
Next, I _0A0 ^(m) is input to the objective lens 9 in step S22. (I _0A0 ^(m) represents I _0A0 when m=0. Here, I _0A0 is equal to the value of the focus correction current I _0A immediately before the start of main routine 1 or main routine 1′. Main routine 1' will be described later.)
Next, in step S23, subroutine 1 (see FIG. 7) is executed to determine the XY difference D _An ^(m) (D _A0 ^(m) ). Here, D _An ^(m) represents D _An obtained by inputting I _0A0 ^(m) to the objective lens 9 . (If subroutine 2 is called from main routine 1, step S23 when m=0 can be replaced with a step of reading the XY difference D _A0 determined in step S3.)
Next, in step S24, it is determined whether or not |D _A0 ^(m) |<ε _D holds. If so, exit subroutine 2 and return to main routine 1 (see FIG. 6) (or return to main routine 1'). If it does not hold, the process proceeds to step S25.
In step S25, m+1 is substituted for m (m=m+1). Then, the process proceeds to step S26.
In step S26, I _0A0 ^(m) is determined. Then, the process returns to step S22.

If the coefficient _{d_IA} is known, steps S1-S7 can be replaced by a step of reading _{d_IA} (step S7'). Main routine 1 simplified by doing so is shown in FIG. 9 as main routine 1'.

(Main routine 1')
As shown in FIG. 9, the main routine 1' consists of the following steps.
First, in step S7', the coefficient _{d_IA} is read.
Next, in step S8', subroutine 2 (see FIG. 8) is executed to determine I _0A0 ^(m) . Then, the main routine 1' ends.

Supplementally, the main routine 1 or main routine 1' is preferably executed after the electron beam 1 is aligned with the center of the objective lens 9. FIG. This is because such alignment of the electron beam 1 reduces the overall aberrations appearing on the knife edge 20 (including the above-described axial defocus). Such alignment of the electron beam 1 can be achieved, for example, by an aligner (not shown) provided between the second shaping aperture plate 7 and the reduction lens 8 or between the reduction lens 8 and the objective lens 9. , is possible.
In the above alignment, whether or not the electron beam 1 passes through the center of the objective lens 9 is determined by increasing or decreasing the acceleration voltage of the electron beam 1 or increasing or decreasing the excitation current of the objective lens 9. This can be seen by looking at the change in position of the electron beam 1 in position. The change is zero if the electron beam 1 passes through the center of the objective lens 9, but not zero if the electron beam 1 does not pass through the center of the objective lens 9. The change can be measured by the knife edge method.
Alignment and measurement as described above are widely carried out not only in variable-shaped electron beam writing systems but also in general charged particle beam systems.

The apparatus of this embodiment then measures and corrects for its own on-axis 2-fold astigmatism. The procedure is the same as the procedure for measuring and correcting the same aberrations in the optical system of FIG. The details are as follows.

The apparatus of this embodiment first measures the X-direction and Y-direction blurring of the electron beam 1 with the knife edge 20 while increasing and decreasing _V2A , thereby determining the X-direction and Y-direction blurring of the electron beam 1 with respect to _V2A . (2 curves in total) are acquired. Supplementally, these curves are acquired immediately after the above-described axial defocus measurement and correction are completed, so it is necessary to move the knife edge 20 again by the material stage 14 in order to acquire these curves. no.
The apparatus of this embodiment then determines the XY average for V2A from these curves, and the XY average thus determined is the _A component of the XY average. Here, the XY average refers to the average of the first two-time astigmatic voltage 51 and the second two-time astigmatic voltage 52 .
The apparatus of this embodiment then updates V _2A based on the A component of the XY average, and inputs the updated V _2A to the objective deflector 13 . As a result, the component of the axial two-fold astigmatism that can be corrected by _V2A is corrected.
The apparatus of this embodiment then measures the X- and Y-direction blurring of the electron beam 1 with the knife edge 20 while increasing and decreasing _V2B , thereby determining the X- and Y-direction blurring of the electron beam 1 with respect to _V2B . Acquire the blur curves (two in total).
The apparatus of this embodiment then determines the XY average for V2B from these curves, and takes the thus determined XY average as the _B component of the XY average.
The apparatus of this embodiment then updates V _2B based on the B component of the XY average, and inputs the updated V _2B to the objective deflector 13 . As a result, the component of the axial two-fold astigmatism that can be corrected by _V2B is corrected.
Supplementally, in the above process, the determination of the _B component of the XY average and the input of V2B are performed after the determination of the _A component of the XY average and the input of V2A. may

The updating of _V2A and _V2B in the above process can be expressed by equations (24) and (25), respectively.
V _2A0 ′=M _A0 (24)
V _2B0 ′=M _B0 (25)
In equations (24) and (25), V _2A0 ′ and V _2B0 ′ represent V _2A0 and V _2B0 updated by the above process, respectively. V _2A0 and V _2B0 represent V _2A and V _2B , respectively, under the 0th focus correction condition, but more specifically V _2A and V _2B , respectively, before the above curves are acquired. M _A0 and M _B0 represent M _{A and M B, respectively, under the 0th focus correction condition, more specifically M A and M B , respectively} , determined from the above curves. M _A and M _B represent the A and B components of the XY average, respectively.

However, even if V _2A0 ′ represented by the equation (24) and V _2B0 ′ represented by the equation (25) are input to the objective deflector 13, the axial two-fold astigmatism is completely does not become zero. That is, the correction residual error of the axial two-fold astigmatism remains.

If the correction residual causes a problem, the evaluation and reduction of the correction residual may be repeated. That is, the sequence of steps from obtaining the X and Y blur curves of the electron beam 1 for V _2A to the input of V _2A into the objective deflector 13 and the X direction of the electron beam 1 for V _2B . , and Y-direction blur curve acquisition to the input of _V2B to the objective deflector 13, a series of steps may be repeated. Meanwhile, _V2A and _V2B are updated by equations (24a) and (25a) respectively. Equations (24a) and (25a) are generalizations of equations (24) and (25), respectively. Henceforth, unless otherwise specified, formulas (24a) and (25a) also serve as formulas (24) and (25), respectively.
V _2A0 ^(m) = M _A0 ^(m-1) (24a)
V _2B0 ^(m) = M _B0 ^(m-1) (25a)
In equations (24a) and (25a), V _2A0 ^(m) and V _2B0 ^(m) represent V _2A0 and V _2B0 updated m (≧1) times, respectively. M _A0 ^(m−1) represents M _A0 obtained after inputting V _2A0 ^(m−1) to the objective deflector 13 . M _B0 ^(m−1) represents M _B0 obtained after inputting V _2B0 ^(m−1) to the objective deflector 13 .

The above series of steps may be repeated until the equations (26) and (27) hold. That is, if │ΔM _A0 ^(m) │ and │ΔM _B0 ^(m) │ are reduced until the equations (26) and (27) hold respectively, the correction of the axial two-fold astigmatism is completed.
|ΔM _A0 ^(m) |<ε _M (26)
|ΔM _B0 ^(m) |<ε _M (27)
In equations (26) and (27), ΔM _A0 ^(m) and ΔM _B0 ^(m) are the amount of change in M _A0 due to the m-th update of V _2A0 and the amount of M A0 due to the m-th update of V _2B0 , respectively. represents the amount of change in _B0 . ε _M (>0) represents the tolerance for |ΔM _A0 ^(m) | and |ΔM _B0 ^(m) |. That is, |ΔM _A0 ^(m) | and |ΔM _B ^(m) | correspond to correction residuals of the axial two-fold astigmatism. ΔM _A0 ^(m) and ΔM _B0 ^(m) are defined by equations (28) and (29), respectively.
ΔM _A0 ^(m) = M _A0 ^(m) - M _A0 ^(m-1) (28)
ΔM _B0 ^(m) = M _B0 ^(m) − M _B0 ^(m−1) (29)

The above two series of steps may in principle be repeated separately. For example, first, a series of steps from acquisition of curves of blurring in the X direction and Y direction of the electron beam 1 with respect to V _2A to input of V _2A to the objective deflector 13 are performed until equation (26) holds. Next, a series of steps from acquisition of the X- and Y-direction blurring curves of the electron beam 1 with respect to _V2B to input of _V2B to the objective deflector 13 are performed by formula (27). You can update until
In practice, however, if the above two series of steps are repeated in this way, there is a possibility that the correction residual of the above-mentioned on-axis two-fold astigmatism will become a problem. That is, there is a possibility that equations (26) and (27) will not hold at the same time.
The cause is the orthogonality error between the two-fold astigmatism _V2A occurs at the height of the knife edge 20 and the two-fold astigmatism _V2B occurs at the same height. This orthogonal error generates a component of the 2-fold deflection astigmatism that can be corrected by _V2B when the component of the ₂ -fold deflection astigmatism that can be corrected by V2A is corrected. When the component of the two-fold deflection astigmatism correctable by _V2B is corrected, the component of the two-fold deflection astigmatism correctable by _V2A is generated.
If the above becomes a problem, the above two series of steps may be alternately repeated. Then, even if the orthogonal error is not zero, the component of the two-fold deflection astigmatism that can be corrected by _V2A and the component of the two-fold deflection astigmatism that can be corrected by _V2B are obtained. becomes smaller alternately and gradually, so that equations (26) and (27) hold simultaneously with each other.

The apparatus of this embodiment then measures and corrects its own deflection curvature of field aberration. The requirements are as follows.

Simply put, measuring and correcting the deflected curvature of field is the same as the axial defocus is measured and corrected at each grid point in the deflection field of the objective deflector 13.
Here, the lattice points in the deflection field of the objective deflector 13 are lattice points obtained by equally or unequally dividing the deflection field in the X and Y directions. If the number of divisions of the deflection field is 4, for example, the number of grid points thus obtained is 25 (=5.times.5) points in total.

In the apparatus of this embodiment, in order to measure and correct the deflection curvature of field aberration, first, the knife edge 20 is moved to the coordinates of the grid point where the measurement and correction are performed, and an electron beam is placed on the grid point. The electron beam 1 is deflected by the objective deflector 13 so that the beam 1 is incident.
The apparatus of this embodiment then blurs the electron beam 1 in the X direction and the Y direction with the knife edge 20 while increasing or decreasing _V2A on the grid point (the grid point where the measurement and correction are performed first). By measuring, curves (two in total) of blurring in the X and Y directions of the electron beam 1 with respect to V _2A are obtained, and from these curves the XY difference D _A0 is determined.
Alternatively, the apparatus of this embodiment measures the X-direction and Y-direction blurring of the electron beam 1 with the knife edge 20 while increasing or decreasing _V2B at the lattice points, thereby measuring the X-direction blurring of the electron beam 1 with respect to _V2B . and Y-direction blur curves (two in total) are obtained, and the XY difference D _B0 is determined from these curves.
When acquiring the above curves, it is necessary to provide a range in which the astigmatism correction voltage is increased and decreased twice to have a width sufficient to confirm the troughs of each curve. For this purpose, it is necessary to roughly grasp the distribution of the deflection field curvature aberration in the deflection field of the objective deflector 13 in advance. The distribution can be predicted from electro-optical calculations for the optical system of FIG. 1, or from the experience of using the same optical system.

Next, the apparatus of this embodiment updates the focus correction voltage _V0A and inputs it to the objective deflector 13 so as to make the XY difference D _A0 or D _B0 zero on the grid point. As a result, the deflection curvature of field aberration on the lattice points is corrected. Here, D _A0 or D _B0 is, more specifically, one of D _A0 and D _B0 that is not zero, and more preferably one of D _A0 and D _B0 that is not smaller in absolute value. be. (It is not one of D _A0 and D _B0 but both D _A0 and D _B0 that becomes zero when the deflection field curvature aberration on the lattice point is corrected. This will be described later. )
The updating of the focus correction voltage V _0A0 (that is, the determination of V _0A0 ′) is according to equation (30) or (31). Formula (30) replaces I _0A0 , I _0A0 ', D _A0 , d _IA , and I _D0 in formula (9) with V _0A0 (x, y), V _0A0 '(x, y), D _A0 (x, y), d _VA , and V _D0 (x, y). Formula (31) replaces I _0A0 , I _0A0 ', D _B0 , d _IB , and I _D0 in formula (10) with V _0A0 (x, y), V _0A0 '(x, y), and D _B0 , respectively. (x, y), d _VB , and V _D0 (x, y). Equations (30) and (31) are used when equations (32) and (33) hold, respectively.
V _0A0 ′(x, y)=V _0A0 (x, y)−D _A0 (x, y)/d _VA
= V _0A0 (x, y) - V _D0 (x, y) (30)
V _0A0 ′(x, y)=V _0A0 (x, y)−D _B0 (x, y)/d _VB
= V _0A0 (x, y) - V _D0 (x, y) (31)
|d _VA |≧|d _VB | (32)
|d _VA |<|d _VB | (33)
In equations (30) and (31), V _0A0 (x, y) represents V _0A0 on the lattice point located at the deflection coordinates (x, y). V _0A0 ′(x,y) represents V _0A0 (x,y) that is updated to zero D _A0 (x,y) or D _B0 (x,y). V _0A0 represents the 0th focus correction voltage. The 0th focus correction voltage is a focus correction voltage under the 0th focus correction condition. D _A0 (x, y) and D _B0 (x, y) respectively represent D _A0 and D _B0 on the grid point located at the deflection coordinates (x, y). d _VA and d _VB represent the partial derivatives of D _A and D _B , respectively, with respect to V _0A .

V _D0 (x, y) in equations (30) and (31) are given by equations (34) and (35), respectively. V _D0 (x, y) in equations (30) and (34) represents the XY difference D _A0 (x, y) converted into the amount of change in focus correction voltage. V _D0 (x, y) in equations (31) and (35) represents the XY difference D _B0 (x, y) converted into the amount of change in the focus correction voltage. Equations (34) and (35) indicate that the apparatus of this embodiment converts the XY difference into the amount of change in the focus correction voltage. The conversion is by the factor _dVA or _dVB .
V _D0 (x, y)=D _A0 (x, y)/d _VA (34)
V _D0 (x, y)=D _B0 (x, y)/d _VB (35)
Here, V _D0 (x, y) also represents the deflection curvature of field aberration converted into the amount of change in the focus correction voltage. This is because the XY differences D _A0 (x, y) and D _B0 (x, y) both represent the deflection field curvature aberration.
Therefore, equations (30) and (31) are obtained when the apparatus of this embodiment subtracts the deflection curvature of field converted into the amount of change in the focus correction voltage from V _0A0 (x, y), so that (ie, determine V _0A0 ′(x,y)) by V _0A0 (x,y).

If V _D0 (x, y) is the measurement value of the deflection curvature of field aberration, the correction value of the deflection curvature of field aberration is −V _D0 (x, y) (=V _0A0 ′(x, y) −V _0A0 (x,y)=−D _A0 (x,y)/d _VA =−D _B0 (x,y)/d _VB ). That is, the measured value and the corrected value of the deflected curvature of field aberration have the same magnitude and opposite signs. Hereinafter, for convenience of explanation, V _D0 (x, y) is used as the measurement value of the deflection field curvature aberration.

Next, the apparatus of this embodiment, on the lattice points, acquires curves of blurring in the X and Y directions of the electron beam 1 with respect to V _2A , and transfers V _0A0 (x, y) to the objective deflector 13. A sequence of steps to input or from acquisition of X and Y blur curves of the electron beam 1 to V _2B to input to the objective deflector 13 of V _0A0 (x,y) repeat the process. This reduces the correction residual of the deflection field curvature aberration on the lattice points.
While the above series of steps are repeated, V _0A0 (x, y) is updated according to equation (30a) or (31a). Equations (30a) and (31a) are generalizations of equations (30) and (31), respectively. Henceforth, unless otherwise specified, formulas (30a) and (31a) also serve as formulas (30) and (31), respectively.
V _0A0 ^(m) (x, y)=V _0A0 ^(m-1) (x, y)-D _A0 ^(m-1) (x, y)/d _VA
= V _0A0 ^(m-1) (x, y) - V _D0 ^(m-1) (x, y) (30a)
V _0A0 ^(m) (x, y)=V _0A0 ^(m-1) (x, y)-D _B0 ^(m-1) (x, y)/d _VB
=V _0A0 ^(m-1) (x, y)-V _D0 ^(m-1) (x, y) (31a)
In equations (30a) and (31a), V _0A0 ^(m) (x, y) represents V _0A0 (x, y) updated for m (≧1) times. D _A0 ^(m−1) (x, y) and D _B0 ^(m−1) (x, y) are obtained by inputting V _0A0 ^(m−1) (x, y) to the objective deflector 13 Denote D _A0 (x, y) and D _B0 (x, y), respectively. V _D0 ^(m−1) (x, y) represents V _D0 (x, y) obtained by inputting V _0A0 ^(m−1) (x, y) to the objective deflector 13 . V _D0 ^(m−1) (x, y) in equations (30a) and (31a) are given by equations (34a) and (35a), respectively.
V _D0 ^(m−1) (x, y)=D _A0 ^(m−1) (x, y)/d _VA (34a)
V _D0 ^(m−1) (x, y)=D _B0 ^(m−1) (x, y)/d _VB (35a)
The series of steps described above is repeated until equation (36) or (37) is established on the lattice points. In equations (36) and (37), ε _D represents the tolerance for |D _A0 ^(m) (x,y)| and |D _B0 ^(m) (x,y)|. |D _A0 ^(m) (x, y)| and |D _B0 ^(m) (x, y)| .
|D _A0 ^(m) (x, y)|<ε _D (36)
|D _B0 ^(m) (x, y)|<ε _D (37)

Determination of coefficients _dVA and _dVB is according to equations (38) and (39), respectively.

（３８）および（３９）式において、Ｄ_Ａ１（ｘ，ｙ）、Ｄ_Ｂ１（ｘ，ｙ）、およびＶ_０Ａ１（ｘ，ｙ）は、偏向座標（ｘ，ｙ）に位置する格子点上におけるＤ_Ａ１、Ｄ_Ｂ１、およびＶ_０Ａ１を、それぞれ表す。Ｖ_０Ａ１は、第１のフォーカス補正電圧を表す。第１のフォーカス補正電圧とは、第１のフォーカス補正条件下におけるフォーカス補正電圧である。Ｄ_Ａｎ（０，０）、Ｄ_Ｂｎ（０，０）、およびＶ_０Ａｎ（０，０）（いずれにおいてもｎ＝０または１）は、対物偏向器１３の偏向フィールドの中央におけるＤ_Ａｎ（ｘ，ｙ）、Ｄ_Ｂｎ（ｘ，ｙ）、およびＶ_０Ａｎ（ｘ，ｙ）をそれぞれ表す。補足すれば、（３８）式におけるＶ_０Ａ１（０，０）、Ｖ_０Ａ０（０，０）、およびΔＶ_０Ａは、それぞれ、（３９）式におけるＶ_０Ａ１（０，０）、Ｖ_０Ａ０（０，０）、およびΔＶ_０Ａと同じであってもよいし、異なっていてもよい。
本実施例の装置は、係数ｄ_ＶＡを、ｄ_ＶＡによるＤ_Ａ０（ｘ，ｙ）のＶ_Ｄ０（ｘ，ｙ）への換算（（３４）および（３４ａ）式を参照）に先立って決定し、記憶する。そうして記憶されたｄ_ＶＡは、ｄ_ＶＡによるＤ_Ａ０（ｘ，ｙ）のＶ_Ｄ０（ｘ，ｙ）への換算のたびに、繰り返し用いられる。本実施例の装置は、あるいは、係数ｄ_ＶＢを、ｄ_ＶＢによるＤ_Ｂ０（ｘ，ｙ）のＶ_Ｄ０（ｘ，ｙ）への換算（（３５）および（３５ａ）式を参照）に先立って決定し、記憶する。そうして記憶されたｄ_ＶＢは、ｄ_ＶＢによるＤ_Ｂ０（ｘ，ｙ）のＶ_Ｄ０（ｘ，ｙ）への換算のたびに、繰り返し用いられる。ここで、ｄ_ＶＡおよびｄ_ＶＢは、それぞれ、（３２）および（３３）式が成立する場合に記憶される。

In equations (38) and (39), D _A1 (x, y), D _B1 (x, y), and V _0A1 (x, y) are given by Denote D _A1 , D _B1 and V _0A1 respectively. V _0A1 represents the first focus correction voltage. The first focus correction voltage is a focus correction voltage under the first focus correction condition. D _An (0,0), D _Bn (0,0), and V _0An (0,0) (where n=0 or 1) are D _An (x , y), D _Bn (x, y), and V _0An (x, y) respectively. Supplementally, V _0A1 (0, 0), V _0A0 (0, 0), and ΔV _0A in equation (38) are V _0A1 (0, 0) and V _0A0 (0, 0, respectively) in equation (39). 0), and ΔV _0A may be the same or different.
The apparatus of this embodiment determines the coefficient d _VA prior to the conversion of D _A0 (x,y) to V _D0 (x,y) by d _VA (see equations (34) and (34a)). ,Remember. The d _VA so stored is used repeatedly each time the d _VA converts D _A0 (x,y) to V _D0 (x,y). Alternatively, the apparatus of this embodiment may _convert the _{coefficient d VB} to Decide and memorize. The d _VB thus stored is used repeatedly for each conversion of D _B0 (x,y) to V _D0 (x,y) by d _VB . Here, d _VA and d _VB are stored when equations (32) and (33) respectively hold.

Supplementally, the coefficients _dVA and _dVB may be recalculated each time the electron beam 1 is made incident on a new grid point, but this is not necessary. That is, the coefficients _{d_VA} and _{d_VB} measured at the center of the deflection field are also valid for the other deflection coordinates. This is because the focus correction sensitivity and the two-fold astigmatism correction sensitivity of the objective deflector 13 do not depend on the deflection coordinates.
Strictly speaking, the dependence of the coefficients _dVA and _dVB on the deflection coordinate is non-zero, so due to that dependence the coefficients _dVA and _dVB may have errors. However, the error does not matter. V _0A0 (x,y) objective deflector 13 updated by equation (30a) from acquisition of the X and Y blur curves of the electron beam 1 versus V _2A , even if the error is not zero or from the acquisition of the X and Y blur curves of the electron beam 1 to V _2B , V _0A0 (x,y) updated by equation (31a) to the objective If the series of steps up to the input to the deflector 13 is repeated on the grid points, D _A0 (x, y) and D _B0 (x, y) approach zero, so on the grid points, Defocus on the knife edge 20 approaches zero.

The apparatus of this embodiment thereafter performs the same measurement and correction as above for the remaining lattice points within the deflection field of the objective deflector 13 . Thereby, the focus correction voltage V _0A0 ^(m) (x, y) on all lattice points within the deflection field of the objective deflector 13 is determined.

The apparatus of this embodiment then determines an approximate surface for V _0A0 ^(m) (x,y). When the apparatus of this embodiment deflects the electron beam 1 toward the deflection coordinates (x, y), the focus correction voltage V _0A corresponding to the deflection coordinates (x, y) is determined from the approximate curved surface. and input the V _0A to the objective deflector 13 . Thereby, the deflection field curvature aberration is corrected over the entire deflection field of the objective deflector 13 .

The approximate curved surface is a function of two or more degrees regarding deflection coordinates (two dimensions). Henceforth, the said approximate curved surface is represented by _V0A0F (x,y).

The approximation accuracy of V _0A0F (x,y) increases with the order of V _0A0F (x,y). However, in order to make V _0A0F (x, y) a higher-order function, it is necessary to increase the number of pairs of deflection coordinates (x, y) and V _0A0 (x, y) accordingly. . That is, it becomes necessary to increase the number of grid points in the deflection field of the objective deflector 13 . More specifically, the number of grid points in the deflection field of the objective deflector 13 should be greater than or equal to the number of coefficients of the polynomial representing V _0A0F (x, y). For example, if V _0A0F (x,y) is a cubic function of X and Y, the number of coefficients of the polynomial expressing V _0A0F (x,y) is ten.

According to the above procedure, the deflection field curvature aberration is measured and corrected based on the blur curve of the electron beam 1 with respect to the two-fold astigmatism correction voltage, not the curve of the blur of the electron beam 1 with respect to the focus correction voltage. can do.
If the deflection curvature of field aberration is measured and corrected based on the blur curve of the electron beam 1 with respect to the two-fold astigmatism correction voltage, the time required for the measurement and correction of the deflection curvature of field aberration can be reduced. It is possible to increase the measurable range of the deflection field curvature aberration. This is due to the speed of response of the objective deflector 13 as an electrostatic two-fold stigmator to the two-fold astigmatism correction voltage, and the speed of the response of the objective deflector 13 as an electrostatic two-fold stigmator. This is because of the size of the range in which the two-fold astigmatism on the knife edge 20 can be changed according to the astigmatism correction voltage.

The size of the range in which the objective deflector 13 as an electrostatic two-fold astigmatism corrector can change the two-fold astigmatism on the knife edge 20 in accordance with the two-fold astigmatism correction voltage includes the electrostatic two-fold astigmatism corrector. The two-fold astigmatism correction sensitivity of the objective deflector 13 as a rotational astigmatism corrector is involved. That is, the two-fold astigmatism correction sensitivity of the objective deflector 13 as an electrostatic two-fold astigmatism corrector is higher than the focus correction sensitivity of the objective deflector 13 as an electrostatic focus corrector.
This involves the dependence of the focus correction sensitivity and the two-fold astigmatism correction sensitivity of the objective deflector 13 on the inner diameter of the objective deflector 13 . More specifically, the focus correction sensitivity of the objective deflector 13 as an electrostatic focus corrector does not depend on the inner diameter of the objective deflector 13. The two-fold astigmatism correction sensitivity is inversely proportional to the square of the inner diameter of the objective deflector 13 . (The dependence of these correction sensitivities on the inner diameter of the objective deflector 13 is derived from the dependence of the potential distribution in the objective deflector 13 on the inner diameter of the objective deflector 13.) Electrostatic Double Astigmatism The two-fold astigmatism correction sensitivity of the objective deflector 13 as a corrector is therefore enhanced by reducing the inner diameter of the objective deflector 13 .

