JPH11307029A - Transmission electron microscope - Google Patents

Abstract

PROBLEM TO BE SOLVED: To avoid waste of utilization efficiency caused by many pieces of analyses by forming the conformal mapping of a reference film from a defocus image of the reference film, conducting orthogonality judgment, and analyzing strain of one piece of defocus image to obtain correct result when judged that this is the conformal mapping of the actual reference film. SOLUTION: This electron microscope has a function mapping a defocusing image of a reference image on a (x+δx , y-δy ) plane, a function judging whether sequence of points corresponding to the lattice point of x = constant in the (x, y) plane appearing in the mapping and sequence of points corresponding to the point of y = constant perpendicularly cross everywhere, and a function selecting decision of magnetic information from δx and δy at that time or decision by a method obtaining a plurality of defocusing images, based on judging results. A plurality of defocusing images can be obtained only when it is required, and as a result, observation efficiency can be enhanced without damaging accuracy and reliability.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electron microscope technique, and more particularly to a transmission electron microscope capable of observing a magnetic field distribution in a minute area with high accuracy and efficiency.

[0002]

2. Description of the Related Art Conventionally, when it is necessary to observe a magnetic field distribution in an extremely small area with high resolution, a transmission electron microscope (T
The following method using EM) has been used. In this method, a TEM image of the reference thin film is reduced and formed in a magnetic field distribution region. The electron beam passing through each part of the magnetic field distribution region is
It is deflected by the force (Lorentz force) applied from the magnetic field while passing through that portion. Therefore, if an image at a position slightly deviated from the magnetic field distribution region (referred to as a defocused image) is enlarged and formed on a photographic plate or an image sensor, an image of the reference film distorted by the magnetic field can be obtained. Then, the magnetic field distribution can be estimated by analyzing the distortion of the image.

In order to determine the magnetic field distribution more accurately, a defocused image of the magnetic field distribution region is observed from various directions (actually, the observation direction is fixed and the magnetic field distribution region is rotated due to the limitations of the apparatus). ), Tomography processing is performed on data extracted from these defocused images to reconstruct a three-dimensional magnetic field distribution. A TEM capable of observing the magnetic field distribution based on such a principle is disclosed in, for example, Japanese Patent Application Laid-Open No.
It is described in No. 8.

Next, a description will be given of a reason why a magnetic field is distorted in the defocused image and a method for obtaining information on the magnetic field distribution from the distortion, with reference to FIG. In the figure, the z-axis is defined in parallel with the traveling direction of the electron beam incident on the magnetic field distribution region, and the x-axis and the y-axis perpendicular to the z-axis are defined in the magnetic field distribution region (here, It is assumed that the x, y, and z axes form a right-handed system as usual). Normally, under the conditions in which the above technique is implemented, the deflection due to the magnetic field is very slight, and the x and y components of this deflection can be expressed by Equations 1 and 2, respectively.

[0005]

(Equation 1)

[0006]

(Equation 2)

Here, e is the absolute value of the electron charge, m is the mass of the electron, V is the acceleration voltage of the electron beam, and Bx, By
Are the x and y components of the magnetic field B in the magnetic field distribution region, respectively.

Accordingly, the point (x, y) in the magnetic field distribution region
After traveling through the distance L, the electron beam deviates in the x-axis direction and the y-axis direction by the distances given by Equations 3 and 4, respectively. Occurs.

[0009]

(Equation 3)

[0010]

(Equation 4)

Therefore, conversely, if these shift amounts are obtained for each point (x, y) in the magnetic field distribution region from the analysis of the defocused image, an integrated value of Bx and By at the point in the z-axis direction is obtained. be able to.

However, in the magnetic field distribution extraction method as described above, an error may occur depending on the shape of the magnetic field distribution region. This will be described with reference to FIG. FIG. 3 shows the trajectories of two electron beams passing through different portions of the magnetic field distribution region. These electron beams travel in parallel to the optical axis (not shown in the figure), enter the magnetic field distribution region, and after passing through, travel linearly again while being deflected with respect to the optical axis. Here, the intersection of the extension lines of the trajectory before and after entering the magnetic field distribution region is called a deflection fulcrum. The above error occurs when the deflection fulcrum is shifted in the optical axis direction by each electron beam. Such a situation appears when the magnetic field spreads in the direction of the optical axis.

At this time, as shown in FIG. 3, the distance from the deflection fulcrum of the defocused image (L and L 'in the figure) does not coincide with each other (L ≠ L'), and the deflection angle is changed. The relationship between (in the figure, θ and θ ′) and the amount of deviation on the defocused image (in the figure, δ and δ ′) is not constant. FIG. 3 shows a case where the deviation amounts on the defocused image match (δ = δ ′) even though the deflection angles are different (θ ≠ θ ′). Similarly, it goes without saying that even when the deflection angles are the same (θ = θ ′), the deviation amounts on the defocused image may not match (δ ≠ δ ′).

In order to avoid such a problem, conventionally, a shift amount of a specific point has been obtained in a plurality of defocus images obtained at different positions from the magnetic field distribution region.

FIG. 3 shows a case where defocused images at positions separated by a distance K are compared. From this, the amount of deviation between the two images of each electron beam (γ and γ ′ in the figure) can be accurately determined regardless of the position of the deflection fulcrum.
(Since only the case of θ, θ′≪1 is assumed in the present invention, γ = Kθ, γ ′ = Kθ ′).

With respect to the magnetic field distribution extraction method as described above,
It is described in Japanese Patent Application Laid-Open No. Hei 7-262951, and is effective in obtaining a highly reliable magnetic field distribution particularly when the magnetic field distribution region is three-dimensionally spread as well as when it is planar.

However, in this method, it is necessary to acquire a plurality of defocused images. In principle, it is sufficient to have two defocused images as shown in FIG. 3, but in practice, in order to ensure reliability, about 10 defocused images at different distances from the magnetic field distribution region It is common to use. Furthermore, in order to reconstruct a three-dimensional magnetic field distribution by tomography processing, a set of these defocused images must be acquired from various directions, so that the number of necessary defocused images is extremely large. I will.

In practice, the magnetic field distribution region is planar, and the analysis using the above-described procedure using a plurality of defocus images is often unnecessary. However, heretofore, it has been assumed that the magnetic field distribution region is planar. Since there is no means for determining whether or not the analysis can be performed, when a highly reliable result is required, the analysis using a plurality of defocus images is generally performed uniformly.

[0019]

As described above, in the above-mentioned prior art, when the magnetic field is distributed in a plane perpendicular to the optical axis, in principle, one defocused image can be obtained. Despite accurate results being obtained by distortion analysis, a plurality of defocused images are obtained uniformly. Therefore, this has been a major factor that hinders the improvement of the utilization efficiency of the above-described conventional technology.

An object of the present invention is to avoid wasteful acquisition of data (defocused image) generated depending on the distribution shape of a magnetic field to be observed, and to improve observation efficiency without reducing accuracy and reliability of the result. That is.

[0021]

According to the present invention, there is provided:
Attention is paid to a general property of a defocused image in which a distortion is generated by a magnetic field.

First, from the integral part of Equations (3) and (4),
Equation 6 is defined.

[0023]

(Equation 5)

[0024]

(Equation 6)

From Equations 5 and 6, Equations 7 and 8 are clear.

[0026]

(Equation 7)

[0027]

(Equation 8)

As is well known, for magnetic field B, equation 9 always holds. In Equation 10 in which both sides of Equation 9 are integrated in the z-axis direction (Bz (x, y, z) is the z component of the magnetic field B), the third term on the left side (Equation 11) is 0 when the strength of the magnetic field component is sufficiently far away. It is 0 because it becomes. Therefore, Equation 1
By swapping the order of differentiation and integration of the first and second terms on the left side of 0, Equation 12 is obtained.

[0029]

(Equation 9)

[0030]

(Equation 10)

[0031]

[Equation 11]

[0032]

(Equation 12)

In general, the equation (13) is expressed as (M, i,
D is the magnetization, current density, electric flux density, and μ, respectively.
₀ and t are respectively the magnetic permeability of vacuum and time), and in a static vacuum, M, i and ∂D / ∂t are all 0, so that Equation 14 holds. In the following, it is assumed that Equation 14 always holds.

[0034]

(Equation 13)

[0035]

[Equation 14]

In Expression 15 obtained by integrating both sides of the z component of Expression 14 as a vector quantity in the z-axis direction, Expression 16 is obtained by changing the order of differentiation and integration.

[0037]

(Equation 15)

[0038]

(Equation 16)

From the above equations 12 and 16, the complex function w ₁ (x, y) defined by equation 17 and the complex function w ₂ (x, y) (= −iW ₁ (x , Y)) are analytic functions that take a complex variable z defined by z≡x + iy as an argument.

[0040]

[Equation 17]

[0041]

(Equation 18)

Further, from the expression 18, the complex function w ₃ (x, y) defined by the expression 19 and the complex function w ₄ (x, y) defined by the expression 20 also have z≡x + iy as an argument. Obviously, it is an analytic function.