The above is the description of the measurement and correction of the deflection curvature of field produced by the apparatus of this embodiment. The measurement and correction routine is shown in FIG. 10 as main routine 2 .
However, both the main routine 2 and the subroutines therein (subroutine 3 and subroutine 4) are based on the premise that equation (32) holds. That is, in these main routines and _subroutines , of the coefficients _dVA and _dVB , only _dVA is _{determined .} is determined.

(Main routine 2)
As shown in FIG. 10, the main routine 2 consists of the following steps.
First, in step S31, n is set to zero (n=0).
Next, in step S 32 , V _0An (V _0A0 ) is read and input to the objective deflector 13 . (Here, V _0A0 is equal to the value of the focus correction voltage V _0A at the center of the deflection field of the objective deflector 13 immediately before the start of main routine 2.)
Next, in step S33, subroutine 1 (see FIG. 7) is executed to determine the XY difference D _An (D _A0 ). (Here, D _An is measured on grid points located in the center of the deflection field of the objective deflector 13. This also applies to step S36.)
Next, in step S34, n is set to 1 (n=1).
Next, in step S 35 , V _0An (V _0A1 ) is read and input to the objective deflector 13 .
Next, in step S36, subroutine 1 (see FIG. 7) is executed to determine the XY difference D _An (D _A1 ).
Next, in step S37, the coefficient d _VA is determined. (If the coefficient d _VA is known, steps S31-S37 can be replaced by reading d _VA .)
Next, in step S38, subroutine 3 (see FIG. 11) is executed to determine V _0A0 ^(m) (x, y).
Next, in step S39, an approximate curved surface V _0A0F (x, y) is determined. Then, the main routine 2 ends.

(Subroutine 3)
Subroutine 3, as shown in FIG. 11, consists of the following steps. However, in the following description n _1A and n _1B represent the loop control variables for V _1A and V _1B respectively. n _1Amin (<0) and n _1Amax (>0) represent the minimum and maximum values of n _1A respectively (|n _1Amin |=|n _1Amax |). n _1Bmin (<0) and n _1Bmax (>0) represent the minimum and maximum values of n _1B , respectively (|n _1Bmin |=|n _1Bmax |). ΔV _1A and ΔV _1B represent the amount of change in V _1A and V _1B , respectively.
First, in step S41, n _1A is set to n _1Amin (n _1A =n _1Amin ) and n _1B is set to n _1Bmin (n _1B =n _1Bmin ).
Next, in step S42, V _1B (=ΔV _1B n _1B ) is input to the objective deflector 13 . where V _1B is proportional to the y component of deflection coordinates (x,y).
Next, in step S43, V _1A (=ΔV _1A n _1A ) is input to the objective deflector 13 . where _V1A is proportional to the x component of deflection coordinates (x,y).
Next, in step S44, subroutine 4 (see FIG. 12) is executed to determine V _0A0 ^(m) .
Next, in step S45, it is determined whether or not n _1A <n _1Amax holds. If it is established, the process proceeds to step S46. If it does not hold, the process proceeds to step S47.
In step S46, n _1A +1 is substituted for n _1A (n _1A =n _1A +1). Then, the process returns to step S43.
In step S47, it is determined whether n _1B <n _1Bmax holds. If it is established, the process proceeds to step S48. If it does not hold, the subroutine 3 is terminated and the process returns to the main routine 2 (see FIG. 10).
In step S48, n _1B +1 is substituted for n _1B (n _1B =n _1B +1). Then, the process returns to step S42.
Supplementally, step S42 and step S43 may be performed before or after each other. However, in doing so, steps S45 and S46 and steps S47 and S48 must be performed one behind the other.

(Subroutine 4)
Subroutine 4, as shown in FIG. 12, is substantially the same as subroutine 2 (see FIG. 8). More specifically, subroutine 4 corresponds to subroutine 2 in which steps S22 and S26 are replaced with steps S52 and S56, respectively. At step S52, V _0A0 ^(m) is input to the objective deflector 13 . In step S56, V _0A0 ^(m) is determined. (V _0A0 ^(m) represents V _0A0 when m=0. Here, V _0A0 is equal to the value of the focus correction voltage V _0A immediately before the start of main routine 2. Its value is the deflection coordinate ( x, y).)
Supplementally, in subroutine 4, the focus correction voltage is determined regardless of the _deflection coordinates ⁽ x, y). V 0A0 (m) instead of V _0A0 ^(m) . Similar considerations apply to subroutine 7 (see FIG. 20).

The apparatus of this embodiment finally measures and corrects its own 2-fold deflection astigmatism. The measurement and correction of the deflection two-fold astigmatism are also performed on grid points (5×5=25 points in total, as described above) in the deflection field of the objective deflector 13 .
The procedure for measuring and correcting the two-fold deflection astigmatism is the same as the procedure for measuring and correcting the same aberration in the optical system of FIG. Simply put, measuring and correcting the deflection 2-fold astigmatism is the same as measuring and correcting the above-described axial 2-fold astigmatism at each of the grid points. The details are as follows.

In order to measure and correct the deflection 2-fold astigmatism, the apparatus of this embodiment first moves the knife edge 20 to the coordinates of the grid point where the measurement and correction are performed, and moves the knife edge 20 onto the grid point. The electron beam 1 is deflected by the objective deflector 13 so that the electron beam 1 is incident.
The apparatus of this embodiment then blurs the electron beam 1 in the X direction and the Y direction with the knife edge 20 while increasing or decreasing _V2A on the grid point (the grid point where the measurement and correction are performed first). By measuring, X-direction and Y-direction blur curves (two curves in total) of the electron beam 1 with respect to V _2A are obtained.
The apparatus of this embodiment determines the XY average M _A0 from these curves, updates V _2A based on M _A0 , and inputs the updated V _2A to the objective deflector 13 . As a result, the component of the two-fold deflection astigmatism on the lattice points that can be corrected by _V2A is corrected.
Next, the apparatus of this embodiment measures the blurring of the electron beam 1 in the X direction and the Y direction with the knife edge 20 while increasing and decreasing _V2B on the lattice points, thereby measuring the electron beam 1 with respect to _V2B . Acquire X and Y blur curves (two in total).
The apparatus of this embodiment determines the XY average M _B0 from these curves, updates V _2B based on M _B0 , and inputs the updated V _2B to the objective deflector 13 . As a result, the component of the two-fold deflection astigmatism on the lattice points that can be corrected by _V2B is corrected.
Supplementally, in the above process, the determination of M _B0 and the input of _V2B are performed after the determination of M _A0 and the input of _V2A , but these orders may be reversed.

The updating of _V2A0 and _V2B0 in the above process can be expressed by equations (40) and (41), respectively.
V _2A0 ′(x, y)=M _A0 (x, y) (40)
V _2B0 ′(x, y)=M _B0 (x, y) (41)
In equations (40) and (41), V _2A0 ′(x, y) and V _2B0 ′(x, y) are V _2A0 (x, y) and V _2B0 (x, y) updated by the above process. , respectively. V _2A0 (x, y) and V _2B0 (x, y) respectively represent V _2A0 and V _2B0 on the grid point located at the deflection coordinates (x, y). M _A0 (x, y) and M _B0 (x, y) respectively represent M _A0 and M _B0 on the grid point located at the deflection coordinates (x, y). M _A (x, y) and M _B (x, y) respectively represent M _A and M _B on the grid point located at the deflection coordinates (x, y).

Next, the apparatus of this embodiment performs a series of steps from acquisition of blur curves in the X direction and Y direction of the electron beam 1 with respect to V _2A to input of V _2A to the objective deflector 13 on the grid points. and the series of steps from acquisition of the X- and Y-direction blur curves of the electron beam 1 to V _2B to input of V _2B to the objective deflector 13 are repeated. Meanwhile, V _2A and V _2B are updated by equations (40a) and (41a) respectively. Equations (40a) and (41a) are generalizations of equations (40) and (41), respectively. Hereinafter, unless otherwise specified, formulas (40a) and (41a) also serve as formulas (40) and (41), respectively.
V _2A0 ^(m) (x, y)=M _A0 ^(m−1) (x, y) (40a)
V _2B0 ^(m) (x, y)=M _B0 ^(m−1) (x, y) (41a)
In equations (40a) and (41a), V _2A0 ^(m) (x, y) and V _2B0 ^(m) (x, y) are V _2A0 (x, y) updated m (≧1) times and V _2B0 (x,y) respectively. M _A0 ^(m−1) (x, y) represents M _A0 (x, y) obtained after inputting V _2A0 ^(m−1) (x, y) to the objective deflector 13 . M _B0 ^(m−1) (x, y) represents M _B0 (x, y) obtained after inputting V _2B0 ^(m−1) (x, y) to the objective deflector 13 .
The above series of steps are repeated until the equations (42) and (43) hold on the lattice points. As a result, the correction of the two-fold deflection astigmatism is completed on the lattice points.
|ΔM _A0 ^(m) (x, y)|<ε _M (42)
|ΔM _B0 ^(m) (x, y)|<ε _M (43)
In equations (42) and (43), ΔM _A0 ^(m) (x, y) and ΔM _B0 ^(m) (x, y) are respectively M _A0 by the m-th update of V _2A0 (x, y) It represents the amount of change in (x, y) and the amount of change in M _B0 (x, y) due to the m-th update of V _2B0 (x, y). ε _M (>0) represents the tolerance for |ΔM _A0 ^(m) (x,y)| and |ΔM _B0 ^(m) (x,y)|. |ΔM _A0 ^(m) (x, y)| and |ΔM _B0 ^(m) (x, y)| do. ΔM _A0 ^(m) (x, y) and ΔM _B0 ^(m) (x, y) are defined by equations (44) and (45), respectively.
ΔM _A0 ^(m) (x, y)=M _A0 ^(m) (x, y)-M _A0 ^(m−1) (x, y) (44)
ΔM _B0 ^(m) (x, y)=M _B0 ^(m) (x, y)−M _B0 ^(m−1) (x, y) (45)

The above two series of steps may in principle be repeated separately. For example, first, a series of steps from acquisition of curves of blurring in the X direction and Y direction of the electron beam 1 with respect to V _2A to input of V _2A to the objective deflector 13 are performed until equation (42) holds. Next, a series of steps from acquisition of the X-direction and Y-direction blur curves of the electron beam 1 with respect to _V2B to input of _V2B to the objective deflector 13 are performed by formula (43). You can update until
However, the above two series of steps are preferably repeated alternately in practice. Then, depending on the orthogonal error between the two-fold astigmatism _V2A generated at the height of the knife edge 20 and the two-fold astigmatism _V2B generated at the same height, (42) and (43) can be established simultaneously with each other.

The apparatus of this embodiment thereafter performs the same measurement and correction as above for the remaining lattice points within the deflection field of the objective deflector 13 . As a result, two-fold astigmatic correction voltages V _2A0 ^(m) (x, y) and V _2B0 ^(m) (x, y) on all lattice points within the deflection field of the objective deflector 13 are determined.

The apparatus of this embodiment then determines an approximate surface for each of _V2A0 ^(m) (x,y) and _V2B0 ^(m) (x,y). After that, when the apparatus of this embodiment _deflects the electron beam 1 toward the deflection _coordinates ⁽ x, y ⁾ , Two-time astigmatism correction voltages V ₂ A and V 2 _B corresponding to the deflection coordinates (x, y) are determined from the approximate curved surface, and both V ₂ A and V 2 _B are input to the objective deflector 13 . As a result, the two-fold deflection astigmatism is corrected over the entire deflection field of the objective deflector 13 .

Hereinafter, the origin of the XY difference will be explained, and the grounds for the fact that the XY difference serves as an index of the degree of defocus on the knife edge 20 will be shown. To that end, first, aberration diagrams representing defocus and 2-fold astigmatism on the knife edge 20 are shown. Then, the XY difference is derived from the aberration diagrams representing these aberrations.

First, assuming that the aberration figure representing the defocus on the knife edge 20 is a _defocus figure _SF (complex number), SF can be expressed by equation (46) using the convergence half angle of the electron beam 1 .
S _F (α, θ)=Fαexp(iθ) (46)
(46), F (real number) is a coefficient relating to the defocus. F corresponds to the deviation of the height position of the projection figure 11 from the height position of the knife edge 20 . α represents the convergence half angle of the electron beam 1 . Here, the convergence half angle of the electron beam 1 means the convergence half angle of rays around each principal ray included in the electron beam 1 . θ represents the rotation angle around each principal ray included in the electron beam 1 . Both α and θ are the angles for the paraxial trajectories of the rays around each of the chief rays. Here, a paraxial trajectory is a trajectory that does not include aberrations.

The coefficient F can be expressed by equation (47).
F _{= FO+a0V0A + b0I0A (} 47)
(47), F ₀ is a constant and represents F under the condition that both V _0A and I _0A are zero. _{a0 and b0 represent coefficients related to defocus generated by the focus correction voltage V0A and the focus correction current I0A , respectively} . Hereinafter, a ₀ and b ₀ are referred to as defocus coefficients.

A defocus figure S _F (α, θ) is defined for all θ in the range 0≦θ<2π and for all α in the range 0≦α≦α _max . Alternatively, it may be considered that an aggregate of them is a defocused figure. where α _max represents the maximum value of α. α _max is determined by the diameter of the electron beam 1 . FIG. 13 shows an example of S _F (α, θ) obtained by appropriately changing θ and α within the ranges of 0≦θ<2π and 0≦α≦ _αmax , respectively.
As can be seen from FIG. 13, the defocused figure S _F (α, θ) forms a circle, and a group of a plurality of S _F with different α has the same center. That is, the group of _SFs form concentric circles.
Each circle forming the concentric circles is drawn counterclockwise once while θ varies from 0 to 2π. This is because the right side of equation (46) makes one counterclockwise rotation while θ changes from 0 to 2π.

The size of the defocus figure S _F (α, θ) is proportional to the coefficient F, as can be seen from the equation (46). Therefore, if the absolute value of _F becomes smaller (or larger), the defocused figure _SF becomes smaller (or larger). become zero.

On the other hand, if the aberration figure representing the 2-fold astigmatism on the knife edge 20 is a 2-fold astigmatism figure S ₂ (complex number), S ₂ can be expressed by equation (48).
S ₂ (α, θ)=T _C αexp(−iθ) (48)
(48), T _C (complex number) is a coefficient relating to the above two-fold astigmatism. T _C has a magnitude equal to half the astigmatic difference (distance between two focal lines) exhibited by the two-fold astigmatism. T _C also has a deflection angle that varies with the direction of the focal line formed by the two-fold astigmatism.

The coefficient T _C can be expressed by the equation (49).
_TC = _TOC + a _2C V _2C (49)
(49), T _OC (complex) is a constant and represents T _C under the condition that V _2C (complex) is zero. a _2C (complex) is the coefficient of V _2C and is a constant. V _2C represents a two-fold astigmatism correction voltage (two-fold astigmatism generation voltage).

_a2C and _V2C are defined by equations (50) and (51), respectively.
a _2C = a _2R + ia _2I (50)
_V2C = _V2A + _iV2B (51)
(50), a _2R and a _2I (both real numbers) represent the real and imaginary parts of a _2C , respectively. a _2C becomes a real number (a _2C =a _2R , that is, a _2I =0) depending on the rotation angle of the objective deflector 13 around the Z-axis. Hereinafter, a _2C , a _2R and a _2I are referred to as 2-fold astigmatism coefficients.

A 2-fold astigmatism figure S ₂ (α, θ) is defined for all θ in the range 0≦θ<2π and for all α in the range 0≦α≦α _max . Alternatively, they may be considered as a two-fold astigmatism figure. FIG. 14 shows an example of S ₂ (α, θ) obtained by appropriately changing θ and α within the ranges of 0≦θ<2π and 0≦α≦ _αmax , respectively.
As can be seen from FIG. 14, S ₂ (α, θ) forms a circle like _SF (α, θ), and a plurality of groups of S ₂ with different α have the same center. That is, the _S2 groups form concentric circles.
Each circle forming the concentric circles is drawn clockwise once while θ varies from 0 to 2π. This is because the right side of equation (48) makes one clockwise rotation while θ changes from 0 to 2π. S ₂ (α, _θ ) differs from SF (α, θ) in this respect.

As can be seen from the equations (48) and (49), the two-fold astigmatism figure S ₂ (α, θ) has its magnitude and rotation angle determined by the two-fold astigmatism correction voltage V _2C (=V _2A +iV _2B ). Therefore, the magnitude and rotation angle of _S2 can be controlled by _V2C . (However, since S2 forms a circle, its rotation angle does not appear in _FIG . 14.)

Equation (52) is obtained from equations (46) and (48). Equation (52) represents an aberration figure obtained by synthesizing the _defocus figure SF and the _two -fold astigmatism figure S2. _Henceforth , the aberration figure is called aberration figure SF ₊ S2.
S _F +S ₂ =Fαexp(iθ)+ _TC αexp(−iθ) (52)
The right-hand side of equation (52) forms an ellipse (or a circle or line segment in some cases) as θ varies from 0 to 2π. Increase or decrease. The ellipse has 2-fold symmetry. This is why the aberration represented by S2 is called ₂ -fold astigmatism.

Even if the coefficient F, the convergence half angle α, and the ₂ - _{fold astigmatism} coefficient a _2C are constant, the ellipse has its own diameter (major axis and minor diameter). This is explained below.
However, for the sake of explanation, a two-fold astigmatism correction voltage different from V _2C is defined, and it is defined as V _2SC (complex number). V _{2SC aligns} its zero value with V _2C satisfying T _C (=T _OC +a _2C V _2C )=0.

Using V _2SC , equation (49) can be rewritten as equation (49a). That is, when V _2SC is zero, T _C is zero and therefore the 2-fold astigmatism on knife edge 20 is zero.
T _C =a _2C V _2SC (49a)
Equation (53) holds from equations (49) and (49a). Equation (53) represents the relationship between _V2SC and _V2C . V _2SC can be expressed by the formula (54). (54), V _2SA and V _2SB represent the real and imaginary parts of V _2SC , respectively. _V2SA and _V2A have a relationship represented by the formula (55). _V2SB and _V2B have a relationship represented by the formula (56).
_V2SC = _{V2C+TOC/a2C ( 53} )
_V2SC = _V2SA + _iV2SB (54)
_V2SA = _V2A + Re( _TOC / _a2C ) (55)
_V2SB = _{V2B + Im(TOC/a2C ) (} 56)

S _F +S ₂ represented by the equation (52) can be expressed by the equation (52a) when V _2SB =0 holds, and by the equation (52b) when V _2SA =0 holds.
S _F +S ₂ ={|a _2C |V _2SA αexp(−i2(θ−ζ/2))+Fα}exp(i(θ−ζ/2))exp(iζ/2) (52a)
S _F + S ₂ = {|a _2C |V _2SB αexp(-i2(θ-ζ/2-π/4))+Fα}exp(i(θ-ζ/2-π/4))exp(i(ζ /2+π/4)) (52b)
In equations (52a) and (52b), ζ represents the argument of _a2C . ζ does not depend on either V _2SA or V _2SB , but depends on the rotation angle of the objective deflector 13 and on the rotation angle of the trajectory of the principal ray in the electron beam 1 due to the magnetic field of the objective lens 9. .

Both the ellipses represented by the right-hand side of equations (52a) and (52b) change their diameters (major and minor axes) as V _2SA and V _2SB change, but their axes (major and minor axes) Do not change the rotation angle. First, the rotation angle of the ellipse represented by the right-hand side of equations (52a) and (52b) is determined by the action of exp(iζ/2) and exp(i(ζ/2+π/4)), respectively. , and secondly, because ζ does not depend on either V _2SA or V _2SB , as described above.

Specifically, the rotation angles of the major axis (or minor axis) of the ellipse represented by the right side of equations (52a) and (52b) are .zeta./2 and .zeta./2+.pi./4, respectively, due to the above action. That is, the aberration figure SF+S2 created by _V2SA when _V2SB = ₀ is established and the aberration figure _SF + _S2 created by _V2SB when _V2SA = ₀ are established are between them. presents a difference of .pi./4 (45.degree.) as the difference in rotation angle of the major axis (or minor axis) of .
The rotation angle difference π/4 is equal to the rotation angle difference of the distribution functions f _2A (ψ) and f _2B (ψ) (see FIG. 2). The difference in these rotation angles corresponds to the difference between the angular position at which f _2A (ψ) is maximum or minimum and the angular position at which f _2B (ψ) is maximum or minimum. The fact that the difference between these angular positions is π/4 can be confirmed from the fact that f _2A (ψ−π/4) matches f _2B (ψ). That is, the electrode voltage V ₂ (ψ) represented by the equation (3) is reflected in the aberration diagram _SF +S ₂ represented by the equations (52a) and (52b). where f _2A (ψ−π/4) and f _2B (ψ) are equal to cos2(ψ−π/4) and sin2ψ, respectively, at each electrode position.
Supplementally, the fact that the angular variables in the right-hand side of equations (52a) and (52b) are θ-ζ/2, θ-ζ/2-π/4, or the like does not distort or rotate the ellipse. This is because the position of each point on the ellipse is determined for each value (from 0 to 2π) indicated by these angle variables regardless of the angle variables. From another point of view, if an additional value (eg, ζ/2 or ζ/2-π/4) is added to θ, the additional value will be the ellipse itself without changing or rotating the ellipse. Each point on the ellipse is moved on the circumference of the ellipse by an amount corresponding to its size.

_FIG . 15A shows an example of the aberration diagram SF+S2 _{created by V 2SA when} SF ≠ 0 and V _2SB = 0, and V _2SB when _{SF ≠ 0 and V 2SA =} 0 _. An example of the aberration diagram S _F +S ₂ produced by is shown in FIG. 15B. 15A and 15B _show how the aberration diagram _{SF+S2 changes its diameter (major and minor) depending on V2SA and V2SB} , respectively. However, the aberration figures _SF + S ₂ shown in FIGS. 15A and 15B are not multiple aberration figures _SF +S ₂ with different α, but the aberration figures _SF +S ₂ with respect to a single α. is.

It should be noted here that if the diameter of the aberration figure S _F +S ₂ changes, the widths of the aberration figure S _F +S ₂ in the X and Y directions also change, as can be seen from FIGS. 15A and 15B. That is, both the X-direction and Y-direction widths of the aberration figure S _F +S ₂ depend on V _2SA and V _2SB . This means that both the X-direction and Y-direction blurring of the electron beam 1 change depending on V _2SA and V _2SB .

This is shown in FIGS. 4A and 4B. FIG. 4A shows how the X _- direction and Y-direction widths of the aberration diagram SF+S2 created by V _2SA when V _2SB =0 is established _depend on V _2SA . (FIG. 4A is a diagram showing an example of curves of the X and Y blurring of the electron beam 1 with respect to V _2A , as described above. In FIG. 4A, when V _2SB =0 holds, the electron beam 1 changes depending on V _2SA .) FIG. 4B shows the aberration diagram S _F +S ₂ created by V _2SB when V _2SA =0 holds. It shows that the X- and Y-direction widths of are dependent on _V2SB . (FIG. 4B is a diagram showing an example of curves of the X and Y blurring of the electron _beam 1 with respect to V _2B , as described above. In FIG. 4B, the electron beam 1 shows how the blurring in the X and Y directions changes depending on V _2SB .)
Here, V _2SA and V _2SB correspond to V _2A and V _2B in FIGS. 4A and 4B, respectively. When the two-fold astigmatism on the knife edge 20 is zero when both V _2A and V _2B are zero, V _2SA and V _2SB are equal to V _2A and V _2B , respectively.

It should be further noted here that if the difference in rotation angle is π/4, it is possible that one of the XY differences D _A and _{D B} becomes zero when SF ≠0 holds. However, it is impossible that both the XY differences D _A and D _B are zero when S _F ≠0. That is, an undesirable situation in which the defocus on the knife edge 20 is erroneously determined to be zero when it is not zero (S _F ≠0) cannot occur.

This can be explained as follows by paying attention to the rotation angle of the aberration diagram S _F +S ₂ .
When the rotation angle difference is π/4, one of the XY differences D _A and D _B may become zero when S _F ≠0. As shown in FIG. 16B, the ellipse formed by the aberration diagram _{SF +} S2 created by _V2SA when _SF ≠0 and _V2SB =0 holds, and the major axis (or minor axis) parallel to the X-axis or Y-axis ) (see FIG. 16A), but the ellipse formed by the aberration diagram _SF + S ₂ created by V _2SB when _{SF ≠ 0 and V 2SA =} 0 holds is π/ In the case of having the major axis (or minor axis) rotated by 4 (45°) (see Figure 16B).
_In the above case, if _{V 2SA} increases or decreases under the condition that _SF ≠ 0 and V _2SB = 0, as can be seen from FIG. minor axis) varies along the X and Y directions. As a result, an XY difference occurs. That is, D _A ≠0 is established.
On the other hand, in the above case, if V _2SB increases or decreases under the condition that _{SF ≠ 0 and V 2SA = 0 ,} as can be seen from FIG. major and minor axes) along directions rotated by π/4 (45°) with respect to the X and Y axes. As a result, no XY difference occurs. That is, D _B =0 holds. This is independent of both the shape and size of the ellipse.
Whether one of D _A and D _B ( _{D B} in the above case) becomes zero depends on the direction of the axis of the ellipse formed by the aberration diagram SF + S ₂ (two directions perpendicular to each other). not only depends, but also on the directions (two mutually orthogonal directions) in which the width of the aberration figure S _F +S ₂ is evaluated. Here, the direction in which the width of the aberration diagram S _F +S ₂ is evaluated corresponds to the direction in which the blurring of the electron beam 1 is measured. These directions are the directions in which the knife edges 20X and 20Y are scanned by the electron beam 1. FIG. (If the direction in which knife edges 20X and 20Y are scanned is changed, the direction of the edges of knife edges 20X and 20Y must be changed accordingly.)

If both D _A and D _B can be zero when S _F ≠0, such a situation can occur because, for example, if the difference in rotation angle is π/2 instead of π/4, and each of the major axis and the minor axis of the aberration figure _SF + S ₂ is the direction in which the width of the aberration figure _SF + S ₂ is evaluated (two directions orthogonal to each other) is rotated by π/4 with respect to .
However, since the difference in rotation angle is π/4 in reality, such a case cannot occur in reality. That is, in reality, an undesirable situation in which the defocus on the knife edge 20 is erroneously determined to be zero when it is not zero cannot occur. In other words, determining the non-zero one of D _A and D _B in this example is a sufficient condition to prevent that undesirable event from occurring.

Supplementally, it is not a necessary condition for preventing the occurrence of the above situation that the directions in which the width of the aberration figure S _F +S ₂ is evaluated are the X direction and the Y direction. The direction in which the width of the aberration figure S _F +S ₂ is evaluated may, for example, be rotated by π/4 with respect to the X and Y directions.
That is, the direction in which the blur of the electron beam 1 is measured may be rotated by π/4 with respect to the X and Y directions. (However, if the direction in which the blurring of the electron beam 1 is measured is so rotated, the edge directions of the knife edges 20X and 20Y are rotated by π/4 with respect to the Y and X directions, respectively. There is a need.)
In other words, it is possible to arbitrarily define the directions in which the blurring of the electron beam 1 is measured and to define them again as the X and Y directions. (This also applies to Examples 2-12.)
However, if the directions in which the blurring of the electron beam 1 is measured are the original X and Y directions, the knife edge 20 also serves as a knife edge for correcting dimensional errors and rotation of the deflection field of the objective deflector 13. There is an advantage of being able to (Here, the deflection field of the objective deflector 13 consists of sides parallel to the original X or Y direction.)

Additionally, it is not a necessary condition for preventing the occurrence of the above situation that the difference in rotation angle is exactly π/4. Therefore, in this example, the difference in rotation angles of the distribution functions f _2A (ψ) and f _2B (ψ) may not be exactly π/4. That is, f _2A (φ) and f _2B (φ) need not be exactly orthogonal to each other. (If f _2A (φ) and f _2B (φ) are equal to cos2φ and sin2φ respectively at each electrode position as before, then f _2A (φ) and f _2B (φ) are orthogonal to each other.) This yields As described above, in this embodiment, it is only necessary to determine either one of the XY differences D _A and D _B , not both the XY differences D _A and D _B. Here, either one of D _A and D _B is more specifically one of D _A and D _B that is not zero, more preferably the absolute value of either one of D _A and D _B is not small.
More specifically, f _2A (ψ) and f _2B (ψ) may be linearly independent of each other. If these are linearly independent of each other, then the two-fold astigmatism caused by _V2A and the two-fold astigmatism caused by _V2B are also linearly independent of each other. 2-fold astigmatism can be generated. (That is, the 2-fold astigmatism on the knife edge 20 can be given a linearly independent 2-component change.) Doing so means that in this embodiment either D _A or D _B is non-zero. A requirement to reliably obtain one, whatever the angle of rotation of the two-fold astigmatism due to V _2A and V _2B , and whatever the direction in which the blur of the electron beam 1 is measured. is. (This also applies to Examples 2-12, except that D _A and D _B are replaced with D _hA and D _hB , respectively, in Examples 3 and 4.)

The XY differences D _A and D _B can be derived from the aberration diagram S _F +S ₂ . More specifically, the relationship between the XY differences D _A and D _B , the coefficient F and the two-fold astigmatism coefficient a _2C can be derived from the aberration diagram S _F +S ₂ . This is explained below.

(52) gives the real part of S _F +S ₂ . If x is the real part of S _F +S ₂ , x is as shown in equation (57).
x=Re(S _F +S ₂ )
=α{(F+Re(T _C ))cos θ+Im(T _C )sin θ} (57)

x depends on θ. When θ makes x the maximum or minimum, the derivative obtained by partially differentiating the equation (57) with θ becomes zero. Its derivative is as shown in equation (58).

（５８）式で示される導関数が零となれば、（５９）式が成立する。
（Ｆ＋Ｒｅ（Ｔ_Ｃ））ｓｉｎθ－Ｉｍ（Ｔ_Ｃ）ｃｏｓθ＝０（５９）
（５９）式を変形すれば、（６０）式が得られる。
Ｐ_１[｛（Ｆ＋Ｒｅ（Ｔ_Ｃ））／Ｐ_１｝ｓｉｎθ－（Ｉｍ（Ｔ_Ｃ）／Ｐ_１）ｃｏｓθ]
＝Ｐ_１（ｃｏｓλ・ｓｉｎθ＋ｓｉｎλ・ｃｏｓθ）
＝Ｐ_１ｓｉｎ（θ＋λ）＝０ （６０）
（６０）式より、（６１）式が成立する。（６１）式において、μは整数である。
θ＝－λ＋μπ （６１）
以降では、（６１）式を満たすθを、θ_μとする。即ち、（６１ａ）式が成立する。
θ_μ＝－λ＋μπ （６１ａ）
（６０）式中のＰ_１、ｃｏｓλ、およびｓｉｎλは、それぞれ、（６２）、（６３）、および（６４）式で与えられる。
Ｐ_１＝｛（Ｆ＋Ｒｅ（Ｔ_Ｃ））^２＋（Ｉｍ（Ｔ_Ｃ））^２｝^１／２ （６２）
ｃｏｓλ＝（Ｆ＋Ｒｅ（Ｔ_Ｃ））／Ｐ_１ （６３）
ｓｉｎλ＝－Ｉｍ（Ｔ_Ｃ）／Ｐ_１ （６４）

If the derivative given by equation (58) becomes zero, then equation (59) holds.
(F+Re(T _C )) sin θ−Im(T _C ) cos θ=0(59)
By transforming the equation (59), the equation (60) is obtained.
P ₁ [{(F+Re(T _C ))/P ₁ } sin θ−(Im(T _C )/P ₁ ) cos θ]
=P ₁ (cos λ·sin θ+sin λ·cos θ)
= P ₁ sin(θ+λ)=0 (60)
From the expression (60), the expression (61) is established. (61), μ is an integer.
θ=−λ+μπ (61)
Hereafter, θ that satisfies the equation (61) is assumed to be θ _μ . That is, the formula (61a) is established.
θ _μ =−λ+μπ (61a)
P ₁ , cos λ, and sin λ in equation (60) are given by equations (62), (63), and (64), respectively.
P ₁ = {(F+Re( _TC )) ² +(Im( _TC )) ² } ^1/2 (62)
cos λ=(F+Re( _TC ))/P ₁ (63)
sinλ=−Im(T _C )/P ₁ (64)

When .theta.=. _theta..mu ., x is as shown in equation (65) from equations (57) and (61a). However, the fact that cos θ _μ =cos(−λ+μπ)=+cos λ or −cos λ and sin θ _μ =sin(−λ+μπ)=−sin λ or +sin λ holds is used in deriving equation (65).
x=α{(F+Re(T _C )) cos θ _μ + Im(T _C ) sin θ _μ }
=±α{(F+Re(T _C )) ² +(Im(T _C )) ² }/P ₁
=±αP ₁ (65)

(65) represents the maximum or minimum value of x. If x _d (≧0) is the difference between the maximum value and the minimum value of x, x _d is given by equation (66).
x _d =2αP ₁ (66)
_xd is equal to the width in the X direction of the ellipse formed by the aberration figure _{SF +} S2, as shown in FIG. _xd thus contributes to the blurring of the electron beam 1 in the X direction.

_xd depends on _V2A . When V _2A minimizes _xd , the derivative obtained by partially differentiating equation (66) with V _2A becomes zero. Its derivative is as shown in equation (67). _xd also depends on _V2B . When _V2B minimizes _xd , the derivative obtained by partially differentiating equation (66) with _V2B is zero. Its derivative is as shown in equation (68). In equations (67) and (68), T _OR and T _OI represent the real and imaginary parts of the constant _TOC , respectively.

即ち、Ｖ_２Ａがｘ_ｄを最小とするとき、（６９）式が成立し、Ｖ_２Ｂがｘ_ｄを最小とするとき、（７０）式が成立する。ただし、（６９）式においては、Ｖ_２ＡがＶ_２ＡＸで表されており、（７０）式においては、Ｖ_２ＢがＶ_２ＢＸで表されている。Ｖ_２ＡＸおよびＶ_２ＢＸは、先述の定義（第０～第２の２回非点発生電圧の定義）から、いずれも、第１の２回非点発生電圧５１に相当する。
Ｖ_２ＡＸ＝－（ａ_２Ｒ／｜ａ_２Ｃ｜^２）Ｆ－（ａ_２ＲＴ_ＯＲ＋ａ_２ＩＴ_ＯＩ）／｜ａ_２Ｃ｜^２ （６９）
Ｖ_２ＢＸ＝（ａ_２Ｉ／｜ａ_２Ｃ｜^２）Ｆ＋（ａ_２ＩＴ_ＯＲ－ａ_２ＲＴ_ＯＩ）／｜ａ_２Ｃ｜^２ （７０）

That is, when _V2A minimizes _xd , equation (69) holds, and when _V2B minimizes _xd , equation (70) holds. However, in equation (69), _V2A is represented by _V2AX , and in equation (70), _V2B is represented by _V2BX . Both V _2AX and V _2BX correspond to the first two-fold astigmatism voltage 51 from the above definition (definition of the 0th to second two-fold astigmatism voltages).
V _2AX =−(a _2R /|a _2C | ² )F−(a _2R T _OR +a _2I T _OI )/|a _2C | ² (69)
V _2BX =(a _2I /|a _2C | ² )F+(a _2I T _OR −a _2R T _OI )/|a _2C | ² (70)

As can be seen from equation (69), V _2AX does not depend on V _2B , and as seen from equation (70), V _2BX does not depend on V _2A . This means that V _2AX and V _2BX do not depend on the components of the two-fold astigmatism on knife edge 20 due to V _2B and V _2A , respectively. However, this assumes that _{V2B is fixed while V2A is increased or decreased to determine V2AX , and} that _V2A is fixed while _V2B is increased or decreased to determine _V2BX . Here, fixing V _2B while increasing or decreasing V _2A and fixing V _2A while increasing or decreasing V _2B mean that the derivatives expressed in equations (67) and (68) are the partial derivatives, respectively. It corresponds to being a function.

(52), the imaginary part of S _F +S ₂ is also obtained. If the imaginary part of S _F +S ₂ is set to y, y is as shown in equation (71).
y=Im(S _F +S ₂ )
=α {(F−Re(T _C )) sin θ+(Im(T _C )) cos θ} (71)

If y _d (≧0) is obtained by the same method as that for obtaining x _d , y _d is as shown in equation (72).
y _d =2αP ₂ (72)
Here, _yd is equal to the Y-direction width of the ellipse formed by the aberration figure _{SF +} S2, as shown in FIG. Therefore, yd contributes to the _blurring of the electron beam 1 in the Y direction. P2 in equation ( ₇₂ ) is given by equation (73).
P ₂ = {(F−Re(T _C )) ² +(Im(T _C )) ² } ^1/2 (73)

_{yd depends on V2A .} When _V2A minimizes _yd , the derivative obtained by partially differentiating equation (72) with _V2A is zero. Its derivative is as shown in equation (74). _{yd also depends on V2B .} When _V2B minimizes _yd , the derivative obtained by partially differentiating equation (72) with _V2B is zero. Its derivative is as shown in equation (75).

即ち、Ｖ_２Ａがｙ_ｄを最小とするとき、（７６）式が成立し、Ｖ_２Ｂがｙ_ｄを最小とするとき、（７７）式が成立する。ただし、（７６）式においては、Ｖ_２ＡがＶ_２ＡＹで表されており、（７７）式においては、Ｖ_２ＢがＶ_２ＢＹで表されている。Ｖ_２ＡＹおよびＶ_２ＢＹは、先述の定義（第０～第２の２回非点発生電圧の定義）から、いずれも、第２の２回非点発生電圧５２に相当する。
Ｖ_２ＡＹ＝｛ａ_２Ｒ／｜ａ_２Ｃ｜^２｝Ｆ－（ａ_２ＲＴ_ＯＲ＋ａ_２ＩＴ_ＯＩ）／｜ａ_２Ｃ｜^２ （７６）
Ｖ_２ＢＹ＝－｛ａ_２Ｉ／｜ａ_２Ｃ｜^２｝Ｆ＋（ａ_２ＩＴ_ＯＲ－ａ_２ＲＴ_ＯＩ）／｜ａ_２Ｃ｜^２ （７７）

That is, when _V2A minimizes _{yd, equation (76) holds, and when V2B minimizes yd} , equation (77) holds. However, in equation (76), _{V2A is represented by V2AY ,} and in equation (77), _{V2B is represented by V2BY .} Both V _2AY and V _2BY correspond to the second two-time astigmatism voltage 52 from the above definition (definition of the 0th to second two-time astigmatism voltages).
V _2AY ={a _2R /|a _2C | ² }F−(a _2R T _OR +a _2I T _OI )/|a _2C | ² (76)
V _2BY =−{a _2I /|a _2C | ² }F+(a _2I T _OR −a _2R T _OI )/|a _2C | ² (77)

As can be seen from equation (76), V _2AY does not depend on V _2B , and as seen from equation (77), V _2BY does not depend on V _2A . This means that V _2AY and V _2BY are independent of the components of the two-fold astigmatism on knife edge 20 due to V _2B and V _2A , respectively. However, this assumes that _V2B is fixed while _V2A is increased or decreased to determine _V2AY , and that _V2A is fixed while _V2B is increased or decreased to determine _V2BY . Here, fixing V _2B while increasing or decreasing V _2A and fixing V _2A while increasing or decreasing V _2B mean that the derivatives expressed by equations (74) and (75) are the partial differential derivatives, respectively. It corresponds to being a function.

As can be seen from equations (69), (70), (76), and (77), V _2AX and V _2AY are different from each other unless a _{2R F} is zero, and V _2BX and V _2BY are different from each other. Here, _a2R and _a2I do not become zero at the same time since the objective deflector 13 functions as a two-time stigmator.
Therefore, in this embodiment, at least one of the difference between V _2AX and V _2AY and the difference between V _2BX and V _2BY has a non-zero value, so long as F is non-zero. This is the origin of the XY difference. That is, D _A and D _B cannot be zero simultaneously with each other unless F is zero.

Note that V _2AX , V _2BX , V _2AY , and V _2BY are all independent of α, as can be seen from equations (69), (70), (76), and (77). That is.
That is, V _2AX and V _2BX each minimize the size of the ellipse formed by the aberration diagram S _F +S ₂ for all α in the X direction, and therefore each X of electron beam 1 measured by knife edge 20X. Minimize directional blur. V _2AY and V _2BY each minimize the Y-direction size of the ellipse formed by the aberration diagram S _F +S ₂ for all α, and therefore each Y-direction Y-direction of the electron beam 1 measured by the knife edge 20Y. Minimize blur.

From equations (69), (70), (76) and (77), equations (78) and (79) are obtained. Equations (78) and (79) give the XY differences D _A and D _B , respectively. Here, D _A (A component of XY difference) and D _B (B component of XY difference) are V _2AX and V _2AY , and between V _2BX and V _2BY . In equations (78) and (79), the signs of D _A and D _B follow the rules previously described.
D _A =V _2AY -V _2AX =(2a _2R /|a _2C | ² )F(78)
D _B =V _2BX -V _2BY =(2a _2I /|a _2C | ² )F(79)

Neither the XY differences D _A and D _B are dependent on the _TOC , as can be seen from equations (78) and (79). Therefore, the XY differences D _A and D _B can each be evaluated regardless of the axial two-fold astigmatism and deflection two-fold astigmatism produced by the apparatus of this embodiment. Here, T _OC is T _C under the condition that V _2C is zero, as described above, and therefore, the axial 2-fold astigmatism and deflection 2-fold astigmatism produced by the apparatus of this embodiment are A constant that characterizes either or both.
Neither the XY differences D _A and D _B are also dependent on V _2A or V _2B , as can be seen from equations (78) and (79). Therefore, the XY difference D _A can be evaluated regardless of what values V _2A and V _2B were each set to before increasing or decreasing V _2A , and the XY difference D _B increases or decreases V _2B . It can be evaluated regardless of what values V _2A and V _2B were each previously set to. However, this assumes that V _2B is fixed while V _2A is increased or decreased to evaluate D _A and that V _2A is fixed while V _2B is increased or decreased to evaluate D _B .

D _A and D _B represented in equations (78) and (79) may each be formally divided by two. With D _A and D _B so amended, equations (78) and (79) are simplified as shown in equations (78a) and (79a), respectively.
D _A =(a _2R /|a _2C | ² )F (78a)
D _B =(a _2I /|a _2C | ² )F (79a)

Equations (78a) and (79a) can be converted into equations (78b) and (79b), respectively, if the rotation angle of the objective deflector 13 about the Z-axis satisfies a _2C =a _2R (that is, a _2I =0). As you can see, it gets even easier.
D _A =(1/a _2R )F (78b)
D _B =0 (79b)

Equation (80) is obtained from equation (78) and equations (5) and (47). Equation (81) is obtained from equation (79) and equations (6) and (47).

(80), the coefficient _dIA is determined from the defocus coefficient _b0 and the two-fold astigmatism coefficients _a2C and _a2R . (81), the coefficient _dIB is determined from the defocus coefficient _{b0 and the two-fold astigmatism coefficients a2C and a2I} . where the defocus coefficient b ₀ and the two-fold astigmatism coefficients a _2C , a _2R and a _2I do not depend on any of I _0A , V _0A , V _2A and V _2B and are therefore unchanged. . This is why the coefficients d _IA and d _IB do not depend on any of I _0A , V _0A , V _2A and V _2B and are therefore invariant.

The same applies to the coefficients _dVA and _dVB . The relationship between d _VA and a ₀ is represented by equation (82). Equation (82) is obtained from equation (78) and equations (38) and (47). The relationship between d _VB and a ₀ is represented by equation (83). Equation (83) is obtained from equation (79) and equations (39) and (47).

（８２）式から分かるように、ｄ_ＶＡは、ａ_０とａ_２Ｃおよびａ_２Ｒとから決まり、また、（８３）式から分かるように、ｄ_ＶＢは、ａ_０とａ_２Ｃおよびａ_２Ｉとから決まる。そのため、ｄ_ＶＡおよびｄ_ＶＢも、Ｉ_０Ａ、Ｖ_０Ａ、Ｖ_２Ａ、およびＶ_２Ｂのいずれにも依存せず、従って不変である。

As can be seen from equation (82), d _VA is determined from a ₀ and a _2C and a _2R , and as seen from equation (83), d _VB is determined from a ₀ and a _2C and a _2I Determined. Therefore, d _VA and d _VB also do not depend on any of I _0A , V _0A , V _2A and V _2B and are therefore unchanged.

Equation (84) is derived from equations (78) and (80) and equation (15). Equation (84) is also derived from equations (79) and (81) and equation (16). In equation (84), _ID is the coefficient F converted into the amount of change in the focus correction current, and is also the XY difference D _A or D _B converted into the amount of change in the focus correction current. However, when deriving equation (84), D _A0 in equation (15) and D _B0 in equation (16) are read as D _A and D _B , respectively.
I _D =F/b ₀ (84)
Equation (84) is derived from equations (78), (80) and (15) as well as from equations (79), (81) and (16) because the XY difference D _A and D _B means that they are equal when converted into the amount of change in the focus correction current.

Equation (85) is derived from equations (78) and (82) and equation (34). Equation (85) is also derived from equations (79) and (83) and (35). In equation (85), V _D (x, y) is the coefficient F(x, y) converted into the amount of change in focus correction current, and the XY difference D _A ( x,y) or D _B (x,y). where F(x,y) represents the coefficient F at deflection coordinates (x,y). However, when deriving formula (85), D _A in formula (78) and D _B in formula (79) are read as D _A (x, y) and D _B (x, y), respectively, V _D0 (x, y) in equations (34) and (35) is replaced with V _D (x, y), and D _A0 (x, y) and (35) in equation (34) D _B0 (x, y) in the formula is read as D _A (x, y) and D _B (x, y) respectively. In the following, no particular distinction is made between F(x, y) and F.
V _D (x, y)=F(x, y)/a ₀ (85)
Equation (85) is derived from equations (78), (82), and (34), and from equations (79), (83), and (35) because the XY difference D _A (x , y) and D _B (x, y) are equal to each other when converted to the amount of change in the focus correction voltage.

It can be seen from equations (78) and (79) that each of the XY differences D _A and D _B has a proportional relationship with the coefficient F. This proportional relationship is the basis for the fact that the XY difference serves as an index of the degree of defocus on the knife edge 20 .
It should be noted here that the proportional coefficient governing the proportional relationship is determined only by the combination of the two-fold astigmatism coefficients a _2C , a _2R and a _2I . As previously mentioned, these aberration coefficients do not depend on any of I _0A , V _0A , V _2A and V _2B and are therefore unchanged. (The proportional coefficient governing the proportional relationship is, for example, 2a _2R /|a _2C | ² in equation (78).)

From equations (78) and (79) it can also be seen that the XY differences D _A and D _B will not be zero at the same time as each other unless the factor F is zero. This is because, as described above, the two-fold astigmatism coefficients _a2R and _a2I do not become zero at the same time since the objective deflector 13 functions as a two-fold astigmatism corrector.
The same applies to the coefficients _dIA , _dIB , _dVA and _dVB . That is, as can be seen from equations (80) to (83), the coefficients _{d_IA} and _{d_IB} do not become zero at the same time, and the coefficients _{d_VA} and _dVB do not become zero at the same time.

Equations (78) and (79) ensure that the magnitude of defocus on knife edge 20 will be zero if any non-zero one of XY differences D _A and D _B is reduced to zero. This means that it is not necessary to evaluate both XY differences _{D A and D B even} if they both have non-zero magnitudes. (In this embodiment, one of D _A and D _B that is not zero may be evaluated, and more preferably, one of D _A and D _B whose absolute value is not small should be evaluated.)

Equations (78) and (79) ensure the above, firstly, by reducing the non-zero one of the XY differences D _A and D _B to zero, as a result of equation (86) or (87) is true, it follows from formula (78) or (79) that formula (88) must be true. ) because the formula holds. (88) and (89) do not depend on the half-angle of convergence α.
D _A =0 (86)
D _B =0 (87)
F=0 (88)
S _F =0 (89)
(86) or (87) by reducing the non-zero one of the XY differences D _A and D _B to zero, as can be seen from Eqs. (78) and (79), a Requires that at least one of _2R and _a2I is non-zero and that F transitions from a non-zero value to zero. That is, if one of the XY differences D _A and D _B that is not zero is reduced to zero, the equation (88) is always established.

Supplementally, the establishment of the equation (86) or (87) by reducing the non-zero one of the XY differences D _A and D _B to zero is actually the result of both equations (86) and (87). means establishment. This means that if the formula (88) is satisfied by the formula (86) or (87), then from the formulas (78) and (79), the XY differences D _A and D _B are both non-zero magnitudes because it cannot have

Supplementally, from equations (78), (79), and (80)-(83), if either or both of _dIA and _dVA have non-zero values, D _A is guaranteed to be non-zero, and either or both of d-- _IB and d-- _VB have non-zero values, then D-- _B is guaranteed to be non-zero when F.noteq.0.
As can be seen from equations (80) to (83), if either or both of d _IA and d _VA are not zero, a _2R is also not zero, and either d _IB and d _VB or Because if both are non-zero, a _2I is also non-zero.

From the equations (80) to (83), it can also be seen that the equation (32) always holds when the equation (19) holds, and the equation (33) holds whenever the equation (20) holds. I understand. That is, when each value of _dIA and _dIB is determined, it is decided which value of _dVA and _dVB should be determined. (Similarly, whenever formula (32) holds, formula (19) holds, and whenever formula (33) holds, formula (20) holds. That is, each of d _VA and d _VB When the value is determined, it is decided which value of _dIA or _dIB should be determined.)

As explained above, the apparatus of this embodiment measures and corrects its own axial defocus, axial two-fold astigmatism, deflection field curvature, and deflection two-fold astigmatism. .
The apparatus of this embodiment is characterized in that the measurement of the axial defocus and deflection curvature of field is based on the blur curve of the electron beam 1 with respect to the two-fold astigmatism correction voltage (two-fold astigmatism generation voltage). and More specifically, the apparatus of this embodiment determines the value of the two-fold astigmatism correction voltage that minimizes the blur in the X direction and the value of the two-fold astigmatism correction voltage that minimizes the blur in the Y direction from the curves. , and the difference (XY difference) is determined, and the difference represents the above-mentioned axial defocus and deflection field curvature aberrations.
Due to the above features, the apparatus of the present embodiment measures the axial defocus and deflection curvature of field aberration in a short time with a sufficient measurable range, and further calculates these aberrations based on the measurement results. can be corrected.
What is important here in acquiring the above curve is the speed of response of the objective deflector 13 as an electrostatic double stigmator to the double stigmator voltage and the speed of response to the double stigmator voltage. Accordingly, the size of the range in which the two-fold astigmatism on the knife edge 20 can be changed is utilized. That is, if the axial defocus and deflection curvature of field aberration are represented by the XY difference, these aberrations can be measured in a short time with a sufficient measurable range.

Supplementally, in this embodiment, all of the axial defocus, axial 2-fold astigmatism, deflection field curvature aberration, and deflection 2-fold astigmatism were measured and corrected, but it depends on the purpose. It is possible, then, to measure and correct some, but not all, of these aberrations. For example, one or both of the axial defocus and deflection field curvature aberration may be measured and corrected. Alternatively, only the axial defocus and axial 2-fold astigmatism may be measured and corrected, or the deflection field curvature and deflection 2-fold astigmatism may be measured and corrected.