[0043]

[Equation 19]

[0044]

(Equation 20)

[0045] Thus, due to the complex function w ₃ (x,
y) to the (x + δ _x , y−δ _y ) plane or from the (x, y) plane by the complex function w ₄ to the (x−δ _x , y +
δ _y ) The mapping onto the plane is a conformal mapping. Each point on the (x + δ _x , y + δ _y ) plane on the defocused image in which the distortion due to the magnetic field has occurred can be converted to the (x + δ _x , y-δ _y ) plane or the (x-δ _x , y + δ _y ) plane by simple deformation. Can be mapped.

The (x + δ _x , y−δ _y ) plane and the (x−δ _x , y + δ _y ) plane thus obtained and the (x, y) plane defined in the magnetic field distribution region are actually mutually conformal mappings. If the relations are satisfied, L and R are the same at each point in the image.
It is supported by the fact that
Assuming that the magnetic field distribution region is planar, it can be determined that accurate results can be obtained from the analysis of the distortion of one defocused image.

According to Equation 17, a similar determination is made as follows:
Such as _{(x + δ y, y +} δ x) plane, generally (Re (x +
iy + α (δ _x −iδ _y )), Im (x + iy + α (δ _x −i
δ _y ))) It is also possible to see if the plane has a conformal mapping relationship with the (x, y) plane. Here, Re (z) and Im (z) are symbols representing a real part and an imaginary part of the complex number z, respectively, and α is an arbitrary complex number. However, in this case, (x
+ Δ _x , y + δ _y ) planes must be transformed, so in practice, the (x + δ _x , y−δ _y ) plane and the (x−δ
_x , y + δ _y ).

[0048] In the _{following, (x + δ x, y} -δ y) performs the mapping only in the description of the _{plane, (x-δ x, y} + δ y) be exactly the same description applies to the mapping of the plane is true needless to say No.

[0049] (x, y) from the plane _{(x + δ x, y-} δ y) for mapping to the plane of determination of whether it is the conformal mapping, the reference thin film to shrink imaged field distribution region TE
And M image _{(x + δ x, y-} δ y) it is desirable that the corresponding coordinates in the plane has an easily patterned. For this purpose, for example, it is convenient to use a reference film having a fine opening at a lattice point. In this case, x-axis direction of the arrangement of the lattice points appearing in the TEM image, and a direction perpendicular thereto is defined as _{y-axis, (x + δ x, y} -δ y) in the plane, (x, y It is only necessary to determine whether or not a point sequence corresponding to x = constant grid point and a point sequence corresponding to y = constant grid point on the plane are orthogonal to each other everywhere.

Therefore, in the present invention, a defocused image (x +
δ _x , y-δ _y ) A function to map onto the plane, which appears
a function of determining whether or not a point sequence corresponding to x = constant grid point and a point sequence corresponding to y = constant grid point on the (x, y) plane are orthogonal to each other everywhere; or to determine from [delta] _x and [delta] _y at that time, or whether to determine a method of acquiring a plurality of defocus image plus a function of selecting based on the result of the determination. As a result, it is possible to obtain a plurality of defocused images only when necessary, and it is possible to improve the observation efficiency without reducing the accuracy and reliability of the result.

[0051]

FIG. 1 shows an embodiment of the present invention. In FIG. 1, an electron beam 2 emitted from an electron gun 1 composed of an electron source, an acceleration unit and the like irradiates a reference film 4 via an irradiation optical system 3. The reference film 4 is provided with a large number of minute openings regularly arranged in a lattice. If the reference film 4 has a shape in which minute regions having different transmittance from the surroundings are arranged in a lattice, the following description applies without any change.

The electron beam 2 transmitted through the reference film 4 reaches the magnetic field distribution region via the reduction optical system 5. Although the source of the magnetic field may be any source, FIG. 1 shows a case where the source is a magnetic head 6 for a magnetic disk. In this magnetic field distribution area, a reduced image of the reference film 4 is formed by the reduction optical system 5. In FIG. 1, this reduction rate is 1 / N.
The portion of the electron beam 2 that is not blocked by the magnetic head 6 reaches the detector 8 via the magnifying optical system 7 after passing through the magnetic field distribution region. The magnifying optical system 7 magnifies an image (defocused image) of a surface separated by a desired distance (a defocus amount; L in FIG. 1) from the magnetic field distribution region and forms an image on the detector 8. In FIG. 1, this enlargement factor is set to M.

Here, the defocus amount can be set to an arbitrary value within a necessary range including the case where L = 0 (positive focus image). The setting of the defocus amount at the time of observation is performed by controlling the magnifying optical system 7 by the observation control device 9.