Supplementally, in this embodiment, when obtaining the above curve, the two-fold astigmatism on the knife edge 20 is generated by the objective deflector 13 as a two-fold astigmatism corrector. This may be done by a dedicated electrostatic two-fold stigmator (not shown).
Alternatively, this may be done by a dedicated magnetic two-fold stigmator (not shown). In that case, instead of the blur curve of the electron beam 1 with respect to the double astigmatism correction voltage, A blur curve is obtained. That is, the double astigmatism correction current is increased or decreased instead of the double astigmatism correction voltage. (Thus, in that case, the XY differences D _A and D _B have dimensions of current rather than voltage.) In general, a magnetic two-fold stigmator operates faster than a magnetic lens, Therefore, the time required to acquire the blur curve of the electron beam 1 with respect to the double astigmatism correction current is shorter than the time required to acquire the blur curve of the electron beam 1 with respect to the focus correction current.

Supplementally, in this embodiment, the projection figure 11 is an image of the aperture of the shaping aperture plate (specifically, the image of the aperture of the first shaping aperture plate 3 and the aperture of the second shaping aperture plate 7). However, the projected figure 11 is a thin film that transmits the electron beam 1 as long as the blurring of the electron beam 1 can be evaluated by means such as a knife edge 20 (for example, Japanese Patent Application Laid-Open No. 08-83741 (see ). That is, either or both of the first shaping aperture plate 3 and the second shaping aperture plate 7 may be replaced with such a thin film. (This also applies to Examples 2-9.)

(Example 2)
Another embodiment of the charged particle beam device of the present invention will be described below as a second embodiment. Also in this embodiment, the charged particle beam apparatus of the present invention is configured as a variable shaping electron beam writing apparatus.

The apparatus of the present embodiment (variably shaped electron beam writing apparatus) has the same configuration as the apparatus of the first embodiment. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this example are the same as the configurations of the optical system, the measurement system, and the control system of the apparatus of Example 1 (see FIG. 1). .

The device of this embodiment basically operates in the same way as the device of the first embodiment. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection field curvature aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.
The above axial defocus and deflection field curvature aberration are represented by the XY difference D _A or D _B in this embodiment as well. Here, D _A or D _B is, more specifically, one of D _A and D _B that is not zero, and more preferably one of D _A and D _B that is not smaller in absolute value. be.

However, the apparatus of this embodiment reduces the total number of curves acquired for measuring the axial defocus, axial two-fold astigmatism, deflection field curvature, and deflection two-fold astigmatism to: It differs from the apparatus of the first embodiment. The total number is less in this embodiment than in the first embodiment.
More specifically, the apparatus of this embodiment first combines these aberrations together to reduce the total number of curves required to measure the axial defocus and axial two-fold astigmatism. and measure. The apparatus of this embodiment also measures the deflection field curvature aberration and the deflection 2-fold astigmatism in pairs to reduce the total number of curves required to measure these aberrations.
By doing so, the apparatus of this embodiment shortens the total time required for the measurement and correction of the above four aberrations. The details are shown below.

The apparatus of this embodiment first measures and corrects the axial defocus and axial two-fold astigmatism. Therefore, the apparatus of this embodiment first moves the knife edge 20 directly under the electron beam 1 by the material stage 14, and then blurs the electron beam 1 in the X and Y directions with respect to the two-time astigmatism correction voltage V _2A . (2 curves in total) are acquired. The apparatus of this embodiment then determines from these curves both the A component of the XY difference and the A component of the XY average under the 0th focus correction condition. That is, both D _A0 and M _A0 are determined. However, here, it is assumed that the formula (19) holds.

Here, M _A0 is given by equation (90). In equation (90), V _2AX0 and V _2AY0 represent V _2AX and V _2AY under the 0th focus correction condition, respectively. Equation (90) is derived from equations (69) and (76).
M _A0 =(V _2AX0 +V _2AY0 )/2=-(a _2R T _OR +a _2I T _OI )/|a _2C | ² (90)
As can be seen from equation (90), M _A0 does not depend on the factor F. That is, M _A0 is independent of defocus on knife edge 20 . (M _A0 does not depend on V _2A0 or V _2B0 .)

The key here is that the XY average M _A0 is determined from the electron beam 1 vs. V _2A X and Y blur curves (two total) to determine the D _A0 . .
That is, in this embodiment, the number of curves that must be acquired again for determining M _A0 is zero. The number was two in the first embodiment.

The apparatus of this embodiment then acquires curves (total one curve) of blurring of the electron beam 1 in the X direction or the Y direction with respect to the double astigmatism correction voltage _V2B . The apparatus of this embodiment then determines the B component of the XY average under the 0th focus correction condition. That is, _MB0 is determined.
M _B0 is given by equation (91). Equation (91) is derived from equations (70) and (77).
M _B0 =(V _2BX0 +V _2BY0 )/2=(a _2I T _OR −a _2R T _OI )/|a _2C | ² (91)
As can be seen from equation (91), M _B0 also does not depend on the factor F. That is, M _B0 is also independent of defocus on knife edge 20 . (M _B0 does not depend on V _2A0 or V _2B0 .)

In the apparatus of this embodiment, one of V _2BX0 and V _2BY0 to be substituted into equation (91) is determined from the blur curve of electron beam 1 with respect to V _2B , while the other is determined by (92) or ( 93) is determined from the formula. Both Equations (92) and (93) are derived from Equations (16) and (79). However, when deriving equations (92) and (93), V _2BX , V _2BY and D _B in equation (79) are read as V _2BX0 , V _2BY0 and D _B0 respectively.
V _2BY0 =V _2BX0 -D _B0 =V _2BX0 -I _{D0 dIB} (92)
V _2BX0 =V _2BY0 +D _B0 =V _2BY0 +I _{D0 dIB} (93)

In equations (92) and (93), _ID0 is the measurement of the axial defocus. I _D0 can be determined from the coefficient d _IA and the XY difference D _A0 as can be seen from equation (15). (D _A0 has been previously determined, together with M _A0 .) I _D0 d _IB (=D _B0 ) is I _D0 converted to D _B0 by the factor d _IB (see equation (16)). . (I _D0 d _IB is also D _A0 converted to D _B0 . As can be seen from equation (15), I _D0 is D _A0 converted to the amount of change in the focus correction current by the coefficient d _IA . by being
Therefore, the equation (92) is obtained from the first two-fold astigmatic voltage 51 (V _2BX0 ) and the measured value of the axial defocus converted to the XY difference D _B0 to obtain the second two-fold astigmatism It shows that the generated voltage 52 (V _2BY0 ) can be determined. Equation (93) shows that the first two-time astigmatism voltage 51 can be determined from the second two-time astigmatism voltage 52 and the measured value of the on-axis defocus converted to D _B0 . showing.
The coefficient _dIA can be determined from equation (5). _{d_IA} is determined and stored prior to the conversion of _{D_A0} to _{I_D0} by _{d_IA} (see equation (15)). The _{d_IA so stored is used repeatedly each time the d_IA converts D_A0} to _{I_D0} . The coefficient _dIB can be determined from equation (6). _dIB is determined and stored prior to the conversion of _ID0 to _DBO by _dIB (see equation (16)). The _dIB so stored is used repeatedly each time the _dIB converts I _D0 to D _B0 .

Substituting equation (92) into equation (91) yields equation (94). Substituting equation (93) into equation (91) yields equation (95). That is, the XY average M _B0 can be determined from equation (94) or (95) in this embodiment.
M _B0 =V _2BX0 −D _B0 /2=V _2BX0 −I _D0 d _IB /2 (94)
M _B0 =V _2BY0 +D _B0 /2 =V _2BY0 +I _{D0 dIB} /2 (95)
Equation (94) indicates that the XY average M _B0 can be determined from the first two-time astigmatic voltage 51 (V _2BX0 ) and the measured value of the axial defocus converted to the XY difference D _B0 . showing. Equation (95) shows that M _B0 can be determined from the second twice astigmatic voltage 52 (V _2BY0 ) and the measured value of the axial defocus converted to D _B0 .

What is important here is that as long as d _IB and I _D0 are known, as can be seen from equations (94) and (95) (and also from equations (91)-(93)), M _B0 is It can be determined from only one of V _2BX0 and V _2BY0 . That is, the curves required here are the curves of the electron beam 1 blurring in the X direction or the Y direction with respect to _V2B (one curve in total).
Therefore, in this embodiment, the number of curves that must be acquired again for determining M _B0 is one. The number was two in the first embodiment.

The apparatus of this embodiment then updates I _0A0 according to equation (9), inputs the updated I _0A0 to the objective lens 9, updates V _{2A0 according} to equation (24), and so V _2A0 updated as above is input to the objective deflector 13 . The apparatus of this embodiment further updates V _{2B0 according} to equation (25) and inputs the updated V _2B0 to the objective deflector 13 .
As a result, the axial defocus is corrected, and the component of the axial two-fold astigmatism that can be corrected by _V2A is corrected. In addition, the component of the on-axis two-fold astigmatism correctable by _V2B is also corrected.

In the above process, _I0A0 updated by equation (9) should be input to the objective lens 9 after obtaining the blur curve of the electron beam 1 in the X direction or the Y direction with respect to _V2B . That is, it is necessary to obtain the curve before inputting I _0A0 ′ to the objective lens 9 .
This is because I _D0 in equation (95) was determined before I _0A0 ′ was input to the objective lens 9, so V 2BX0 in equation (94) and V _2BX0 in equation (95). This is because V _2BY0 also needs to be determined before I _0A0 ′ is input to objective lens 9 .
On the other hand, V _2A0 updated by equation (24) may be input to the objective deflector 13 before or after acquisition of the curve of the blur in the X direction or Y direction of the electron beam 1 with respect to V _2B . .

From the acquisition of the above curves (the X- and Y-direction blur curves of the electron beam 1 for V _2A and the X- or Y-direction blur curves of the electron beam 1 for V _2B ), I A series of steps from the input of _0A0 to the objective lens 9 and the inputs of _V2A0 and _V2B0 to the objective deflector 13 are repeated. The series of steps is repeated until equations (22), (26) and (27) hold.
Here, the updates of I _0A0 , V _2A0 and V _2B0 are according to equations (9a), (24a) and (25a) respectively. M _A0 ^(m−1) in formula (24a) is given by formula (90a). M _B0 ^(m−1) in formula (25a) is given by formula (94a) or (95a). Equations (90a), (94a), and (95a) are generalizations of equations (90), (94), and (95), respectively.
M _A0 ^(m-1) = (V _2AX0 ^(m-1) +V _2AY0 ^(m-1) )/2 (90a)
M _B0 ^(m-1) = V _2BX0 ^(m-1) - D _B0 ^(m-1) /2
=V _2BX0 ^(m-1) -I _D0 d _IB ^(m-1) /2 (94a)
M _B0 ^(m−1) =V _2BY0 ^(m−1) +D _B0 ^(m−1) /2
=V _2BY0 ^(m-1) +I _D0 ^(m-1) d _IB /2 (95a)
In the equation (90a), V _2AX0 ^(m-1) inputs I _0A0 (m ^-1) to the objective lens 9 and inputs V _2A0 ^(m-1) and V _2B0 ^(m-1) to the objective deflector 13. represents V _2AX0 obtained after inputting V _2AY0 ^(m−1) is obtained after inputting I _0A0 ( ^m−1) into the objective lens 9 and inputting V _2A0 ^(m−1) and V _2B0 ^(m−1) into the objective deflector 13 represents V _2AY0 . In equation (94a), V _2BX0 ^(m−1) inputs I _0A0 (m ⁻¹⁾ into the objective lens 9 and inputs V _2A0 ^(m−1) and V _2B0 ^(m−1) into the objective deflector 13. represents V _2BX0 obtained after inputting In the equation (95a), V _2BY0 ^(m−1) inputs I _0A0 (m ⁻¹⁾ to the objective lens 9 and inputs V _2A0 ^(m−1) and V _2B0 ^(m−1) to the objective deflector 13. represents the V _2BY0 obtained after inputting
If the above series of steps is repeated, the correction residual of the axial defocus, the correction residual of the component of the axial two-fold astigmatism that can be corrected by _V2A , and the axial two-fold astigmatism The correction residual of the component of aberration that can be corrected by _V2B is reduced.

The above is an explanation of the procedure for measuring and correcting axial defocus and axial two-fold astigmatism produced by the apparatus of this embodiment. Measurement and correction of these aberrations follow routines similar to main routine 1 (see FIG. 6) or main routine 1' (see FIG. 9). Hereinafter, for convenience of explanation, it is assumed that the measurement and correction of these aberrations follow main routine 1 or main routine 1'. However, subroutine 2 (see FIG. 8) is replaced with subroutine 5.

Subroutine 5 is shown in FIG. Both the subroutine 5 and the subroutine (subroutine 6) therein are based on the premise that the formula (19) holds.

(Subroutine 5)
As shown in FIG. 18, subroutine 5 consists of the following steps.
First, in step S61, n is set to 0 (n=0) and m is set to 0 (m=0).
Next, I _0A0 ^(m) is input to the objective lens 9 in step S62. (I _0A0 ^(m) represents I _0A0 when m=0. Here, I _0A0 is equal to the value of the focus correction current I _0A immediately before the start of main routine 1 or main routine 1′.)
Next, in step S63, V _2A0 ^(m) and V _2B0 ^(m) are input to the objective deflector 13 . (V _2A0 ^(m) and V _2B0 ^(m) represent V _2A0 and V _2B0 , respectively, when m=0. Here, V _2A0 and V _2B0 are the are equal to the values of the two-fold astigmatism correction voltages _V2A and _V2B , respectively.)
Next, in step S64, subroutine 6 (see FIG. 19) is executed to determine the XY difference D _A0 ^(m) and the XY averages M _A0 ^(m) and M _B0 ^(m) .
Next, in step S65, it is determined whether or not |D _A0 ^(m) |<ε _D , |ΔM _A0 ^(m) |<ε _M and |ΔM _B0 ^(m) |<ε _M hold. If so, exit subroutine 5 and return to main routine 1 (see FIG. 6) or main routine 1' (see FIG. 9). If it does not hold, the process proceeds to step S66. (ΔM _A0 ^(m) and ΔM _B0 ^(m) in step S65 are given by equations (28) and (29), respectively. When m=0, M _A0 ^{(m−1) in equation (28)} and M _B0 ^(m−1) in formula (29) represent V _2A0 and V _2B0 , respectively.)
In step S66, m+1 is substituted for m (m=m+1). Then, the process proceeds to step S67.
In step S67, I _0A0 (m ⁾ , V _2A0 ^(m) and V _2B0 ^(m) are determined. Then, the process returns to step S62.

(Subroutine 6)
Subroutine 6, as shown in FIG. 19, consists of the following steps.
First, in step S71, the astigmatism correction voltage _V2A is increased and decreased twice to measure the blurring of the electron beam 1 in the X and Y directions. By doing so, curves of X and Y blurring of the electron beam 1 against V _2A are obtained.
Next, in step S72, the XY difference D _A0 (D _A0 ^(m) in step S64) and the XY average M _A0 (M _A0 ^(m) in step S64) are determined.
Next, in step S73, the astigmatism correction voltage _V2B is increased and decreased twice to measure the blur of the electron beam 1 in the X direction (or Y direction). By doing so, we obtain a curve of the X-direction (or Y-direction) blurring of the electron beam 1 against V _2B .
Next, in step S74, the XY average M _B0 (M _B0 ^(m) in step S64) is determined.
Then, the subroutine 6 is terminated and the subroutine 5 (see FIG. 18) is returned to.

The apparatus of this embodiment then measures and corrects the deflected curvature of field aberration and the deflected two-fold astigmatism produced by itself based on the same procedure as described above. To put it simply, the above-mentioned axial defocus and axial 2 The same is true for measuring and correcting the rotational astigmatism at each of the grid points in the deflection field of the objective deflector 13 .

More specifically, the apparatus of this embodiment first moves the knife edge 20 to the coordinates of the grid point where the above measurement and correction are performed, and moves the electron beam 1 so that the electron beam 1 is incident on the grid point. The electron beam 1 is deflected by the objective deflector 13 .
Next, the apparatus of this embodiment obtains curves (total of two lines) of blurring in the X and Y directions of the electron beam 1 with respect to the two-time astigmatic correction voltage V _2A by the knife edge 20 on the lattice points. Then, from these curves, both the A component of the XY difference and the A component of the XY average under the 0th focus correction condition are determined. That is, both D _A0 (x,y) and M _A0 (x,y) are determined. However, here, it is assumed that the formula (32) holds.
The apparatus of this embodiment further obtains curves (total one curve) of blurring in the X direction or Y direction of the electron beam 1 with respect to the two-time astigmatism correction voltage _V2B on the lattice points, and from the curves, The B component of the XY average under the 0th focus correction condition is determined. That is, M _B0 (x,y) is determined.
The apparatus of this embodiment then updates V _0A0 (x, y) according to equation (30), inputs the updated V _0A0 (x, y) to the objective deflector 13, and (40 ) update V _2A0 (x, y) according to the formula, and input the updated V _2A0 (x, y) to the objective deflector 13 . The apparatus of this embodiment further updates V _2B0 (x, y) according to equation (41) and inputs the updated V _2B0 (x, y) to the objective deflector 13 .
From the acquisition of the above curves (the X- and Y-direction blur curves of the electron beam 1 for V _2A and the X- or Y-direction blur curves of the electron beam 1 for V _2B ), V A series of steps is repeated up to the input of _0A0 (x, y), V _2A0 (x, y), and V _2B0 (x, y) to the objective deflector 13 . The series of steps is repeated on the lattice points until the equations (36), (42) and (43) hold. where V _0A0 (x,y), V _2A0 (x,y), and V _2B0 (x,y) are updated according to equations (30a), (40a), and (41a), respectively.
The apparatus of this embodiment thereafter performs the same measurements and corrections as above for the remaining grid points in the deflection field.
The apparatus of this embodiment then determines an approximate curved surface V _0A0F (x, y) from V _0A0 ^(m) (x, y) on all grid points in the deflection field, and From each of V _2A0 ^(m) (x, y) and V _2B0 ^(m) (x, y) on the lattice points, an approximate curved surface is determined for each of them.
The apparatus of this embodiment corrects the deflection field curvature aberration and the deflection two-fold astigmatism over the entire deflection field of the objective deflector 13 based on the approximate curved surface thus determined. .

In the above manner, M _A0 (x,y) is determined by equation (96) and substituted into equation (40). M _B0 (x,y) is determined by equation (97) or (98) and substituted into equation (41).
M _A0 (x, y)=(V _2AX0 (x, y)+V _2AY0 (x, y))/2 (96)
M _B0 (x, y)=V _2BX0 (x, y)−V _D0 (x, y)d _VB /2 (97)
M _B0 (x, y)=V _2BY0 (x, y)+V _D0 (x, y)d _VB /2 (98)
In equations (96) to (98), V _2AX0 (x, y) and V _2BX0 (x, y) are V _2AX0 and V _2BX0 on the grid point located at the deflection coordinates (x, y) (both A first two-fold astigmatic voltage 51) is represented respectively. V _2AY0 (x, y) and V _2BY0 (x, y) are V _2AY0 and V _2BY0 on the grid point located at the deflection coordinates (x, y) (both are the second double astigmatism voltage 52 ), respectively.
Formula (96) replaces M _A0 , V _2AX0 , and V _2AY0 in formula (90) with M _A0 (x, y), V _2AX0 (x, y), and V _2AY0 (x, y), respectively. is obtained by Equation (97) is derived from equations (99) and (100). Equation (98) is derived from equations (99) and (101).
M _B0 (x, y)=(V _2BX0 (x, y)+V _2BY0 (x, y))/2 (99)
V _2BY0 (x,y)=V _2BX0 (x,y)−D _B0 (x,y)
=V _2BX0 (x, y)-V _D0 (x, y)d _VB (100)
V _2BX0 (x, y)=V _2BY0 (x, y)+D _B0 (x, y)
= _V2BY0 (x,y)+ _VD0 (x,y) _dVB (101)
Formula (99) replaces M _B0 , V _2BX0 , and V _2BY0 in formula (91) with M _B0 (x, y), V _2BX0 (x, y), and V _2BY0 (x, y), respectively. is obtained by Equations (100) and (101) are both derived from equations (35) and (79). However, when deriving equations (100) and (101), D _B , V _2BX and V _2BY in equation (79) are respectively D _B0 (x, y), V _2BX0 (x, y), and V _2BY0 (x, y).

In equations (97), (98), (100), and (101), V _D0 (x,y) is a measurement of the deflected field curvature aberration. V _D0 (x, y) can be determined from the coefficient d _VA and the XY difference D _A0 (x, y), as can be seen from equation (34). (D _A0 (x, y) has been previously determined together with M _A0 (x, y).) V _D0 (x, y)d _VB (=D _B0 (x, y)) is the coefficient d It is V _D0 (x, y) converted to D _B0 (x, y) by _VB (see equation (35)). (V _D0 (x, y)d _VB is also D _A0 (x, y) converted to D _B0 (x, y), which is V _D0 (x, y) is D _A0 (x, y) converted into the amount of change in the focus correction voltage by the coefficient d _VA ).
Therefore, the equation (97) is obtained by the measurement of the first two-fold astigmatic voltage 51 (V _2BX0 (x, y)) and the deflection curvature of field converted into the XY difference D _B0 (x, y). values, the XY average M _B0 (x, y) can be determined. Equation (98) is obtained from the second two-time astigmatic voltage 52 (V _2BY0 (x, y)) and the measurement value of the deflection curvature of field converted into D _B0 (x, y), It shows that M _B0 (x,y) can be determined. Equation (100) is derived from the first two-fold astigmatic voltage 51 and the measured deflection field curvature aberration converted to D _B0 (x, y) to obtain the second two-fold astigmatic voltage 52 can be determined. Equation (101) is derived from the second two-fold astigmatic voltage 52 and the measured deflection field curvature aberration converted to D _B0 (x, y) to obtain the first two-fold astigmatic voltage 51 can be determined.
The coefficient _dVA can be determined from equation (38). d _VA is determined and stored prior to the conversion of D _A0 (x,y) to V _D0 (x,y) by d _VA (see equation (34)). The d _VA so stored is used repeatedly each time the d _VA converts D _A0 (x,y) to V _D0 (x,y). The coefficient _dVB can be determined from equation (39). d _VB is determined and stored prior to conversion of V _D0 (x,y) to D _B0 (x,y) by d _VB (see equation (35)). The d _VB thus stored is used repeatedly for each conversion of V _D0 (x,y) to D _B0 (x,y) by d _VB .

Generalizing Eqs. (96), (97), and (98), Eqs. (96), (97), and (98) become Eqs. (96a), (97a), and (98a), respectively. . M _A0 (x,y) ^(m−1) determined by equation (96a) is substituted into equation (40a). M _B0 ^(m−1) (x,y) determined by equation (97a) or (98a) is substituted into equation (41a).
M _A0 ^(m−1) (x, y)
=(V _2AX0 ^(m−1) (x, y)+V _2AY0 ^(m−1) (x, y))/2 (96a)
M _B0 ^(m−1) (x, y)
=V _2BX0 ^(m-1) (x, y)-V _D0 ^(m-1) (x, y)d _VB /2 (97a)
M _B0 ^(m−1) (x, y)
=V _2BY0 ^(m-1) (x, y)+V _D0 ^(m-1) (x, y)d _VB /2 (98a)
In equation (96a), V _2AX0 ^(m-1) (x, y) is V _0A0 (m-1 ⁾ (x, y), V _2A0 ^(m-1) (x, y), and V _2B0 ^{( m−1)} represents V _2AX0 (x,y) obtained after inputting all of (x,y) into the objective deflector 13 . V _2AY0 ^(m-1) (x, y) is V _0A0 (m-1 ⁾ (x, y), V _2A0 ^(m-1) (x, y), and V _2B0 ^(m-1) (x , y) are all input to the objective deflector 13, V _2AY0 (x, y). (97a), V _2BX0 ^(m-1) (x, y) is V _0A0 (m-1 ⁾ (x, y), V _2A0 ^(m-1) (x, y), and V _2B0 ^{( m−1)} represents V _2BX0 (x,y) obtained after inputting all of (x,y) into the objective deflector 13 . (98a), V _2BY0 ^(m−1) (x, y) is V _0A0 (m−1 ⁾ (x, y), V _2A0 ^(m−1) (x, y), and V _2B0 ^{( m−1)} represents V _2BY0 (x,y) obtained after inputting all of (x,y) into the objective deflector 13 .

What is important in the above procedure is that at each grid point in the deflection field of the objective deflector 13, the XY average M _A0 (x, y) _is V is determined from the X and Y blur curves (two total) of electron beam 1 for _2A , and the XY average M _B0 (x,y) is determined from the X or Y direction of electron beam 1 for V _2B . It is to be determined from the blur curves (one in total).
That is, in this embodiment, the number of curves that must be acquired again for determining M _A0 (x, y) and M _B0 (x, y) on each grid point is zero and one, respectively. is. On the other hand, in Example 1, the number of curves that must be acquired again for determining M _A0 (x, y) and M _B0 (x, y) on each grid point is two. there were.

However, in the above procedure, inputting V _0A0 (x, y) updated by the formula (30a) to the objective deflector 13 is to acquire the curve of the blurring of the electron beam 1 in the X direction or the Y direction with respect to V _2B . must be after That is, it is necessary to acquire the curve at each lattice point before inputting V _0A0 ^(m) (x, y) to the objective deflector 13 .
This is because V _D0 ^(m−1) (x, y) in equations (97a) and (98a) is determined before V _0A0 ^(m) (x, y) is input to the objective deflector 13. Therefore, V 2BX0 (m−1) (x, y) in equation (97a) and V _2BY0 ^(m−1) (x, y) in equation (98a) _are also V _0A0 ^{( m )} (x, y) must be determined before input to the objective deflector 13 .

The above is a description of the procedure for measuring and correcting deflection curvature of field aberration and deflection 2-fold astigmatism produced by the apparatus of this embodiment. Measurement and correction of these aberrations follow a routine similar to main routine 2 (see FIG. 10). Hereinafter, for convenience of explanation, it is assumed that the measurement and correction of these aberrations follow the main routine 2. However, subroutine 4 (see FIG. 12) in subroutine 3 (see FIG. 11) is replaced with subroutine 7. FIG. Further, step S39 (determination of the approximate curved surface V _0A0F (x, y)) is performed by determining the approximate curved surface V _0A0F (x, y), the two-time astigmatism correction voltages V _2A0 ^(m) (x, y) and V _2B0 ^{( m)} determining an approximate surface for each of (x,y).

Subroutine 7 is shown in FIG. Both the subroutine 7 and the subroutine (subroutine 6) therein are based on the premise that the formula (32) holds.

(Subroutine 7)
As shown in FIG. 20, subroutine 7 is substantially the same as subroutine 5 (see FIG. 18). More specifically, subroutine 7 corresponds to steps S62 and S67 in subroutine 5 replaced with steps S82 and S87, respectively. At step S82, V _0A0 ^(m) is input to the objective deflector 13 . In step S87, V _0A0 (m ⁾ , V _2A0 ^(m) and V _2B0 ^(m) are determined. (V _0A0 ^(m) in step S82 represents V _0A0 when m=0. Here, V _0A0 is equal to the value of the focus correction voltage V _0A immediately before the start of main routine 2. Its value is Depends on the deflection coordinates (x,y) V _2A0 ^(m) and V _2B0 ^(m ) in step S83 represent V _2A0 and V _2B0 respectively when m=0, where V _2A0 and V _2B0 are equal to the values of the two-time astigmatism correction voltages V2A and V2B, respectively, immediately before the start of main routine 2. _These values depend on the deflection coordinates (x, y) _{.ΔM A0} ^(m) in step S85 and ΔM _B0 ^(m) denote ΔM _A0 ^(m) (x, y) and ΔM _B0 ^(m) (x, y), respectively, given by equations (44) and (45), respectively.(44) M _A0 ^(m−1) (x, y) in the formula and M _B0 ^(m−1) (x, y) in the formula (45) are respectively V _2A0 (x, y) when m=0 and V _2B0 (x, y).)

As described above, the apparatus of this embodiment, like the apparatus of the first embodiment, has its own axial defocus, axial two-fold astigmatism, deflection curvature of field aberration, and deflection two-fold non-aberration. Astigmatism is both measured and corrected based on the blur curve of the electron beam 1 versus twice astigmatic voltage.
However, the apparatus of this embodiment measures the axial defocus and the axial 2-fold astigmatism in pairs. The apparatus of this embodiment further measures the deflection curvature of field aberration and the deflection 2-fold astigmatism in pairs.
More specifically, the apparatus of this embodiment measures the X-direction and Y-direction blurring of the electron beam 1 with respect to the two-fold astigmatism correction voltage V _2A for the measurement of the axial defocus and the axial two-fold astigmatism. , both the XY difference D _A0 and the XY average M _A0 are determined from the curves (total of two curves), and the curve of the _blurring of the electron beam 1 in the X direction or the Y direction (total 1), determine the XY average M _B0 . In order to measure the deflection curvature of field and the deflection 2-fold astigmatism, the apparatus of this embodiment also measures the XY difference D _A0 (x, y) and XY average M _A0 (x, y) are determined, and XY average M _B0 (x, y) is determined.
By doing so, the apparatus of this embodiment reduces the total number of curves required for measuring the axial defocus and the axial 2-fold astigmatism, and also reduces the deflection curvature of field and deflection 2 Reduce the total number of curves required to measure rotational astigmatism. More specifically, the apparatus of this embodiment reduces the number of curves that must be reacquired to determine the XY averages M _A0 and M _B0 and also reduces the XY averages M _A0 (x,y) and M _B0 Reduce the number of curves that must be acquired anew for the determination of (x,y).
Therefore, in this embodiment, the total number of curves required for measuring and correcting the four aberrations is reduced, and the total time required for measuring and correcting the four aberrations is shortened.

Supplementally, in this embodiment, all of the axial defocus, axial 2-fold astigmatism, deflection field curvature aberration, and deflection 2-fold astigmatism were measured and corrected, but it depends on the purpose. But it doesn't have to be. For example, the axial defocus and axial 2-fold astigmatism may only be measured and corrected, or the deflection field curvature and deflection 2-fold astigmatism may only be measured and corrected.

Supplementally, in this embodiment, M _B0 and M _B0 (x, y) on each grid point are respectively defined as the blurring of the electron beam 1 in the X direction or the Y direction with respect to the two-time astigmatism correction voltage V _2B . (one total), each determined from the X and Y blur curves (two total) of the electron beam 1 versus V _2B as in Example 1. good too. That is, M _B0 and M _B0 (x, y) on each grid point may be determined from equations (91) and (99), respectively.

(Example 3)
Still another embodiment of the charged particle beam apparatus of the present invention (variably shaped electron beam writing apparatus) will be described below as Embodiment 3. FIG.

The apparatus of this embodiment (variably shaped electron beam writing apparatus) has the same configuration as the apparatuses of the first and second embodiments. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this example are the same as the configurations of the optical system, the measurement system, and the control system of the apparatus of Example 1 (see FIG. 1). .

The device of this embodiment basically operates in the same way as the device of the first embodiment. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.

The apparatus of this embodiment, however, represents the magnitude of defocus on the knife edge 20 not by the XY difference, but by the difference between the first two-fold astigmatic voltage 51 and the XY average. (The XY difference is the difference between the first two-fold astigmatic voltage 51 and the second two-fold astigmatic voltage 52. The XY average is the difference between the first two-fold astigmatic voltage 51 and the second two-fold astigmatic voltage. This is the average with the two-fold astigmatism voltage 52.)

If the A component and the B component of the difference between the first double astigmatic voltage 51 and the XY average are _DhA and _DhB respectively, then _DhA and _DhB are (102) and (103) respectively. can be expressed by the formula
_DhA = M _A - V _2AX (102)
D _hB = V _2BX - M _B (103)
Equation (104) can be derived from equation (102) and equations (78) and (90), and equation (105) can be derived from equation (103) and equations (79) and (91). However, when deriving equation (104), M _A0 , V _2AX0 and V _2AY0 in equation (90) are read as M _A , V _2AX and V _2AY respectively. When deriving equation (105), M _B0 , V _2BX0 and V _2BY0 in equation (91) are read as M _B , V _2BX and V _2BY , respectively.
D _hA =D _A /2 (104)
D _hB =D _B /2 (105)

That is, D _hA represented by formulas (102) and (104) is the same as D _A represented by formula (78a), and D hB represented by formulas (103) and (105) is the same as D _hB represented by formula (79a). It is the same as the D _B that represents In this sense, this embodiment is substantially the same as the first embodiment.

It is not necessary to evaluate both D _hA and D _hB in this example, just as it was not necessary to evaluate both D _A and D _B in Example 1 (and Example 2). More specifically, in this embodiment, one of D _hA and D _hB that is not zero should be evaluated, and more preferably one of D _hA and D _hB that has a non-small absolute value should be evaluated. Just do it. That is, in this embodiment, the defocus on the knife edge 20 is represented by either _{DhA or DhB that is not zero, and more preferably, whichever of DhA and DhB has a non} -small absolute value. is represented by

D _hA and D _hB are determined at the center of the deflection field of the objective deflector 13 for the measurement of the axial defocus, and for the measurement of the deflection field curvature aberration, the multiple values within that deflection field. Determined on grid points. If D _{hA (x, y) and D hB (x, y) are D hA ( x} , y) and D _hB (x, y) on the grid point located at the deflection coordinates (x, y), then D _hA (x, y) and D _hB (x, y) can be represented by equations (106) and (107), respectively. D _hA (x,y) and D _hB (x,y) can also be expressed in equations (108) and (109), respectively.
D _hA (x, y)=M _A (x, y)−V _2AX (x, y) (106)
D _hB (x, y)=V _2BX (x, y)−M _B (x, y) (107)
D _hA (x, y)=D _A (x, y)/2 (108)
D _hB (x, y)=D _B (x, y)/2 (109)

I _0A0 with D _hA and D _hB as zero is determined by equations (110) and (111) respectively. V _0A0 (x, y) that makes D _hA (x, y) and D _hB (x, y) zero is determined by equations (112) and (113), respectively. Equations (110) and (111) are used when equations (114) and (115) hold, respectively. Equations (112) and (113) are used when equations (116) and (117) hold, respectively.
I _0A0 ′=I _0A0 −D _hA0 /d _hIA =I _0A0 −I _D0 (110)
I _0A0 ′=I _0A0 −D _hB0 /d _hIB =I _0A0 −I _D0 (111)
V _0A0 ′(x, y)=V _0A0 (x, y)−D _hA0 (x, y)/d _hVA
= V _0A0 (x, y) - V _D0 (x, y) (112)
V _0A0 ′(x, y)=V _0A0 (x, y)−D _hB0 (x, y)/d _hVB
= V _0A0 (x, y) - V _D0 (x, y) (113)
|d _hIA |≧|d _hIB | (114)
|d _hIA |<|d _hIB | (115)
|d _hVA |≧|d _hVB | (116)
|d _hVA |<|d _hVB | (117)
Formula (110) is a formula obtained by replacing D _A0 and d _IA in formula (9) with D _hA0 and d _hIA , respectively. Formula (111) is a formula obtained by replacing D _B0 and d _IB in formula (10) with D _hB0 and d _hIB , respectively. D _hA0 and D _hB0 represent D _hA and D _hB under the 0th focus correction condition, respectively. Equation (112) is obtained by replacing D _A0 (x, y) and d _VA in Equation (30) with D _hA0 (x, y) and d _hVA , respectively. Equation (113) is obtained by replacing D _B0 (x, y) and d _VB in Equation (31) with D _hB0 (x, y) and d _hVB , respectively. D _hA0 (x, y) and D _hB0 (x, y) represent D _hA (x, y) and D _hB (x, y) under the 0th focus correction condition, respectively.

I _D0 in equations (110) and (111) represents D _hA0 and D _hB0 converted into the amount of change in the focus correction current, and is given by equations (118) and (119), respectively. Equations (118) and (119) indicate that the apparatus of this embodiment converts the difference between the first two-time astigmatic voltage 51 and the XY average into the amount of change in the focus correction current. . The conversion is by factor d _hIA or d _hIB .
I _D0 =D _hA0 /d _hIA (118)
I _D0 =D _hB0 /d _hIB (119)
Here, _ID0 also represents the axial defocus converted into the amount of change in the focus correction current. This is because D _hA0 and D _hB0 both represent the axial defocus. (I _D0 is also the measured on-axis defocus above.)
Therefore, equations (110) and (111) show that the apparatus of this embodiment subtracts the above-described axial defocus converted to the amount of change in the focus correction current from _I0A0 , and thereby updates _I0A0 ( That is, it indicates that I _0A0 ′ is determined).
V _D0 (x, y) in equations (112) and (113) represents D _hA0 (x, y) and D _hB0 (x, y) converted to the amount of change in the focus correction voltage, respectively, and (120 ) and (121), respectively. Equations (120) and (121) indicate that the apparatus of this embodiment converts the difference between the first two-time astigmatic voltage 51 and the XY average into the amount of change in the focus correction voltage. . The conversion is by the coefficient d _hVA or d _hVB .
V _D0 (x, y)=D _hA0 (x, y)/d _hVA (120)
V _D0 (x, y)=D _hB0 (x, y)/d _hVB (121)
Here, V _D0 (x, y) also represents the deflection curvature of field aberration converted into the amount of change in the focus correction voltage. This is because both D _hA0 (x, y) and D _hB0 (x, y) represent the deflection field curvature aberration. (V _D0 (x, y) is also the measured value of the deflection field curvature aberration.)
Therefore, equations (112) and (113) are obtained by subtracting the deflection curvature of field converted into the amount of change in the focus correction voltage from V _0A0 (x, y) by the apparatus of this embodiment, (ie, determine V _0A0 ′(x,y)) by V _0A0 (x,y).

Coefficients d _hIA and d _hIB , like coefficients d _IA and d _IB , can be determined from equations (5) and (6), respectively. The coefficients _{d_hVA} and _{d_hVB} , like the coefficients _{d_VA} and _{d_VB} , can be determined from equations (38) and (39), respectively. However, when determining d _hIA , d _IA and D _An (n=0 or 1) in formula (5) are read as d _hIA and D _hAn , respectively. When determining d _hIB , d _IB and D _Bn (n=0 or 1) in formula (6) are read as d _hIB and D _hBn , respectively. When determining d _hVA , d _VA and D _An (0,0) (n=0 or 1) in formula (38) are read as d _hVA and D _hAn (0,0), respectively. When determining d _hVB , d _VB and D _Bn (0,0) (n=0 or 1) in equation (39) are read as d _hVB and D _hBn (0,0), respectively.
The apparatus of this embodiment determines and stores the coefficient d _hIA prior to the conversion of D _hA0 to I _D0 by d _hIA (see equation (118)). The d _hIA thus stored is used repeatedly each time the d _{hIA converts} D _hA0 to I _D0 . Alternatively, the apparatus of this embodiment determines and stores the coefficient d _hIB prior to the conversion of D _hB0 to I _D0 by d _hIB (see equation (119)). The d _hIB so stored is used repeatedly for each conversion of D _hB0 to I _D0 by d _hIB . where d _hIA and d _hIB are stored when equations (114) and (115) hold, respectively.
The apparatus of this embodiment further determines and stores a coefficient d _{hVA prior} to conversion of D _hA0 (x,y) to V _D0 (x,y) by d _hVA (see equation (120)). do. The d _hVA thus stored is used repeatedly each time the d _{hVA converts} D _hA0 (x,y) to V _D0 (x,y). Alternatively, the apparatus of this embodiment determines and stores the coefficient d _{hVB prior} to conversion of D _hB0 (x,y) to V _D0 (x,y) by d _hVB (see equation (121)). do. The d _hVB thus stored is used repeatedly each time the d _{hVB converts} D _hB0 (x,y) to V _D0 (x,y). where _{d_hVA} and _{d_hVB} are stored when equations (116) and (117) hold, respectively.

The coefficients d _hIA , d _hIB , d _hVA , and d _hVB and the coefficients d _IA , d _IB , d _VA , and d _VB have relationships represented by equations (122) to (125). Therefore, if the coefficients _dIA , _dIB , _dVA , and _dVB are known, the coefficients _dhIA , _dhIB , _dhVA , and _dhVB are derived from _dIA , _dIB , _dVA , and _dVB , respectively. Desired.
d _hIA =d _IA /2 (122)
d _hIB =d _IB /2 (123)
d _hVA =d _VA /2 (124)
d _hVB =d _VB /2 (125)

Generalizing the equations (110) to (113) and (118) to (121), the equations (110) to (113) and (118) to (121) are, respectively, (110a) to (113a) ), and formulas (118a) to (121a). Formulas (110a) to (113a) and (118a) to (121a) also serve as formulas (110) to (113) and (118) to (121), respectively, unless otherwise specified.
I _0A0 ^(m) =I _0A0 (m ^-1) -D _hA0 ^(m-1) /d _hIA =I _0A0 ^(m-1) -I _D0 ^(m-1) (110a)
I _0A0 ^(m) =I _0A0 (m ^-1) -D _hB0 ^(m-1) /d _hIB =I _0A0 ^(m-1) -I _D0 ^(m-1) (111a)
V _0A0 ^(m) (x, y)=V _0A0 ( ^m-1) (x, y)-D _hA0 ^(m-1) (x, y)/d _hVA
= V _0A0 ^(m-1) (x, y) - V _D0 ^(m-1) (x, y) (112a)
V _0A0 ^(m) (x, y)=V _0A0 ( ^m-1) (x, y)-D _hB0 ^(m-1) (x, y)/d _hVB
=V _0A0 ^(m-1) (x, y)-V _D0 ^(m-1) (x, y) (113a)
I _D0 ^(m−1) = D _hA0 ^(m−1) /d _hIA (118a)
I _D0 ^(m-1) = D _hB0 ^(m-1) /d _hIB (119a)
V _D0 ^(m−1) (x, y)=D _hA0 ^(m−1) (x, y)/d _hVA (120a)
V _D0 ^(m−1) (x, y)=D _hB0 ^(m−1) (x, y)/d _hVB (121a)

In order to reduce the correction residual error of the axial defocus, the apparatus of this embodiment acquires curves of blurring in the X and Y directions of the electron beam 1 with respect to the two-time astigmatism correction voltage V _2A (110a). A series of steps up to the input of _I0A0 updated by the equation to the objective lens 9 are repeated until the equation (126) is established. Alternatively, for the same purpose, the device of the present embodiment can be obtained from obtaining curves of blurring in the X and Y directions of the electron beam 1 with respect to the twice astigmatic voltage V _2B , I _0A0 updated by equation (111a). A series of steps up to the input to the objective lens 9 are repeated until the equation (127) holds.
|D _hA0 ^(m) |<ε _D /2 (126)
|D _hB0 ^(m) |<ε _D /2 (127)
Further, in order to reduce the correction residual error of the deflection field curvature aberration, the apparatus of the present embodiment is configured such that, on each grid point in the deflection field of the objective deflector 13, the electron beam is applied twice to the astigmatism correction voltage _V2A . 1 to the input of V _0A0 (x, y) updated by equation (112a) to the objective deflector 13 by equation (128) repeat until is established. Alternatively, for the same purpose, the apparatus of the present embodiment can obtain curves of blurring in the X and Y directions of the electron beam 1 with respect to the twice astigmatic voltage V _2B on each grid point (113a). A series of steps up to the input of V _0A0 (x, y) updated by the equation to the objective deflector 13 are repeated until the equation (129) is established. The apparatus of this embodiment then determines an approximate curved surface V _0A0F (x, y) from V _0A0 ^(m) (x, y) on all lattice points in the deflection field, and thereafter V _0A0F (x, y), the deflection field curvature aberration is corrected over the entire deflection field of the objective deflector 13 .
|D _hA0 ^(m) (x, y)|<ε _D /2 (128)
|D _hB0 ^(m) (x, y)|<ε _D /2 (129)

The manner in which the apparatus of this example measures and corrects the axial two-fold astigmatism and deflection two-fold astigmatism is the same as the manner in which the apparatus of Example 1 measures and corrects the same aberrations. can be That is, the determination of the XY average and the update and input of the two-time astigmatism correction voltage may be performed in the following manner. (Hereinafter, a description will be given of reduction of correction residuals for the axial two-fold astigmatism and deflection two-fold astigmatism, and two-fold astigmatism correction voltage V _2A0 ^{(m )} (x, y) and V _2B0 ^(m) (x, y) are omitted.)
First, the XY average M _A0 is determined from the blur curves in the two directions (X direction and Y direction) of the electron beam 1 with respect to the double astigmatism correction voltage V _2A , and V _2A0 is updated by the equation (24). V _2A0 updated as above is input to the objective deflector 13 . Furthermore, the XY average M _B0 is determined from the curves of blurring in the two directions (X direction and Y direction) of the electron beam 1 with respect to the double astigmatism correction voltage V _2B , and V _2B0 is updated by the equation (25). V _2B0 updated as above is input to the objective deflector 13 .
As a result, the component of the on-axis two-fold astigmatism that can be corrected by _V2A is corrected, and the component of the on-axis two-fold astigmatism that can be corrected by _V2B is corrected.
Next, steps similar to the series of steps described above are performed on each lattice point within the deflection field of the objective deflector 13 . That is, on each grid point, the XY average M _A0 (x, y) is determined from the curves of blurring in the two directions (X direction and Y direction) of the electron beam 1 with respect to V _2A , and V _2A0 (x, y) is updated, and the updated V _2A0 (x, y) is input to the objective deflector 13 . On each grid point, the XY average M _B0 (x, y) is further determined from the blur curves in the two directions (X direction and Y direction) of the electron beam 1 with respect to V _2B , and V _2B0 (x, y) and then input the updated V _2B0 (x, y) to the objective deflector 13 .
As a result, the component of the 2-fold deflection astigmatism that can be corrected by V2A is corrected on each grid point, and the component of the ₂ -fold deflection astigmatism that can be corrected by _V2B is corrected. be done.

However, in this embodiment, M A0 is required to determine D _hA0 as seen from equation (102), and M _A0 is required to determine D _hA0 (x, y) as seen from equation (106). _A0 (x,y) is required. Alternatively, as can be seen from equation (103), determination of D _hB0 requires M _B0 , and as seen from equation (106), determination of D _hB0 (x, y) requires M _B0 (x, y) Is required.
Therefore, in this embodiment, when _DhA0 is determined from the curve of the blurring of the electron beam 1 in two directions with respect to _V2A , M _A0 is also determined from the same curve, and _V2A When D _hA0 (x,y) is determined from the two-way blur curve of the electron beam 1 against , M _A0 (x,y) is also determined from the same curve. Alternatively, when D _hB0 is determined from the two-way blur curve of electron beam 1 for V _2B , M _B0 is also determined from the same curve, _and the 2 When D _hB0 (x,y) is determined from the directional blur curve, M _B0 (x,y) is also determined from the same curve.
Thus, D _hA0 and M _A0 are determined and D _hA0 (x, y) and M _A0 (x, y) are determined, or D _hB0 and M _B0 are determined and D _hB0 (x , y) and M _B0 (x, y) are determined by determining D _A0 and M _A0 in Example 2 and determining D _A0 (x, y) and M _A0 (x, y). is essentially the same as

Of the XY averages M _A0 , M _B0 , M _A0 (x, y), and M _B0 (x, y), M _B0 and M _B0 (x, y) may be determined as follows. Determining M _{B0 and M B0 (x, y) in the following manner is essentially the same as determining M B0 and M B0 ( x} , y) in Example 2. However, here, it is assumed that the formulas (114) and (116) hold.
First, the XY average M _B0 is determined from the blur curve in one direction (X direction or Y direction) of the electron beam 1 with respect to V _2B . Here, it is assumed that D _hA0 is determined before M _B0 is determined.
Next, the XY average M _B0 (x, y) is determined from the blur curve in one direction (X direction or Y direction) of the electron beam 1 with respect to V _2B on each grid point. The assumption here is that D _hA0 (x,y) is determined before M _B0 (x,y) is determined.

In the above manner, M _B0 is determined from equation (130) or (131) and substituted into equation (25). M _B0 (x,y) is determined from equation (132) or (133) and substituted into equation (41). Equation (130) is derived from equations (94) and (123). Equation (131) is derived from equations (95) and (123). Equation (132) is derived from equations (97) and (125). Equation (133) is derived from equations (98) and (125).
M _B0 =V _2BX0 -I _D0 d _hIB (130)
M _B0 =V _2BY0 +I _D0 d _hIB (131)
M _B0 (x, y)=V _2BX0 (x, y)−V _D0 (x, y)d _hVB (132)
M _B0 (x, y)=V _2BY0 (x, y)+V _D0 (x, y)d _hVB (133)

In equations (130) and (131), _ID0 is a measure of the axial defocus. I _D0 can be determined from coefficients d _hIA and D _hA0 as can be seen from equation (118). (D _hA0 has been previously determined, together with M _A0 .) I _D0 d _hIB (=D _hB0 ) is I _D0 converted to D _hB0 by the factor d _hIB (see equation (119)). . (I _D0 d _hIB is also D _hA0 converted to D _hB0 . As can be seen from equation (118), I _D0 is D _hA0 converted to the amount of change in the focus correction current by the coefficient d _hIA . by being
In equations (132) and (133), V _D0 (x, y) is the measured value of the deflected field curvature aberration. V _D0 (x, y) can be determined from coefficients d _hVA and D _hA0 (x, y) as can be seen from equation (120). (D _hA0 (x,y) has been previously determined together with M _A0 (x,y).) V _D0 (x,y)d _hVB (=D _hB0 (x,y)) is the coefficient d V _D0 (x, y) converted to D _hB0 (x, y) by _hVB (see equation (121)). (V _D0 (x, y)d _hVB is also D _hA0 (x, y) reduced to D _hB0 (x, y), which is V _D0 (x, y) is D _hA0 (x, y) converted into the amount of change in the focus correction voltage by the coefficient d _hVA .)
Therefore, equation (130) indicates that the XY average M _B0 can be determined from the first twice astigmatic voltage 51 (V _2BX0 ) and the measured value of the axial defocus converted to D _hB0 . showing. Equation (131) shows that M _B0 can be determined from the second twice astigmatic voltage 52 (V _2BY0 ) and the measured value of the axial defocus converted to D _hB0 . Equation (132) is obtained from the first two-time astigmatic voltage 51 (V _2BX0 (x, y)) and the measurement value of the deflection curvature of field converted into D _hB0 (x, y), It shows that the XY average M _B0 (x,y) can be determined. Equation (133) is derived from the second two-time astigmatic voltage 52 (V _2BY0 (x, y)) and the measurement value of the deflection curvature of field converted into D _hB0 (x, y), It shows that M _B0 (x,y) can be determined.

The coefficient d _hIB is determined and stored prior to the conversion of I _D0 to D _hB0 by d _hIB (see equation (119)). The d _hIB so stored is used repeatedly for each conversion of I _D0 to D _hB0 by d _hIB . The coefficient d _hVB is determined and stored prior to conversion of V _D0 (x,y) to D _hB0 (x,y) by d _hVB (see equation (121)). The d _hVB thus stored is used repeatedly each time the d _{hVB converts} V _D0 (x,y) to D _hB0 (x,y).

As described above, the apparatus of this embodiment, like the apparatus of the first embodiment, has its own axial defocus, axial two-fold astigmatism, deflection curvature of field aberration, and deflection two-fold non-aberration. Astigmatism is both measured and corrected based on the blur curve of the electron beam 1 versus twice astigmatic voltage.
However, the apparatus of this embodiment determines D _hA or D _hB instead of the XY difference D _A or D _B for the measurement of axial defocus and deflection field curvature aberration produced by itself, and D _hA or Defocus on the knife edge 20 is denoted by D _hB . Here, _DhA and _DhB are both the difference between the first two-time astigmatic voltage 51 and the XY average. D _hA and D _hB may alternatively both be the difference between the second twice astigmatic voltage 52 and the XY average.

Supplementally, in this embodiment, all of the axial defocus, axial 2-fold astigmatism, deflection field curvature aberration, and deflection 2-fold astigmatism were measured and corrected, but it depends on the purpose. But it doesn't have to be. For example, one or both of the axial defocus and deflection field curvature aberration may be measured and corrected. Alternatively, only the axial defocus and axial 2-fold astigmatism may be measured and corrected, or the deflection field curvature and deflection 2-fold astigmatism may be measured and corrected.

(Example 4)
Still another embodiment (variably shaped electron beam writing apparatus) of the charged particle beam apparatus of the present invention will be described below as Embodiment 4. FIG.

The apparatus of this embodiment (variably shaped electron beam writing apparatus) has the same configuration as the apparatuses of the first to third embodiments. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this example are the same as the configurations of the optical system, the measurement system, and the control system of the apparatus of Example 1 (see FIG. 1). .

The device of this embodiment basically operates in the same way as the device of the first embodiment. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.

However, the apparatus of this embodiment can detect the magnitude of defocus on the knife edge 20 by the XY difference (difference between the first two-fold astigmatism voltage 51 and the second two-time astigmatism voltage 52). It is expressed by the difference between the first two-fold astigmatism voltage 51 and the 0th two-fold astigmatism voltage 50 . (The 0th double astigmatism voltage 50 is V _2A before increasing or decreasing V _2A to obtain a curve of electron beam 1 blur against double astigmatism voltage V _2A , or double astigmatism means V _2B before increasing or decreasing V _2B to obtain a curve of electron beam 1 blur against correction voltage V _2B ).

From this, the curves obtained by the apparatus of this embodiment for measuring the defocus on the knife edge 20 are the X- or Y-direction blur curves of the electron beam 1 against V _2A , or the electron beam 1 against V _2B . 4 is the blur curve in the X or Y direction for beam 1; That is, the total number of curves acquired at one time for the measurement is one.
On the other hand, the curves obtained by the apparatus of Example 1 for the measurement of defocus on the knife edge 20 are the X and Y blur curves of electron beam 1 for V _2A , or electron beam 1 for V _2B . was the blur curve in the X and Y directions of . That is, the total number of curves acquired at one time for the measurement was two.

Supplementally, the direction in which the blur of the electron beam 1 is measured for obtaining the difference between the first two-fold astigmatic voltage 51 and the zeroth two-fold astigmatic voltage 50 is either the X direction or the Y direction. It doesn't have to be. The direction in which the blurring of the electron beam 1 is measured can therefore be rotated by .pi./4 with respect to the X or Y direction, for example. (However, if the direction in which the blurring of the electron beam 1 is measured is so rotated, the edge directions of the knife edges 20X and 20Y are rotated by π/4 with respect to the Y and X directions, respectively. There is a need.)
In other words, the direction in which the blurring of the electron beam 1 is measured can be arbitrarily determined and defined again as the X direction or the Y direction.

The difference between the first two-fold astigmatic voltage 51 and the zeroth two-fold astigmatic voltage 50 is, by definition, the first two-fold astigmatic voltage 51 and Close to the difference from the XY average. Here, the XY average is the average of the first double astigmatic voltage 51 and the second double astigmatic voltage 52 . The difference between the first two-time astigmatic voltage 51 and the XY average is D _hA or D _hB in the third embodiment.
Therefore, the manner in which the apparatus of this embodiment measures and corrects the axial defocus and deflection field curvature aberrations is similar to the manner in which the apparatus of Example 3 measures and corrects the same aberrations, and essentially are the same.
Hereinafter, the difference between the first two-fold astigmatism voltage 51 and the 0th two-fold astigmatism voltage 50 is defined as the difference between the first two-fold _astigmatism voltage 51 and the XY average. or _DhB .

That is, also in this embodiment, the defocus on the knife edge 20 (including either or both of the axial defocus and deflection field curvature aberration) is represented by D _hA or D _hB . Here, D _hA or D _hB is, more specifically, one of D _hA and D _hB that is not zero, more preferably one of D _hA and D _hB that is not smaller in absolute value be.
More specifically, the apparatus of this embodiment updates the focus correction current I _0A0 according to the equation (110a) to make D _hA zero, or updates the focus correction current I _0A0 to make D _hB zero. is updated by equation (111a). D _hA or D _hB is then converted to I _D0 by a factor d _hIA or d _hIB , respectively (see equations (118a) or (119a)).
The apparatus of this embodiment determines and stores the coefficient d _hIA prior to the conversion of D _hA0 to I _D0 by d _hIA (see equation (118a)). The d _hIA thus stored is used repeatedly each time the d _{hIA converts} D _hA0 to I _D0 . Alternatively, the apparatus of this embodiment determines and stores the coefficient d _hIB prior to the conversion of D _hB0 to I _D0 by d _hIB (see equation (119a)). The d _hIB so stored is used repeatedly for each conversion of D _hB0 to I _D0 by d _hIB . where d _hIA and d _hIB are stored when equations (114) and (115) hold, respectively. The manner in which the coefficients d _hIA and d _hIB are determined in this embodiment is due to the difference in definition of D _hA and D _hB difference, or the difference between the first two-fold astigmatism voltage 51 and the XY average), the manner in which d _hIA and d _hIB are determined in Example 3 is the same, respectively.
The apparatus of this embodiment further _{updates the focus correction voltage V 0A0 (} x, y) according to the equation (112a) or D _hB (x, y ) is made zero, the focus correction voltage V _0A0 (x, y) is updated by the equation (113a). Then D _hA (x, y) or D _hB (x, y) is converted to V _D0 (x, y) by coefficients d _hVA or d _hVB , respectively (using equation (120a) or (121a) as reference).
The apparatus of this embodiment determines and stores the coefficient d _{hVA prior} to conversion of D _hA0 (x,y) to V _D0 (x,y) by d _hVA (see equation (120a)). The d _hVA thus stored is used repeatedly each time the d _{hVA converts} D _hA0 (x,y) to V _D0 (x,y). Alternatively, the apparatus of this embodiment determines and stores the coefficient d _{hVB prior} to conversion of D _hB0 (x,y) to V _D0 (x,y) by d _hVB (see equation (121a)). do. The d _hVB thus stored is used repeatedly each time the d _{hVB converts} D _hB0 (x,y) to V _D0 (x,y). where _{d_hVA} and _{d_hVB} are stored when equations (116) and (117) hold, respectively. The manner in which coefficients d _hVA and d _hVB are determined in this example is the same as the manner in which d _hVA and d _hVB are determined in Example 3, except for the difference in definitions of D _hA and D _hB . .

However, strictly speaking, the difference between the first two-fold astigmatic voltage 51 and the 0th two-fold astigmatic voltage 50 is the difference between the first two-fold astigmatic voltage 51 and the XY average. ,different. More specifically, these differences are equal to each other when the two-fold astigmatism appearing on the knife edge 20 is zero (T _C =0), but as the two-fold astigmatism increases, deviate from each other.
Therefore, if the two-fold astigmatism increases, the measurement errors of the axial defocus and deflection curvature of field aberration increase, and along with that, the correction residual error of the axial defocus and deflection curvature of field aberration increases. growing.

For this reason, in this embodiment, in order to reduce the correction residual error, the axial two-fold astigmatism is measured and corrected before the axial defocus is measured and corrected, and the deflection image plane It is advisable to measure and correct the deflection 2-fold astigmatism before measuring and correcting the curvature aberration. (Hereinafter, a description will be given of reduction of correction residuals for the axial two-fold astigmatism and deflection two-fold astigmatism, and a two-fold astigmatism correction voltage V _2A0 ^{(m )} (x, y) and V _2B0 ^(m) (x, y) are omitted.)

Here, the difference between the first two-fold astigmatism voltage 51 and the 0th two-fold astigmatism voltage 50 can be matched to the difference between the first two-fold astigmatism voltage 51 and the XY average. ,Desired. That is, it is required to match the 0th double astigmatic voltage 50 with the XY average. More specifically, premised on the establishment of equations (114) and (116), the following two things are required.
First, the on-axis two-fold astigmatism described above is corrected by V _2A before obtaining the curve of blurring in one direction (X direction or Y direction) of the electron beam 1 versus V _2A to determine D _hA0 . Possible components are measured and corrected. That is, before obtaining the curve, the XY average M _A0 is determined, V _2A0 is updated by equation (24), and the updated V _2A0 is input to the objective deflector 13 . The component of the axial two-fold astigmatism correctable by _{V2B may be measured and corrected before or after determining DhA0 .}
Second, before obtaining the blur curve in one direction (X direction or Y direction) of the electron beam 1 versus V _2A to determine D _hA0 (x,y), the deflection two-fold astigmatism, The components correctable by _V2A are measured and corrected. That is, before obtaining the curve, determine the XY average M _A0 (x,y) and update V _2A0 (x,y) according to equation (40), so that the updated V _2A0 (x , y) are input to the objective deflector 13 . This process is performed on each grid point within the deflection field of the objective deflector 13 . The component of the deflection 2-fold astigmatism correctable by V _2B may be measured and corrected before or after determining D _hA0 (x,y).

As far as these things are concerned, the manner in which the apparatus of this example measures and corrects the axial two-fold astigmatism and the deflection two-fold astigmatism is similar to that the apparatus of Example 1 measures the same aberrations. , and the correction procedure may be the same. That is, the determination of the XY average and the update and input of the two-time astigmatism correction voltage may be performed in the following manner.
First, the XY average M _A0 is determined from the blur curves in the two directions (X direction and Y direction) of the electron beam 1 with respect to the double astigmatism correction voltage V _2A , and V _2A0 is updated by the equation (24). V _2A0 updated as above is input to the objective deflector 13 . Furthermore, the XY average M _B0 is determined from the curves of blurring in the two directions (X direction and Y direction) of the electron beam 1 with respect to the double astigmatism correction voltage V _2B , and V _2B0 is updated by the equation (25). V _2B0 updated as above is input to the objective deflector 13 .
As a result, the component of the axial two-fold astigmatism that can be corrected by _V2A is corrected, and the component of the axial two-fold astigmatism that can be corrected by _V2B is corrected.
Next, steps similar to the series of steps described above are performed on each lattice point within the deflection field of the objective deflector 13 . That is, on each grid point, the XY average M _A0 (x, y) is determined from the curves of blurring in the two directions (X direction and Y direction) of the electron beam 1 with respect to V _2A , and V _2A0 (x, y) is updated, and the updated V _2A0 (x, y) is input to the objective deflector 13 . On each grid point, the XY average M _B0 (x, y) is further determined from the blur curves in the two directions (X direction and Y direction) of the electron beam 1 with respect to V _2B , and V _2B0 (x, y) and then input the updated V _2B0 (x, y) to the objective deflector 13 .
As a result, the component of the 2-fold deflection astigmatism that can be corrected by V2A is corrected on each lattice point, and the component of the ₂ -fold deflection astigmatism that can be corrected by _V2B is corrected. be done.

Of the XY averages M _A0 , M _B0 , M _A0 (x, y), and M _B0 (x, y), M _B0 and M _B0 (x, y) may be determined as follows. Determining M _{B0 and M B0 (x, y) in the following manner is essentially the same as determining M B0 and M B0 ( x} , y) in Example 2. However, here, it is assumed that the formulas (114) and (116) hold.
First, the XY average M _B0 is determined from the blur curve in one direction (X direction or Y direction) of the electron beam 1 with respect to V _2B . Here, it is assumed that D _hA0 is determined before M _B0 is determined.
Next, the XY average M _B0 (x, y) is determined from the blur curve in one direction (X direction or Y direction) of the electron beam 1 with respect to V _2B on each grid point. The assumption here is that D _hA0 (x,y) is determined before M _B0 (x,y) is determined.

In the above manner, the XY average M _B0 is determined from equation (130) or (131). That is, M _B0 is the first twice astigmatic voltage 51 (V _2BX0 ) or the second twice astigmatic voltage 52 (V _2BY0 ) and the measurement of the on-axis defocus converted to D _hB0 . can be determined from the value Here, the above measured on-axis defocus converted to D _hB0 is also D _hA0 converted to D _hB0 . M _B0 determined from equation (130) or (131) is substituted into equation (25).
The XY average M _B0 (x,y) is determined from equation (132) or (133). That is, M _B0 (x, y) is the first two-fold astigmatic voltage 51 (V _2BX0 (x, y)) or the second two-fold astigmatic voltage 52 (V _2BY0 (x, y)). and the measured deflection field curvature aberration converted to D _hB0 (x,y). Here, the measured deflection field curvature aberration converted to D _hB0 (x, y) is also D _hA0 (x, y) converted to D _hB0 (x, y). M _B0 (x, y) determined from equation (132) or (133) is substituted into equation (41).
The coefficient d _hIB in equations (130) and (131) is determined and stored prior to the conversion of I _D0 to D _hB0 by d _hIB (see equation (119)). The d _hIB so stored is used repeatedly for each conversion of I _D0 to D _hB0 by d _hIB . The coefficient d _hVB in equations (132) and (133) is determined prior to the conversion of V _D0 (x,y) to D _hB0 (x,y) by d _hVB (see equation (121)), remembered. The d _hVB thus stored is used repeatedly each time the d _{hVB converts} V _D0 (x,y) to D _hB0 (x,y).

As described above, the apparatus of this embodiment, like the apparatus of the first embodiment, has its own axial defocus, axial two-fold astigmatism, deflection curvature of field aberration, and deflection two-fold non-aberration. Astigmatism is both measured and corrected based on the blur curve of the electron beam 1 versus twice astigmatic voltage.
However, the apparatus of this embodiment determines D _hA or D _hB for the measurement of its own axial defocus and deflection field curvature aberration, and D _hA or D _hB determines the defocus on the knife edge 20 represents Here, D _hA and D _hB are both the difference between the first two-fold astigmatic voltage 51 and the zeroth two-fold astigmatic voltage 50 . Alternatively, D _hA and D _hB may both be the difference between the second two-fold astigmatic voltage 52 and the zeroth two-fold astigmatic voltage 50 .

Supplementally, in this embodiment, all of the axial defocus, axial 2-fold astigmatism, deflection field curvature aberration, and deflection 2-fold astigmatism were measured and corrected, but it depends on the purpose. But it doesn't have to be. For example, one or both of the axial defocus and deflection field curvature aberration may be measured and corrected. Alternatively, only the axial defocus and axial 2-fold astigmatism may be measured and corrected, or the deflection field curvature and deflection 2-fold astigmatism may be measured and corrected.

For convenience of explanation, in the following examples (Examples 5-12), D _hA and D _hB are considered essentially the same as the XY differences D _A and D _B , respectively. More specifically, in the examples that follow, D _hA and D _hB are included in D _A and D _B , respectively. In the examples that follow, the coefficients d _hIA , d _hIB , d _hVA , and d _hVB are also included in the coefficients d _IA , d _IB , d _VA , and d _VB , respectively.

(Example 5)
Still another embodiment (variably shaped electron beam writing apparatus) of the charged particle beam apparatus of the present invention will be described below as Embodiment 5. FIG.

The apparatus of this embodiment (variably shaped electron beam writing apparatus) has the same configuration as the apparatuses of the first to fourth embodiments. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this example are the same as the configurations of the optical system, the measurement system, and the control system of the apparatus of Example 1 (see FIG. 1). .

The apparatus of this embodiment basically operates in the same way as the apparatuses of the first to fourth embodiments. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.

However, the apparatus of this embodiment, when measuring the XY difference D _A or D _B , compensates for the secondary defocus that occurs with the increase and decrease of the two-time astigmatism correction voltages _V2A and _V2B . Cancel by generating a new defocus. By doing so, the device of this embodiment reduces the measurement error of D _A or D _B .

The above secondary defocus is caused by the 0th-order potential component around each principal ray included in the electron beam 1 . The zero-order potential component is derived from the second-order potential component around the central axis of the objective deflector 13 .
The secondary potential component is a potential component generated within the objective deflector 13 due to the double astigmatism correction voltage, and is uniform within the objective deflector 13 . Therefore, the secondary potential component does not depend on the positional deviation of each principal ray from the central axis of the objective deflector 13 .
On the other hand, the 0th-order potential component is one of the low-order potential components (see Patent Document 2) obtained by polynomial expansion of the potential around each principal ray in the presence of the second-order potential component (more specifically, is the lowest order component). These low-order potential components depend on the displacement of the principal rays from the central axis of the objective deflector 13 . As such positional deviation, what is important in this embodiment is the positional deviation common to the principal rays. The positional deviation is the overall positional deviation of all principal rays contained in the electron beam 1 with respect to the central axis of the objective deflector 13 .

If the secondary defocus occurs during the measurement of the XY difference D _A or D _B , what is represented by D _A or D _B is the difference between the desired defocus and the secondary defocus. become peace. Here, the intended defocus corresponds to the true measurement of D _A or D _B , while the secondary defocus corresponds to the measurement error of D _A or D _B.

Cancellation of the secondary defocus is performed in a manner similar to that described in Patent Document 2. Here, the point similar to the point described in Patent Document 2 is, more specifically, the point of canceling secondary two-fold astigmatism that occurs as the three-fold astigmatism correction voltage increases and decreases. This corresponds to replacing the astigmatism correction voltage and the double astigmatism correction voltage with the double astigmatism correction voltage and the focus correction voltage, respectively.
That is, the secondary defocus cancellation is performed by adding to the focus correction voltage _V0A an addition value given by a linear combination of the variation amounts of the two-time astigmatism correction voltages _V2A and _V2B . The added value is the sum of the product of the coefficient ρ _0A2A and the variation of the two-fold astigmatism correction voltage V _2A and the product of the coefficient ρ _0A2B and the variation of the two-fold astigmatism correction voltage V _2B . If the added value is added to the focus correction voltage _V0A , the new defocus occurs, which cancels the secondary defocus.
Here, the coefficient ρ _0A2A is a partial differential coefficient with respect to V _2A of the focus correction voltage V _0A for canceling secondary defocus that occurs as the twice astigmatism correction voltage V _2A increases and decreases. The coefficient ρ _0A2B is a partial differential coefficient with respect to V _2B of the focus correction voltage V _0A for canceling secondary defocus that occurs as the two-time astigmatism correction voltage V _2B increases or decreases.
Both the coefficients ρ _0A2A and ρ _0A2B depend on the deflection coordinates (x,y). This is because the zero-order potential component depends on the positional deviation of each principal ray from the central axis of the objective deflector 13, as described above. Therefore, ρ _0A2A and ρ _0A2B must be determined for each lattice point within the deflection field of the objective deflector 13 when measuring the deflection field curvature aberration.

(Example 6)
Still another embodiment (variably shaped electron beam writing apparatus) of the charged particle beam apparatus of the present invention will be described below as Embodiment 6. FIG.

The apparatus of this embodiment (variably shaped electron beam writing apparatus) has the same configuration as the apparatuses of the first to fourth embodiments. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this example are the same as the configurations of the optical system, the measurement system, and the control system of the apparatus of Example 1 (see FIG. 1). .

The apparatus of this embodiment basically operates in the same way as the apparatuses of the first to fourth embodiments. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.

However, the apparatus of this embodiment corrects both the axial defocus and the deflected curvature of field by updating the focus correction voltage V _{0A 0} and inputting V _{0A 0} to the objective deflector 13 .
On the other hand, in the devices of Examples 1 to 4, the axial defocus is corrected by updating the focus correction current _I0A0 and inputting the current _I0A0 to the objective lens 9, while the deflection curvature of field aberration was corrected by updating the focus correction voltage V _{0A 0} and inputting V _{0A 0} to the objective deflector 13 .

More specifically, the apparatus of this embodiment updates V _0A0 ^(m) (0, 0) according to equation (30a) or (31a) to correct the axial defocus, and then updates The obtained V _0A0 ^(m) (0, 0) is input to the objective deflector 13 . Here, V _0A0 ^(m) (0,0) means the focus correction voltage V _0A updated at the center of the deflection field of the objective deflector 13 . The apparatus of this embodiment further corrects the deflection field curvature aberration by V _0A0 ^(m) (x, y) on each grid point in the deflection field according to equation (30a) or (31a). is updated, and the thus updated V _0A0 ^(m) (x, y) is input to the objective deflector 13 . (After that, an approximate curved surface V _0A0F (x, y) is determined from V _0A0 ^(m) (x, y) on all lattice points in the deflection field, and thereafter based on V _0A0F (x, y), the above The deflection field curvature aberration is corrected over the entire deflection field.)

Therefore, in this embodiment, the coefficient _dVA or _dVB can be used for both the measurement and correction of the axial defocus and the measurement and correction of the deflection curvature of field.

However, it is allowed to correct the axial defocus by updating the focus correction voltage _V0A0 and inputting _V0A0 to the objective deflector 13 when the magnitude of the axial defocus is relatively small. In. This is because the focus correction sensitivity of the objective deflector 13 as an electrostatic focus corrector is low as described above.
Here, when the magnitude of the axial defocus is relatively small, for example, after the axial defocus is corrected by updating the focus correction current _I0A0 and inputting _I0A0 to the objective lens 9, This is the case when the on-axis defocus exhibits a small change over time.

Supplementally, in principle, both the correction of the axial defocus and the correction of the deflection curvature of field can be performed by a magnetic focus corrector (not shown) or the objective lens 9 . Here, the magnetic focus corrector is simply a small weak magnetic field lens. Furthermore, both the correction of the axial two-fold astigmatism and the correction of the deflection two-fold astigmatism can be performed by a magnetic field type two-fold astigmatism corrector (not shown). That is, it is possible to correct all of the axial defocus, deflection curvature of field aberration, axial 2-fold astigmatism, and deflection 2-fold astigmatism by the magnetic field corrector. However, magnetic compensators generally operate slower than electrostatic compensators.
If all these aberrations are corrected by the magnetic field corrector, the electron beam 1 versus the two-fold astigmatism current is used instead of the blur curve of the electron beam 1 versus the two-fold astigmatism voltage when measuring these aberrations. A blur curve of 1 is obtained. That is, instead of the double astigmatism correction voltage, the double astigmatism correction current is increased or decreased. Therefore, in that case, the XY differences D _A and D _B have current dimensions instead of voltage dimensions. In that case, furthermore, the partial derivative of the XY difference D _A or D _B thus measured with respect to the focus correction current I _0A is determined and stored instead of the coefficients d _VA or d _VB respectively, and used repeatedly.

(Example 7)
Still another embodiment (variably shaped electron beam writing apparatus) of the charged particle beam apparatus of the present invention will be described below as Embodiment 7. FIG.

The apparatus of this embodiment (variably shaped electron beam writing apparatus) has the same configuration as the apparatuses of the first to fourth embodiments. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this example are the same as the configurations of the optical system, the measurement system, and the control system of the apparatus of Example 1 (see FIG. 1). .

The apparatus of this embodiment basically operates in the same way as the apparatuses of the first to fourth embodiments. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.

However, the apparatus of this embodiment measures and corrects the deflection curvature of field in the following manner.
First, at each grid point in the deflection field of the objective deflector 13, the curve of the blurring of the electron beam 1 with respect to the twice astigmatism correction voltage _V2A is obtained, the XY difference D _A0 is determined from the curve, and ( 30) Update V _0A0 (x,y) (ie, determine V _0A0 ′(x,y)) by equation. Alternatively, at each of the grid points, a curve of blurring of the electron beam 1 with respect to the double astigmatism correction voltage _V2B is obtained, the XY difference D _B0 is determined from the curve, and V _0A0 (x , y) (ie, determine V _0A0 ′(x,y)).
Next, an approximate curved surface V _0A0F (x, y) is determined from V _0A0 (x, y)' at all of the lattice points.
Then, V _0A0F (x,y) thus determined is updated. _{Specifically, V 0A0F ( x} , y) and the vertical axis Gives a shift in direction. Here, the value of V _0A0F (x, y) at any deflection coordinate (x, y) falls within the variable range of V _0A means that the maximum and minimum values of V _0A0F (x, y) are It refers to being below the upper limit and above the lower limit of the variable range of _V0A . Here, however, V _0A0F (x, y) is a function defined only within the deflection field of the objective deflector 13 .
Next, the new defocus generated at the height position of the material 10 (and the knife edge 20) due to the shift is eliminated. Its cancellation is due to the updating of the focus correction current _I0A0 and its input to the objective lens 9. FIG. The details will be described later.
Next, under the condition that the focus correction voltage V _0A based on V _0A0F (x, y) updated in the above process is input to the objective deflector 13, at each of the grid points, the electron beam 1 for V _2A A series of steps from acquisition of the blur curve to input of V _0A0 (x, y) updated by the equation (30a) to the objective deflector 13 are repeated until the equation (36) is established. Alternatively, under the same conditions, at each of the grid points above, from the acquisition of the blur curve of the electron beam 1 against V _2B , V _0A0 (x,y) updated by equation (31a) to the objective deflector 13 A series of steps up to the input of are repeated until the formula (37) holds.
Next, V _0A0F (x, y) is determined from V _0A0 ( ^m) (x, y) at all of the lattice points. That is, V _0A0F (x,y) is updated again.
Thereafter, based on V _0A0F (x, y) updated in this way, the deflected curvature of field aberration is corrected over the entire deflection field of the objective deflector 13 .

In the manner described above, it is important that the approximate curved surface V _0A0F (x, y) is given the above shift. If the above shift is not given to V _0A0F (x, y), the condition that the difference between the maximum value and the minimum value of V _0A0F (x, y) is equal to or less than the width of the variable range of the focus correction voltage V _0A is satisfied. However, it is possible that the value at some or all of the deflection coordinates (x,y) of V _0A0F (x,y) falls outside the variable range of V _0A . This involves the magnitude of the difference, as well as the initial value of V _0A (ie, V _0A0 (x,y)).

More specifically, the above shift includes a margin from the maximum value of the approximate curved surface V _0A0F (x, y) to the upper limit of the variable range of the focus correction voltage V _0A and the minimum value of V _0A0F (x, y) to V It is a vertical, uniform shift to equalize the margin to the lower end of the variable range of _0A .
To give such a shift to V _0A0F (x, y), the average of the maximum and minimum values of the approximate curved surface V _0A0F (x, y), the average of the upper and lower limits of the variable range of the focus correction voltage V _0A from V _0A0F (x,y).

Alternatively, the approximate surface V _0A0F (x,y) may be given a different shift. Here, another shift is, for example, a shift that makes the maximum value or minimum value of V _0A0F (x, y) coincide with the upper limit or lower limit of the variable range of the focus correction voltage V _0A , respectively.

If the approximate curved surface V _0A0F (x, y) is shifted as in the two examples described above, the difference between the maximum and minimum values of V _0A0F (x, y) is equal to or less than the width of the variable range of the focus correction voltage V _0A . As long as the above _condition that _.theta .

However, if some shift is given to the approximate curved surface V _0A0F (x, y), the above new defocus will occur as long as the magnitude is not zero. That is, it is necessary to eliminate the new defocus.

The new defocus cancellation is based on one of the following two methods.
First, the new defocus is first measured by the null method. That is, the blurring of the electron beam 1 is measured by the knife edge 20 while increasing or decreasing the focus correction current _I0A , and the _I0A that minimizes the blurring is determined. Next, I _0A thus determined is input to the objective lens 9 as an updated value of I _0A0 .
Second, first, obtain a curve of the blur of the electron beam 1 with respect to the double astigmatism correction voltage V _2A (or V _2B ), and determine the XY difference D _A0 (or D _B0 ) from the curve. Next, update I _0A0 according to equation (9a) (or (10a)) and input the updated I _0A0 to the objective lens 9 .

Supplementally, when the difference between the maximum value and the minimum value of V _0A0F (x, y) is sufficiently larger than the width of the variable range of the focus correction voltage V _0A , V _0A0 When updating (x, y) (that is, determining V _0A0 ′ (x, y)) and determining the approximate surface V _0A0F (x, y) from V _0A0 (x, y)′, V _0A0F ( x, y) need not be highly accurate. That is, in that case, even if the approximation accuracy of V _0A0F (x, y) is not very high, the occurrence of the above-described undesirable situation can be sufficiently prevented.
Therefore, in the above case, to update V _0A0 (x, y) by equation (30), obtain the curve of the blur of the electron beam 1 against the double astigmatic voltage V _2A , or by equation (31), obtain V The number of grid points can be reduced when acquiring the blur curve of the electron beam 1 with respect to the double astigmatism correction voltage _V2B to update _0A0 (x,y). If the number of lattice points is reduced, the time required for measuring and correcting the deflection curvature of field aberration can be shortened.
The number of grid points depends on the order and symmetry of V _0A0F (x,y). For example, if V _0A0F (x, y) is limited to a paraboloid of revolution having a minimum value at the center of the deflection field, the number of lattice points should be at least two. Specifically, these two points may be the central point and the corner point of the deflection field.

(Example 8)
Still another embodiment (variably shaped electron beam writing apparatus) of the charged particle beam apparatus of the present invention will be described below as an eighth embodiment.

The apparatus of this embodiment (variably shaped electron beam writing apparatus) has the same configuration as the apparatuses of the first to fourth embodiments. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this example are the same as the configurations of the optical system, the measurement system, and the control system of the apparatus of Example 1 (see FIG. 1). .

The apparatus of this embodiment basically operates in the same way as the apparatuses of the first to fourth embodiments. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.

The apparatus of this embodiment then corrects the deflection distortion caused by the focus correction voltage _V0A . The deflection distortion aberration is caused by changes in the deflection sensitivity of the objective deflector 13 due to changes in the focus correction voltage. The apparatus of this embodiment improves its own drawing accuracy (positional accuracy) by correcting the deflection distortion aberration.

However, in the apparatus of this embodiment, the deflection distortion is corrected in the following manner. First, the deflection distortion aberration is predicted from the distribution function of the focus correction voltage _V0A . Next, the value of the distortion correction voltage is determined for each deflection coordinate so as to cancel the deflection distortion aberration thus predicted. Then, the biased distortion correction voltage thus determined is input to the objective deflector 13 according to the deflection coordinates (added to the deflection signal).
According to this manner of correction of the deflection distortion aberration, there are cases where the focus correction voltage _V0A is represented by a continuous distribution function with respect to the deflection coordinates, and where _V0A has a discontinuous distribution with respect to the deflection coordinates. The correction residual of the deflection distortion aberration can be reduced both in the case of the function and in the case of the function.

Here, when the focus correction voltage V _0A is represented by a continuous distribution function with respect to the deflection coordinates, for example, the approximate curved surface V _0A0F (x, y) is a polynomial ( 2nd order or higher). In such a case, the value of V _0A0F (x,y) is determined from the coefficients of the polynomial and the deflection coordinates (x,y), and the value of the focus correction voltage thus determined is the objective deflection input to the device 13.
On the other hand, when the focus correction voltage V _0A is represented by a discontinuous distribution function with respect to the deflection coordinates, for example, the deflection field of the objective deflector 13 has a large number of sections (each of which is several μm to several tens of μm square). ) and each of those partitions is assigned the value of V _0A0F (x,y) as a discrete value. In such a case, the value of each section is referenced according to the deflection coordinates, and the focus correction voltage of the referenced value is input to the objective deflector 13 . Therefore, in such a case, V _0A0F (x,y) will end up with a discrete (step-like discontinuity) with respect to the deflection coordinates (x,y) regardless of its original continuity. a) function. In other words, V _0A0F (x, y) contains many high-order components related to deflection coordinates.

When the focus correction voltage _V0A is represented by a continuous distribution function with respect to the deflection coordinates, the deflection distortion aberration also changes continuously with respect to the deflection coordinates, similarly to _V0A . Therefore, in that case, the deflection distortion aberration is measured in the deflection field together with the deflection distortion aberration not caused by the focus correction voltage _V0A by the null method, and the value of the distortion correction voltage obtained by doing so is Once the approximate curved surface is determined for the distribution, the value of the distortion correction voltage for each deflection coordinate can be determined with high accuracy from the approximate curved surface. That is, the approximation accuracy of the value of the distortion correction voltage by the approximate curved surface is high, and therefore the correction residual of the deflection distortion aberration is small. (Similarly to the deflection distortion aberration, the deflection distortion aberration not caused by the focus correction voltage _V0A also changes continuously with respect to the deflection coordinates. The correction residual of the deflection distortion aberration not caused by _V0A is also small, provided that each of the above sections is not assigned a discrete value of the correction signal other than the focus correction voltage. do.)
Here, measuring the deflection distortion aberration together with the deflection distortion aberration not caused by the focus correction voltage V _0A by the zero point method means that the focus correction voltage V _0A0F (x, y) is input to the objective deflector 13. , the deflection distortion aberration on the grid points in the deflection field of the objective deflector 13 is directly measured by the knife edge 20, and each grating is adjusted so that the measured deflection distortion aberration is canceled on each grid point. Refers to determining the value of the distortion correction voltage on the point. Here, it is important to confirm cancellation of the deflection distortion aberration on each lattice point. That is, it is important to repeat the measurement of the deflection distortion aberration until the correction residual of the deflection distortion aberration becomes sufficiently small on each grid point.

On the other hand, when the focus correction voltage _V0A is represented by a discontinuous distribution function with _respect to the deflection coordinates, the deflection distortion aberration is also discrete (discontinuous to) change. In other words, the distribution function of the deflection distortion aberration contains many high-order components related to deflection coordinates. (The deflection distortion aberration not caused by the focus correction voltage _V0A changes continuously with respect to the deflection coordinates even in this case.)
In this case, the deflection distortion aberration is measured in the deflection field together with the deflection distortion aberration not caused by the focus correction voltage _V0A by the null method, and the distribution of the distortion correction voltage values obtained by doing so is Determining an approximate curved surface is not a good idea. This is because the high-order components that the distribution function of the distortion correction voltage should have are lost when such an approximate curved surface is determined. That is, the approximation accuracy of the value of the distortion correction voltage by the approximate curved surface is low, and therefore the correction residual of the deflection distortion aberration is large.

When the focus correction voltage V _0A is represented by a discontinuous distribution function with respect to the deflection coordinates, first, the deflection distortion aberration is expressed as V _0A0F (x, y) (or V _0A0 ^(m) (x, y )), and then determine the value of the distortion correction voltage for each deflection coordinate such that the deflection distortion aberration thus predicted is canceled. Then, the value of the distortion correction voltage will faithfully reflect the higher-order components of V _0A0F (x, y) (or V _0A0 ^(m) (x, y)). Aberration correction residuals become smaller.

The apparatus of this embodiment predicts the deflection distortion aberration and determines the value of the distortion correction voltage according to a procedure similar to that described in Japanese Patent Application Laid-Open No. 2002-200032. Here, the point similar to the point described in Patent Document 2 is, more specifically, n-times astigmatism correction voltage in the point of canceling deflection distortion aberration derived from n (≧2) times astigmatism correction voltage. , is replaced by the focus correction voltage.
That is, in this embodiment, the product of the coefficient ρ _1A0A and the approximate curved surface V _0A0F (x, y) is added to the A component of the distortion correction voltage, and the coefficient ρ _1B0A and the approximate curved surface are added to the B component of the distortion correction voltage. The product with V _0A0F (x, y) should be added. Then, where the values of V _0A0F (x,y) are assigned to each of the sections as discrete values, the value of the distortion correction voltage is determined for each of the sections.
Here, the A component and the B component of the distortion correction voltage mean the distortion correction voltage components added to the deflection voltages _V1A and _V1B , respectively. The coefficients ρ _1A0A and ρ _1B0A are partial differential coefficients of the A component and the B component of the distortion correction voltage for canceling the deflection distortion aberration caused by the focus correction voltage V _0A with respect to V _0A , respectively.

However, if the deflection distortion is corrected according to the above procedure (prediction of the deflection distortion and determination of the value of the distortion correction voltage), the larger the absolute value of V _0A0F (x, y) is, the more the deflection distortion becomes. Correction residuals can be large. This is because cancellation of the deflection distortion aberration is not confirmed in the above procedure.

If this is a problem, do the following:
First, the focus correction voltage V _0A0F (x, y) is divided into low-order components (for example, up to 2nd order) and high-order components (for example, 3rd-order or higher) regarding deflection coordinates, and the low-order and high-order components are divided into , V _0A0FL (x,y) and V _0A0FH (x,y), respectively. Hereinafter, the values of V _0A0FL (x,y) are determined from the coefficients of the polynomial representing V _0A0FL (x,y) and the deflection coordinates, while the values of V _0A0FH (x,y) are determined from V Let the discrete values of _0A0FH (x,y) be determined by looking up according to the deflection coordinates. where the discrete values of V _0A0FH (x,y) are once determined from the coefficients of the polynomial of V _0A0FH (x,y) and the deflection coordinates, then assigned to each of the above partitions and stored, and shall be referenced.
Next, V _0A0FL (x, y) is input to the objective deflector 13 according to the deflection coordinates, and the deflection distortion aberration appearing in the deflection field of the objective deflector 13 under this condition is calculated by the null method as An approximation curved surface is determined for the distribution of the distortion correction voltage values obtained by the measurement. The deflection distortion aberration appearing in the deflection field of the objective deflector 13 under this condition is the deflection distortion aberration caused by V _0A0FL (x, y), V _0A0FL (x, y) and V _0A0FH (x, It is the sum of deflection distortion aberration caused by none of y). Thereafter, when the electron beam 1 is deflected toward the deflection coordinates (x, y), the value of the distortion correction voltage corresponding to the deflection coordinates (x, y) is determined from the approximate curved surface. It is assumed that the distortion correction voltage having the value obtained is input to the objective deflector 13 .
Then, V _0A0FL (x, y) and V _0A0FH (x, y) are input to the objective deflector 13 according to the deflection coordinates, and the deflection distortion appearing in the deflection field under these conditions is corrected. do. The deflection distortion appearing in the deflection field under this condition is the deflection distortion caused by V _0A0FH (x, y). (At this point, the sum of the deflection distortion caused by V _0A0FL (x, y) and the deflection distortion caused by neither V _0A0FL (x, y) nor V _0A0FH (x, y) is already corrected.)
Correction of deflection distortion aberration caused by V _0A0FH (x, y) is performed by determining a value of a distortion correction voltage for each deflection coordinate so as to cancel deflection distortion aberration predicted from V _0A0FH (x, y), Then, the biased distortion correction voltage having the value thus determined is input to the objective deflector 13 according to the deflection coordinates. That is, the product of the coefficient ρ _1A0A and V _0A0FH (x, y) is added to the A component of the distortion correction voltage, and the coefficient ρ _1B0A and V _0A0FH (x, y) is added to the B component of the distortion correction voltage. is added.

In this way, the deflection distortion caused by V _0A0FL (x, y) can be accurately corrected together with the deflection distortion caused by neither V _0A0FL (x, y) nor V _0A0FH (x, y). Furthermore, the deflection distortion caused by V _0A0FH (x, y) is also corrected with high accuracy. That is, the correction residuals of all deflection distortion aberrations within the deflection field are reduced.
What is important here is that |V _0A0FH (x, y)| is smaller than |V _0A0F (x, y)| _. is smaller than the correction residual of the deflection distortion caused by V _0A0F (x, y).

(Example 9)
Still another embodiment (variably shaped electron beam writing apparatus) of the charged particle beam apparatus of the present invention will be described below as the ninth embodiment.

The apparatus of this embodiment (variably shaped electron beam writing apparatus) basically has the same configuration as the apparatuses of the first to fourth embodiments. That is, the configurations of the optical system, the measurement system, and the control system of the apparatus of this embodiment are respectively the configurations of the optical system, the measurement system, and the control system of the apparatus of Embodiment 1 (see FIG. 1), and basically is the same as

However, as shown in FIG. 21, the measurement system of the apparatus of this embodiment includes a knife edge 20' and a Faraday cup 21' as a knife edge and a Faraday cup separate from the knife edge 20 and the Faraday cup 21. The measurement system of the apparatus of this embodiment further includes an optical height detector (not shown, see Japanese Patent Laid-Open No. 02-241021, for example). The apparatus of this embodiment measures the height position of the surface on which the electron beam 1 is incident by its height detector. The height position is, more specifically, the height position of the region on each surface of the knife edge 20, the knife edge 20', and the material 10 on which the electron beam 1 is incident.

The apparatus of this embodiment basically operates in the same way as the apparatuses of the first to fourth embodiments. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.

The apparatus of the present embodiment then corrects the axial defocus for each region within the material 10 . In other words, the apparatus of the present embodiment reduces the axial defocus correction residual for each region within the material 10 .
The correction residual is caused by the height position of the material 10 not matching the height position of the knife edge 20 . Here, the fact that the height position of the material 10 does not match the height position of the knife edge 20 specifically means, firstly, that there is a height difference in the material 10, and secondly, that the material 10 has a height difference. This means that the height position of the material 10 does not match the height position of the knife edge 20, although there is no height difference.
On the other hand, the apparatuses of Examples 1 to 4 did not correct the axial defocus for each region within the material 10 . That is, in Examples 1 to 4, the difference between the height position of the material 10 and the height position of the knife edge 20 was zero, or even if the difference was not zero, the difference was irrelevant. was done.

The apparatus of this embodiment also corrects dimensional errors and rotation of the deflection field of the objective deflector 13 for each region within the material 10 . The dimensional error and rotation are also caused by the fact that the height position of the material 10 does not match the height position of the knife edge 20, similar to the correction residual.

The manner in which the apparatus of this embodiment corrects the axial defocus and the dimensional error and rotation of the deflection field of the objective deflector 13 for each region within the material 10 will be described below.

Prior to drawing on the material 10, the apparatus of this embodiment first moves the knife edge 20' directly below the electron beam 1 by the material stage 14, and then detects the height of the knife edge 20' by the height detector. The knife edge 20' measures and corrects the defocus on the knife edge 20' while measuring the edge position. For the measurement and correction of the defocus, obtain the curve of blurring in the X direction and the Y direction of the electron beam 1 with respect to the double astigmatism correction voltage _{V2A, and update the focus correction current I0A0 according} to the equation (9a), or Or, obtain the curves of blurring in the X and Y directions of the electron beam 1 with respect to the twice astigmatism correction voltage _{V2B, and update the focus correction current I0A0 according} to equation (10a), and then update By inputting _I0A0 into the objective lens 9; What is important here is that the defocus on the knife edge (specifically, the knife edge 20') whose height position is different from that of the knife edge 20 is corrected.
In the apparatus of this embodiment, next, the knife edge 20 is moved directly below the electron beam 1 by the material stage 14, and then the height detector measures the height position of the knife edge 20. 20 measures the defocus on the knife edge 20 . (At this point, the defocus on the knife edge 20 is not corrected.) The defocus measurement obtains curves of the X and Y blurring of the electron beam 1 against the twice astigmatic voltage _V2A . , determine the XY difference D _A from those curves, or obtain curves of the blurring of the electron beam 1 in the X and Y directions for the twice astigmatism correction voltage V _2B and determine the XY difference D _B from those curves. By doing.
Here, it is not necessary to determine both D _A and D _B , and either one of D _A and D _B may be determined. More specifically, if the formula (19) (or (32)) holds, D _A should be determined, and if the formula (20) (or (33)) holds, D _B should be determined.

Supplementally, the correction of the defocus on the knife edge 20' is, in principle, performed by updating the focus correction current _I0A0 and inputting _I0A0 to the objective lens 9, updating the focus correction voltage _V0A0 and inputting _V0A0 . This is possible either with the input to the objective deflector 13 . This also applies to the region-by-region correction of the axial defocus within the material 10 in subsequent steps.
However, in this embodiment, both of these corrections are based on updating the focus correction current _I0A0 and inputting _I0A0 to the objective lens 9. FIG.

The apparatus of this embodiment then obtains the partial differential coefficient of D _A or D _B with respect to the height position of the surface on which the electron beam 1 is incident. These partial differential coefficients are obtained from the XY difference D _A or D _B obtained in the above process and the height positions of the knife edges 20 and 20'. Here, changes in the XY differences D _A and D _B due to changes in the focus correction current I _0A (updating I _0A0 according to equation (9a) or (10a)) are respectively It is considered the change in the XY differences D _A and D _B due to the change in height position.
_{Let dzA} and _dzB be the partial differential coefficients of D _A and D _B with respect to the height position of the plane on which the electron beam 1 is incident _. given by the formula

（１３４）および（１３５）式において、ｚは、電子ビーム１の入射する面の高さ位置を表す。ｚ_０およびｚ_１は、それぞれ、ナイフエッジ２０および２０’の高さ位置を表す。Ｄ_Ａ０およびＤ_Ａ１は、それぞれ、ナイフエッジ２０および２０’上で測定されるＤ_Ａを表す。Ｄ_Ｂ０およびＤ_Ｂ１は、それぞれ、ナイフエッジ２０および２０’上で測定されるＤ_Ｂを表す。Ｄ_Ａ１およびＤ_Ｂ１は、ナイフエッジ２０’上のデフォーカスが補正されていれば、いずれも零となる。Δｚは、ｚ_１のｚ_０からの差であり、例えば、１０μｍ程度の大きさを持つ。
ここでは、ｄ_ｚＡおよびｄ_ｚＢの両方を求める必要はなく、ｄ_ｚＡおよびｄ_ｚＢのいずれか一方を求めればよい。より具体的には、（１９）式が成立する場合には、ｄ_ｚＡを求めればよく、（２０）式が成立する場合には、ｄ_ｚＢを求めればよい。

In equations (134) and (135), z represents the height position of the surface on which the electron beam 1 is incident. _z0 and z1 represent the height positions of knife edges 20 and 20', _respectively . D _A0 and D _A1 represent D _A measured on knife edges 20 and 20', respectively. D _B0 and D _B1 represent D _B measured on knife edges 20 and 20', respectively. Both D _A1 and D _B1 are zero if the defocus on the knife edge 20' is corrected. Δz is the difference of z ₁ from z ₀ and has a magnitude of, for example, about 10 μm.
Here, it is not necessary to obtain both _dzA and _dzB , but one of _dzA and _dzB may be obtained. More specifically, _dzA should be obtained when the formula (19) holds, and _dzB should be obtained when the formula (20) holds.

The apparatus of the present embodiment then measures and corrects for dimensional error and rotation of the deflection field of the objective deflector 13 at the height position of each of the knife edges 20 and 20'.
Specifically, the apparatus of this embodiment first deflects the electron beam 1 toward a point on the boundary of the deflection field of the objective deflector 13 by the objective deflector 13. Positional deviations (dimensional errors and rotations of the deflection field of the objective deflector 13) are measured and corrected by the knife edge 20. FIG. At that time, correction values for dimensional error and rotation of the deflection field of the objective deflector 13 are determined, and these correction values are added to the deflection voltage. The apparatus of this embodiment then makes similar measurements and corrections with knife edge 20'. However, before these steps are performed on each knife edge, the defocus on each knife edge is measured and corrected.

The apparatus of this embodiment then obtains the differential coefficient of the dimensional error of the deflection field of the objective deflector 13 and the rotational correction value with respect to the height position z. These differential coefficients are obtained from the amount of change in the dimensional error and rotation correction value of the deflection field of the objective deflector 13 and the amount of change in the height position z (that is, Δz) in the above process.
Supplementally, these differential coefficients include not only changes in the dimensional error and rotation correction values of the deflection field of the objective deflector 13 due to the height position z, but also changes in these correction values due to the focus correction current _I0A . reflected.

The apparatus of this embodiment then establishes a reference point within the material 10 . The reference point may be, for example, a point indicating a height position corresponding to the average of the maximum and minimum height positions within the material 10, or, more simply, a point located in the center of the material 10. may be

In the apparatus of this embodiment, next, the material stage 14 moves the reference point directly below the electron beam 1, and then the height detector measures the height position of the reference point. The apparatus of this embodiment then converts the difference in height position from the height position of the knife edge 20 to a target value for the XY difference D _A0 or D _B0 at the reference point.
The target value of D _A0 or D _B0 at the reference point is given by equation (136) or (137), respectively. In equations (136) and (137), D _A0T and D _B0T represent target values of D _A0 and D _B0 at the reference point, respectively. _zT represents the height position of the reference point. D _A0T may be determined when the formula (19) holds, and D _B0T may be determined when the formula (20) holds.
D _A0T =d _zA (z _T −z ₀ ) (136)
D _B0T =d _zB (z _T −z ₀ ) (137)

The apparatus of this embodiment then moves the knife edge 20 directly under the electron beam 1 by the material stage 14 and then measures the XY difference D _A0 or D _B0 on the knife edge 20 . The apparatus of this embodiment then updates I _0A0 (ie determines I _0A0 ^(m)) according to equation (138) or (139). The updating continues until the formula (140) or (141) holds.
I _0A0 ^(m) = I _0A0 ^(m-1) - (D _A0 ^(m-1) - D _A0T )/d _IA (138)
I _0A0 ^(m) = I _0A0 ^(m-1) - (D _B0 ^(m-1) - D _B0T )/d _IB (139)
|D _A0 ^(m) −D _A0T |<ε _D (140)
|D _B0 ^(m) −D _B0T |<ε _D (141)

That is, the apparatus of this embodiment updates I _0A0 until D _A0 ^(m) converges to D _A0T or until D _B0 ^(m) converges to D _B0T . The update is based on the formula (138) when the formula (19) holds, and according to the formula (139) when the formula (20) holds. Here, convergence of D _A0 ^(m) to D _A0T or convergence of D _B0 ^(m) to D _B0T corresponds to correction of the axial defocus at the reference point.
In contrast, the devices of Examples 1 to 4 updated I _0A0 by equation (9a) or (10a). The updating was continued until equation (22) or (23) was established. In other words, in Examples 1-4, D _A0T and D _B0T were zero.

The apparatus of the present embodiment then begins to write onto material 10 . Thereafter, the apparatus of this embodiment corrects its own axial defocus and the dimensional error and rotation of the deflection field of the objective deflector 13 for each region within the material 10 . Here, each region in the material 10 is, more specifically, each deflection field of the objective deflector 13 . Details of these corrections are given below.

In the apparatus of this embodiment, first, of the regions in the material 10, the region whose height is to be measured first is moved by the material stage 14 directly below the electron beam 1, and the height position of that region is Measured by the height detector described above.
At this point, drawing to that region has not yet started. At this point, the electron beam 1 is blocked by the blanker 15 (see FIG. 24).

The apparatus of this embodiment then corrects for the axial defocus in the region (where the height measurement is first made). Therefore, the apparatus of the present embodiment updates I _0A0 (x _w , y _w ) in the above region according to equation (142) or (143) and inputs it to the objective lens 9 . Even at this point, drawing to the area has not yet started.
I _0A0 ′(x _w ,y _w )=I _0A0 (x _w ,y _w )+(d _zA /d _IA )Δz(x _w ,y _w )
= _I0A0 ( _xw , _yw )+ _ID0 ( _xw , _yw ) (142)
I _0A0 ′(x _w ,y _w )=I _0A0 (x _w ,y _w )+(d _zB /d _IB )Δz(x _w ,y _w )
= _I0A0 ( _xw , _yw )+ _ID0 ( _xw , _yw ) (143)
In equations (142) and (143), I _0A0 (x _w , y _w ) represents the focus correction current I _0A0 at the in-material coordinates (x _w , y _w ). In-material coordinates (x _w , y _w ) represent the coordinates of the area within material 10 where the height measurement is taken. I _0A0 ′(x _w , y _w ) represents I _0A0 (x _w , y _w ) updated to zero the axial defocus at in-material coordinates (x _w , y _w ). The apparatus of this embodiment matches I _0A0 (x _w , y _w ) before being updated with I _0A0 at the reference point. Here, I _0A0 at the reference point means I 0A0 updated by the formula (138) until the formula (140) holds, or I _0A0 updated by the formula (139) until the formula (141) holds. _0A0 . Δz(x _w , y _w ) represents the difference of the height position of the material 10 at the in-material coordinates (x _w , y _w ) from the height position of the reference point.

I _D0 (x _w , y _w ) in equations (142) and (143) are given by equations (144) and (145), respectively. I _D0 (x _w , y _w ) in equations (142) to (145) represents Δz(x _w , y _w ) converted into the amount of change in the focus correction current. (I _D0 (x _w , y _w ) also represents the measurement of Δz(x _w , y _w ).)
I _D0 ( _xw , _{yw)=(dzA / dIA} ) Δz( _xw , _yw ) (144)
I _D0 ( _xw , _{yw)=(dzB / dIB} ) Δz( _xw , _yw ) (145)
Therefore, equations (142) and (143) show that the apparatus of this embodiment adds Δz(x _w , y _w ) converted to the amount of change in the focus correction current to I _0A0 (x _w , y _w ). , thereby updating I _0A0 (x _w ,y _w ) (ie determining I _0A0 ′(x _w ,y _w )).

However, the above process assumes that the height difference in material 10 is sufficiently small (ie, |Δz(x _w , y _w )| is sufficiently small). Under that assumption, the nonlinearity and non-reproducibility of the defocus due to the focus correction current with respect to the focus correction current can be neglected. That is, the correction residual of the axial defocus due to its nonlinearity and non-reproducibility can be ignored. The nonlinearity and non-reproducibility involve the nonlinearity of the lens strength of the objective lens 9 with respect to the excitation current of the objective lens 9 and the magnetic hysteresis of the magnetic material of the objective lens 9 .

The apparatus of the present embodiment then determines the addition value to the deflection voltage in the region. The added value is the difference of the height position of the region from the height position of the knife edge 20 and the above-mentioned differential coefficient (that is, the correction value of the dimensional error and rotation of the deflection field with respect to the height position z is determined from the differential coefficient).
The added value is determined according to the deflection coordinates (coordinates within the deflection field of the objective deflector 13). More specifically, the added value is determined according to a linear function of the deflection coordinates (see, for example, JP-A-01-077937).

The apparatus of the present embodiment then begins drawing in the above area (the area where height measurements are first made). The apparatus of this embodiment adds the above-mentioned added value to the deflection voltage for each deflection coordinate during the drawing.
Dimensional errors and rotations of the deflection field of the objective deflector 13 in this region are then corrected. That is, drawing in the above area proceeds with these corrected.

The apparatus of this embodiment then performs the same writing, measuring, and correcting operations on the remaining regions within the material 10 as described above.
By doing so, the axial defocus produced by the apparatus of this embodiment and the dimensional error and rotation of the deflection field of the objective deflector 13 are adjusted for each region in the material 10 (more specifically, the Writing on the material 10 progresses while being corrected for each deflection field.
However, if there is no height difference in the material 10, Δz( _xw , _yw ) will be zero, so updating _I0A0 ( _xw , _yw ) according to equations (142) or (143) is unnecessary. Become.

In the above description, it was assumed that the height difference within the material 10 is sufficiently small (that is, |Δz(x _w , y _w )| is sufficiently small). It is possible that the correction residual for axial defocus cannot be ignored. This is because if |Δz(x _w , y _w )| is large, the non-linearity and non-reproducibility of the defocus caused by the focus correction current with respect to the focus correction current cannot be ignored.
If the above becomes a problem, the maximum value of |Δz(x _w , y _w )| that can occur on the material 10 should be reduced. By doing so, manifestation of the non-linearity and non-reproducibility is suppressed.
To do so, the material 10 is first appropriately divided into portions, each of which has a reference point, and then the height of each reference point is measured before the material 10 is drawn. , a target value _{D-- A0T} or _{D-- B0T} for each reference point. Then, after starting drawing on the material 10, update _I0A0 by equation (138) or (139) for each of the above portions, and define Δz(x _w , y _w ) within each of the above portions. Just do it.

(Example 10)
Still another embodiment (spot electron beam drawing apparatus) of the charged particle beam apparatus of the present invention will be described below as the tenth embodiment. For the sake of explanation, FIG. 22 shows the configuration of the optical system of the device (spot electron beam drawing device) of this embodiment.

The apparatus of this embodiment has the same configuration as the apparatus of the first to ninth embodiments (variably shaped electron beam writing apparatus). However, in this embodiment, the electron beam 1 is a spot beam instead of a variable shaped beam.

In order to convert the electron beam 1 into a spot beam in the apparatus of this embodiment, simply adjust the excitation currents of the irradiation lens 2, shaping lens 6, reduction lens 8, and objective lens 9, as shown in FIG. As shown, a light source image 16 may be focused between reduction lens 8 and objective lens 9, and a light source image 19 may be focused at the height of material 10 (and knife edge 20).
In this case, none of the first shaping aperture plate 3, the second shaping aperture plate 7, and the shaping deflector 12 (see FIG. 24) are necessary. The first shaping aperture plate 3 or the second shaping aperture plate 7 can be effectively used as an aperture stop. Therefore, in this embodiment, the shape of the openings of these aperture plates may be circular. (Generally, the shape of the apertures in these aperture plates is rectangular.)

The apparatus of this embodiment basically operates in the same way as the apparatuses of the first to ninth embodiments. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. The blurring of the electron beam 1 is measured by the knife edge method.
However, in this embodiment knife edge 20 (or knife edges 20 and 20') is scanned by electron beam 1 as a spot beam rather than by electron beam 1 as a variable shaped beam.

(Example 11)
Still another embodiment (scanning electron microscope) of the charged particle beam device of the present invention will be described below as an eleventh embodiment.

The apparatus (scanning electron microscope) of this embodiment basically has the same configuration as the apparatus (spot electron beam drawing apparatus) of the tenth embodiment. However, in this embodiment, an observation sample (not shown) is placed on the material stage 14 (see FIG. 22) instead of the material 10 .

The apparatus of this embodiment is provided with a secondary electron detector (not shown) or a transmission electron detector (not shown). Among these detectors, the secondary electron detector is provided above the observation sample (on the upstream side of the electron beam 1). On the other hand, the transmission electron detector is provided below the observation sample (on the downstream side of the electron beam 1).

The device of this embodiment basically operates in the same way as the device of the tenth embodiment. That is, the apparatus of this embodiment also measures and corrects its own axial defocus, axial 2-fold astigmatism, deflection curvature of field aberration, and deflection 2-fold astigmatism. Both the measurement and correction of these aberrations are based on the curve of the blurring of the electron beam 1 versus the two-fold astigmatism voltage. An image 19 of the light source is formed at the height of the material 10 .

The apparatus of this embodiment, however, uses the electron beam 1 not for writing on the material 10, but for observing the observation sample. For this purpose, the apparatus of this embodiment scans the observation sample with the electron beam 1 .
More specifically, the apparatus of this embodiment operates as a scanning electron microscope. More specifically, the apparatus of this embodiment is a scanning electron microscope (SEM) if secondary electrons generated from the observation sample during scanning of the observation sample are detected by the secondary electron detector. ). Alternatively, the apparatus of this embodiment operates as a scanning transmission electron microscope (STEM) if the transmitted electrons transmitted through the observation sample at that time are detected by the transmission electron detector.

The apparatus of this embodiment measures the blurring of the electron beam 1 by the knife-edge method or by another method.

Here, the other technique specifically refers to image analysis (two-dimensional Fourier analysis) of an electron image (scanning electron image) obtained by scanning the observation sample. Japanese Patent Application Laid-Open No. 10-106469), or signal analysis of electronic signals obtained by scanning the observation sample (see Japanese Patent Application Laid-Open No. 7-153407, for example).

Such a technique for measuring the blurring of the electron beam 1 eliminates the need for knife-edges (specifically knife-edge 20, knife-edge 20', or electron reflectors). Furthermore, a height detector (see Example 9) of the kind is also unnecessary. (In Example 9, neither an electronic image nor an electronic signal is acquired when the material 10 is scanned by the electron beam 1, so the height position of the material 10 is measured using the height detector described above. was required.)

In principle, the above two methods make it possible to measure the magnitude of blur in two orthogonal directions. More specifically, in the image analysis, the magnitude of the blur appearing in the electronic image in two arbitrary orthogonal directions can be evaluated independently of each other as the smallness of the wave numbers in the two orthogonal directions in Fourier space. . In the signal analysis, if the observation specimen is scanned along each of two arbitrary orthogonal directions, the two electron signals thus obtained can be used to determine the magnitude of the blur on the observation specimen in the two orthogonal directions. can be evaluated independently of each other.

Therefore, the above two techniques can be applied to the measurement and correction of any of the axial defocus, axial 2-fold astigmatism, deflection field curvature aberration, and deflection 2-fold astigmatism. That is, the electron image is acquired while increasing and decreasing the astigmatism correction voltage twice, and the image analysis is applied to the electron image thus obtained, or the image analysis is performed while increasing and decreasing the astigmatism correction voltage twice. By obtaining the electronic signal and applying the signal analysis to the electronic signal thus obtained, the blurring of the electron beam 1 in two orthogonal directions (or in a single direction) with respect to the two-fold astigmatism correction voltage curve can be obtained.
Based on the curves thus obtained, the axial defocus, the axial two-fold astigmatism, the deflection field curvature aberration, and the deflection two-fold astigmatism are measured and corrected. Except for the difference in means for measuring the blur of 1 (analysis of the electronic image or electronic signal obtained by scanning the observation sample, or analysis of the Faraday cup absorption current signal obtained by scanning the knife edge 20), It is the same that the same aberrations are measured and corrected in Examples 1-10.

(Example 12)
Still another embodiment (transmission electron microscope) of the charged particle beam apparatus of the present invention will be described as embodiment 12 below.

FIG. 23 shows the configuration of the optical system of the apparatus (transmission electron microscope) of this example. In FIG. 23, optical system components on the upstream side of the electron beam 1 from the objective lens 9 are omitted.

The apparatus of the present embodiment has partially the same configuration as the apparatus (scanning electron microscope) of the eleventh embodiment. More specifically, the apparatus of the present embodiment has the same configuration as that of the apparatus of the eleventh embodiment in the configuration of the optical system on the upstream side of the electron beam 1 from the specimen 33 to be observed.
However, unlike the apparatus of the eleventh embodiment, the apparatus of this embodiment includes optical system components below the observation sample 33 (on the downstream side of the electron beam 1). More specifically, the apparatus of this embodiment includes a magnifying optical system 34 and a transmission electron image detector 35 below the observation sample 33, as shown in FIG. Here, the observation sample 33 is a thin film sample. The magnifying optical system 34 consists of an objective lens 9' and magnifying lenses 36 and 37, as can be seen from FIG.
The apparatus of this embodiment includes a kind of knife edge (specifically, knife edge 20, knife edge 20', or an electron reflector) and a kind of deflector used for scanning (specifically, an objective lens). A deflector 13) is also not provided. The apparatus of this example is also not equipped with any kind of height detector (see Example 9).

The magnifying optical system 34 is provided with an objective diaphragm 38 as a diaphragm for converting the image (transmission electron image) formed on the transmission electron image detector 35 into a bright field image or a dark field image. Henceforth, the image is called a bright-field image.
The height position of the objective diaphragm 38 is set to a height position that is not optically conjugate with the transmission electron image detector 35 . More specifically, the height position of the objective diaphragm 38 is the height position of the back focal plane of the objective lens 9', or the downstream side of the electron beam 1 from the focal plane. The height position is optically conjugate with the height position. In the optical system of FIG. 23, the height position of the objective diaphragm 38 is the height position of the back focal plane of the objective lens 9'.

The enlarging optical system 34 is further provided with a two-fold astigmatism corrector 30 . Although the two-fold astigmatism corrector 30 is a magnetic multipole element (for example, an octopole element) in FIG. 23, it may be an electrostatic multipole element.
The height position of the two-fold stigmator 30 is set at the same height as (or near) the objective aperture 38, or at the downstream side of the electron beam 1 from the objective aperture 38. The height position is optically conjugate with the height position. In the optical system of FIG. 23, the height position of the two-fold stigmator 30 is downstream of the electron beam 1 from the objective aperture 38 and optically conjugate with the height position of the objective aperture 38. be.
The two-fold stigmator 30 is connected to the stigmator controller 31 . The stigmator controller 31 is connected to the central controller 28 (not shown in FIG. 23).

Supplementally, the height position of the two-fold stigmator 30 is the height position described above (that is, the same height position as the objective diaphragm 38, or a height position optically conjugate with that height position). The reason for this is to reduce the distortion aberration that is secondary to the generation of two-fold astigmatism by the two-fold astigmatism corrector 30 . The distortion aberration depends on the coordinates in the image 39 (transmission electron image) of the observation sample.
The above-mentioned distortion aberration is caused by positional deviation of each principal ray included in the electron beam 1 from the central axis of the electron beam 1 at the height position of the two-fold astigmatism corrector 30 . This involves the magnetic field distribution around each of its chief rays. Assuming that the height position of the two-fold stigmator 30 is the above height position, all principal rays contained in the electron beam 1 intersect at one point at the height position of the two-fold stigmator 30. Therefore, The positional deviation is minimized, and the distortion aberration is also minimized.

The apparatus of this embodiment operates as a transmission electron microscope (TEM) under the above configuration. Its operation is as follows.

The apparatus of this embodiment first irradiates the observation sample 33 with the electron beam 1 . Here, in the apparatus of this embodiment, the height position of the image 19 of the light source is matched with the height position of the object-side focal plane of the objective lens 9 in order to irradiate the observation sample 33 with the electron beam 1 as a parallel beam. Let
In the apparatus of this embodiment, the electron beam 1 transmitted through the observation sample 33 is then transmitted to the transmission electron image detector 35 by the magnifying optical system 34 . Here, the magnifying optical system 34 forms an image 39 of the observation sample on the transmission electron image detector 35 (image plane).
In the apparatus of this embodiment, the transmission electron image detector 35 acquires an image 39 of the observed sample as an electronic image (electronic data).
The apparatus of this embodiment also measures and corrects defocus and two-fold astigmatism on the transmission electron image detector 35 . This is for sharpening the image 39 of the observation sample.
Here, the defocus and 2-fold astigmatism on the transmission electron image detector 35 are the axial defocus and 2-fold astigmatism produced by the apparatus of the present embodiment. It occurs up to the transmission electron image detector 35 . (Supplementally, the apparatus of this embodiment does not deflect the electron beam 1 during observation of the observation sample 33. Therefore, the apparatus of this embodiment is capable of correcting deflection curvature of field and deflection two-fold astigmatism. neither will occur.)

The above axial defocus and axial two-fold astigmatism measurements are based on the blur curve of the electron beam 1 versus the two-fold astigmatic current. However, the apparatus of this embodiment acquires these curves from the image 39 of the observation sample. For this purpose, the apparatus of this embodiment obtains an image 39 of the observation specimen by means of the transmission electron image detector 35 while increasing and decreasing the astigmatism correction current twice. In contrast, the aforementioned image analysis (see Example 11) is applied.

The axial defocus and the axial two-fold astigmatism are corrected by the magnifying lens 37 (or the magnifying lens 36) and the two-fold astigmatism corrector 30, respectively. That is, in this embodiment, the focus correction current and the two-fold astigmatism correction current are input to the magnifying lens 37 (or the magnifying lens 36) and the two-fold stigmator 30, respectively.

The above-described axial defocus and axial two-fold astigmatism are measured and corrected in the manner described above, because the difference in the correction signal (either the two-fold astigmatism correction current or the , or two-time astigmatism correction voltage), the difference in the acquired electron image (transmission electron image or scanning electron image), and the difference in means for correcting these aberrations (magnifying lens 37 and two-time astigmatism correction 30, or objective lens 9 and objective deflector 13), the same aberrations are measured and corrected by the image analysis described above in Example 11.

The present invention can adopt not only the configuration of each embodiment described above, but also any configuration within the scope defined in the claims. For example, in Examples 1 to 12, the electron beam 1 may be replaced with an ion beam (not shown). However, in that case, an electron gun (not shown) for generating the electron beam 1 is replaced with an ion beam source (not shown) for generating a desired ion beam.

1 electron beam, 2 irradiation lens, 3 first shaping aperture plate, 4 light source, 5, 16, 19 image of light source, 6 shaping lens, 7 second shaping aperture plate, 8 reduction lens, 9, 9' objective lens , 10 material, 11 projected figure, 12 shaping deflector, 13 objective deflector, 14 material stage, 15 blanker, 17 blanking aperture plate, 20, 20' knife edge, 21 Faraday cup, 24 lens control section, 25 objective deflection instrument control unit 26 material stage control unit 27 Faraday cup absorption current signal processing unit 28 central control unit 29 storage unit 30 two-time stigmator 31 stigmator control unit 33 observation sample 34 enlargement optical system 35 transmission electron image detector 36, 37 magnifying lens 38 objective diaphragm 39 image of observation sample 50 0th double astigmatic voltage 51 first double astigmatic voltage 52 th 2 double astigmatic voltage

Claims (11)

a light source for generating a charged particle beam, an optical system for the charged particle beam, and blur measurement means for the charged particle beam;
The optical system comprises at least one stage lens for transmitting the charged particle beam to a surface to be irradiated by the charged particle beam,
The optical system may be an image of the light source, a transmitted image of an aperture in an aperture plate arranged in the optical system and illuminated by the charged particle beam, or an image of the aperture plate arranged in the optical system and illuminated by the charged particle beam. connecting a transmission image of a thin film irradiated by the at least one stage lens onto the irradiated surface;
The optical system further includes focus correction means for correcting defocus appearing at a height position of the surface to be illuminated, and two-fold astigmatism correction for correcting two-fold astigmatism appearing at a height position on the surface to be illuminated. means and
The focus correction means receives a signal for correcting the defocus as a focus correction signal, changes the defocus according to the focus correction signal,
The two-fold astigmatism correction means receives a signal for correcting the two-fold astigmatism as a two-fold astigmatism correction signal, and changes the two-fold astigmatism in accordance with the two-fold astigmatism correction signal. let
The blur measuring means measures the blur of the charged particle beam appearing at a height position of the illuminated surface, the blur varying in magnitude depending on the defocus and the two-fold astigmatism.
A charged particle beam device configured as
While increasing or decreasing the intensity of the double astigmatism correction signal, the blur measuring means measures the magnitude of the blur along each of two directions orthogonal to each other in the irradiated surface, The intensity of the double astigmatic correction signal that minimizes the magnitude in one of the two directions, and the double astigmatism correction that minimizes the magnitude of the blur in the other one of the two directions. a first method of determining a difference from the intensity of a signal as a first difference, said first difference representing said defocus;
While increasing or decreasing the intensity of the double astigmatism correction signal, the blur measuring means measures the magnitude of the blur along each of two directions orthogonal to each other in the irradiated surface, The intensity of the double astigmatic correction signal that minimizes the magnitude in one of the two directions, and the double astigmatism correction that minimizes the magnitude of the blur in the other one of the two directions. the intensity of the two-time astigmatism-corrected signal that minimizes the magnitude of one of the two directions of the blur, and the first a second method of determining as a second difference a difference from the average of and said second difference representing said defocus;
Alternatively, the magnitude of the blur is measured along a single direction within the illuminated surface by the blur measuring means while increasing or decreasing the intensity of the twice astigmatic correction signal, The magnitude of the blur is measured along the single direction while increasing or decreasing the strength of the twice astigmatically corrected signal in one direction and the strength of the twice astigmatically corrected signal. obtaining the defocus by a third method, wherein the difference from the intensity of the two-fold astigmatism-corrected signal before is determined as a third difference, and the third difference represents the defocus; A charged particle beam apparatus characterized by being configured as follows. A knife-edge-shaped blur measurement medium for measuring the blur is provided at a height position of the irradiated surface,
scanning means for causing the charged particle beam to scan the blur measurement medium by deflecting the charged particle beam between the light source and the illuminated surface;
using the blur measuring medium and the scanning means as the blur measuring means,
The blur measurement by the blur measurement means is carried out by scanning the blur measurement medium with the scanning means, while the current of the charged particle beam in the portion not blocked by the blur measurement medium or the current in the blur measurement medium is detected. By detecting the current in the occluded portion and evaluating the blur based on the blunting of the waveform of the detected current,
A charged particle beam device according to claim 1 . The two-fold astigmatism correction signal is composed of two mutually independent components,
two independent components constituting the two-fold astigmatism correction signal are applied to the two-fold astigmatism correcting means, whereby the two-fold astigmatism is given a change composed of the two linearly independent components,
having a non-zero magnitude among the two types of the first, the second, or the third difference obtained by increasing or decreasing each of the two mutually independent components constituting the two-time astigmatism correction signal , the defocus is represented by any one of the first, the second, or the third difference;
A charged particle beam apparatus according to claim 1 or 2. any one of said first, said second, or said third difference having said non-zero magnitude, of said two of said first, said second, or said third difference, one of said first, said second, or said third difference, whichever is not less in absolute value;
The defocus is represented by one of the first, second, or third differences, whichever is not smaller in absolute value.
A charged particle beam apparatus according to claim 3 . one of the two mutually independent components constituting the two-time astigmatism correction signal, which is one of the first, second, or third differences for representing the defocus; One component that exhibits is referred to as the first component,
determining said first difference exhibited by said first component by said first method, determining said second difference exhibited by said first component by said second method, or determining the third difference exhibited by the first component by the third method;
representing the defocus by the first, the second, or the third difference exhibited by the first component;
giving a predetermined amount of change to the intensity of the focus correction signal, wherein the first, the second, or the third difference exhibited by the first component is the predetermined change in the intensity of the focus correction signal; From the difference between the first, the second, or the third before and after the amount is given and the predetermined amount of change, the first, the second, or the third determining the partial derivative of the difference with respect to the intensity of the focus correction signal as a first partial derivative;
The defocus represented by the first, the second, or the third difference exhibited by the first component is converted into an amount of change in intensity of the focus correction signal by the first partial differential coefficient. do,
A charged particle beam apparatus according to claim 3 or 4. After determining the first partial derivative, storing the determined value of the first partial derivative;
of the defocus represented by the first, second, or third difference exhibited by the first component after storing the determined first partial derivative value; repeatedly using the first partial differential coefficient of the stored value for conversion into an amount of change in intensity of the focus correction signal;
A charged particle beam device according to claim 5 . After converting the defocus into the amount of change in the intensity of the focus correction signal by the first partial differential coefficient, the defocus converted into the amount of change in the intensity of the focus correction signal is converted to the amount of change in the intensity of the focus correction signal, subtracting from the intensity of the focus correction signal before the predetermined amount of change is given to the intensity of the signal, and providing the focus correction signal having an intensity equal to the result of the subtraction to the focus correction means;
A charged particle beam apparatus according to claim 5 or 6. when the defocus is represented by the second difference exhibited by the first component, providing the first component having an intensity equal to the first average to the twice astigmatism correction means;
When the defocus is represented by the first difference exhibited by the first component, the intensity of the first component that minimizes the magnitude of the blur in one of the two directions, and determining the average of the intensity of the first component that minimizes the magnitude of the other of the two directions of blur as the first average; providing one component to the two-fold astigmatism correction means;
A charged particle beam apparatus according to claim 7 . When the defocus is represented by the third difference exhibited by the first component,
The single and the first component of the blur that minimizes the magnitude in the direction perpendicular to the single direction in the illuminated surface. determining the average with the intensity as a second average, and providing the first component of intensity equal to the second average to the two-fold astigmatic correction means;
A charged particle beam device according to claim 7 . Of the two mutually independent components constituting the two-time astigmatism correction signal, one component different from the first component is referred to as a second component, and the first, determining the second or the third difference, applying the predetermined amount of change, or a predetermined amount of change other than the predetermined amount, to the intensity of the focus correction signal; is the first, the second, or the third difference exhibited by the component of the first difference before and after the predetermined amount of change or the another predetermined amount of change is given to the intensity of the focus correction signal , the second or the third difference and the predetermined amount of change or the another predetermined amount of change, the first, the second or the third difference exhibited by the second component determining the partial differential coefficient with respect to the intensity of the focus correction signal as a second partial differential coefficient,
When the defocus is represented by the first difference exhibited by the first component, the first difference exhibited by the second component is determined by the first method, and the second A partial derivative is determined from the first difference exhibited by the second component, and where the defocus is represented by the second difference exhibited by the first component, the second component exhibits The second difference is determined by the second method, the second partial derivative is determined from the second difference exhibited by the second component, and the defocus is the first When expressed by the third difference exhibited by the components, the third difference exhibited by the second component is determined by the third method, and the second partial differential coefficient is calculated by the second determined from the third difference exhibited by the components of
after expressing the defocus by the first, the second, or the third difference exhibited by the first component, wherein the defocus so expressed is the first partial derivative; After converting into the amount of change in the intensity of the focus correction signal by a coefficient, the defocus converted into the amount of change in the intensity of the focus correction signal is further converted into the amount of change in the intensity of the focus correction signal by the second partial differential coefficient. converted to the first, second, or third difference exhibited by the two components,
The intensity of the second component that minimizes the magnitude of one of the two directions of the blur, and the second component that minimizes the magnitude of the other of the two directions of the blur determining the average of the intensity of the component as another first average, or the intensity of the second component that minimizes the magnitude of said single direction of said blur and determining, as another second average, the average of the intensity of the second component that minimizes the magnitude in the direction orthogonal to the single direction in the illuminated surface; or
When the defocus is represented by the first difference exhibited by the first component, the intensity of the second component that minimizes the magnitude of the blur in one of the two directions, and , the defocus converted to the first difference exhibited by the second component, and the another first average is determined by a fourth method. death,
When the defocus is represented by the second difference exhibited by the first component, the intensity of the second component that minimizes the magnitude of one of the two directions of the blur and , and the defocus converted to the second difference exhibited by the second component. death,
When the defocus is represented by the third difference exhibited by the first component, the magnitude of the blur in the single direction is minimized, or the blur in the illuminated surface is From the intensity of the second component that minimizes the magnitude in the direction orthogonal to the single direction and the defocus converted to the third difference exhibited by the second component, the another determining the another second average by a sixth method of determining the average of the two;
providing said second component of intensity equal to said another first or said another second average determined to said two-fold stigmator;
A charged particle beam apparatus according to claim 8 or 9. After determining the second partial derivative, storing the determined value of the second partial derivative;
After storing the determined value of the second partial differential coefficient, the second component of the defocus converted into an amount of change in intensity of the focus correction signal by the first partial differential coefficient repeatedly using the second partial derivative of the stored value to convert the first, second, or third difference represented by
Charged particle beam device according to claim 10 .

Images