The defocused image formed on the detector 8 is sent to the image analyzer 10. Then, information on the magnetic field distribution is extracted from the defocus image. For this purpose, the image analysis device 10 performs the distortion (δ _x , δ _y ) due to the magnetic field.
The function of storing the defocused image where the image is generated, and the (x + δ _x , y + δ _y ) plane defined on the defocused image is represented by (x
+ Δ _x , y−δ _y ) plane (conformal mapping generation function), which appears on this (x + δ _x , y−δ _y ) plane (x, y)
A function of determining whether or not a point sequence corresponding to x = a fixed lattice point and a point sequence corresponding to a y = constant lattice point on a plane are orthogonal to each other everywhere; A function of selecting from _x and δ _y or a method of obtaining a plurality of defocus images based on the result of the above determination, and a function and a plurality of functions of determining magnetic field distribution information from δ _x and δ _y The function to determine from the defocused image is provided. Note that a defocused image required for the analysis is obtained under the control of the observation control device 9. And the data on the magnetic field distribution determined by the above method is:
The information is stored in the magnetic field distribution information storage device 11.

Next, the relationship between images used in image analysis performed in the image analysis apparatus 10 will be described with reference to FIG. The image on the (x, y) plane defined in the magnetic field distribution region is a reduced image of the reference film 4 without distortion. Reference film 4 in the present embodiment
In this example, a large number of minute openings are arranged in a grid. In FIG. 4, among the lattice patterns appearing in the reduced image, 4 × 4 =
16 are shown. The grid pattern, on the defocus image _{(x + δ x, y +} δ y) distorted by the influence of the magnetic field in the plane. Grid points on the (x, y) plane and (x + δ _x , y +
From [delta] _y) between the lattice point corresponding to the plane of the lattice points by a simple image transformation (x + [delta] _x, it is possible to determine the position of at _y-δ y) plane.

Next, the flow of processing from the start of observation to the output of magnetic field distribution information in the present invention will be described with reference to FIG.

First, with respect to the defocused image acquired first, (x + δ _x , y−δ _y ) is obtained by the above-described image conversion.
Generate a grid pattern on a plane (conformal mapping generation).
Then, the _{(x + δ x, y-} δ y) lattice pattern on the plane, to make sure that it is actually (x, y) conformal mapping of the grating pattern of the plane. This includes (x +
On the [delta] _x , y- [delta] _y ) plane, a sequence of points corresponding to x = constant grid points and a sequence of points corresponding to y = constant grid points on the (x, y) plane are orthogonal to each other everywhere. (Orthogonality judgment).

As a result, if the grid pattern on the (x + δ _x , y−δ _y ) plane is actually a conformal mapping of the grid pattern on the (x, y) plane, the distortion amount δ _x immediately magnetic field distribution by number 3 and number 4 from [delta] _y is obtained. On the other hand, when the mapping is not conformal, it can be determined that the magnetic field distribution region cannot be regarded as a plane perpendicular to the optical axis. In this case, as described with reference to FIG. 3, a plurality of defocused images are acquired, the movement of each lattice point between the images is tracked, and the inclinations θ _x and θ _{y of the} locus indicating the movement with respect to the optical axis are obtained. Is determined, and the magnetic field distribution is determined from Expressions 1 and 2.

The configuration and operation of the embodiment of the present invention have been described above. The transmission electron microscope image of the reference film 4 is once formed in the magnetic field distribution region, and the image is formed with an arbitrary defocus amount. Function to compare the defocused image and the confocal image obtained thereby, and distortion (δ _x , δ _y ) at each point (x, y) in the confocal image that appears in the defocused image due to the influence of the magnetic field. , Each point (x, y) in the focus image is a point (Re (x + iy + α (δ _x −iδ _y )), Im (x + iy
+ Α (δ _x −iδ _y ))), a function of determining whether the generated image is a conformal mapping of a positive focus image, and as a result, this is a conformal If it is a mapping, a function to obtain the magnetic field distribution from Equations 3 and 4 from the distortion (δ _x , δ _y ). If this is not a conformal mapping, a plurality of defocused images can be obtained. Obtain and track the movement of each point in the positive focus image between images, determine the inclination θ _x , θ _y of the locus indicating the movement with respect to the optical axis, and determine the magnetic field distribution from Equations 1 and 2. It goes without saying that a transmission electron microscope having a function can be implemented without impairing the essence of the present invention even if it has a configuration different from that of the present embodiment.

[0060]

According to the present invention, when the magnetic field is distributed in a plane in the direction perpendicular to the optical axis, an accurate result can be obtained in principle by analyzing the distortion of one defocused image. Nevertheless, it is possible to avoid the waste of utilization efficiency in the conventional technology that multiple defocused images must be acquired, and it is possible to improve the observation efficiency without reducing the accuracy and reliability of the result. became.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing an apparatus configuration according to an embodiment of the present invention.

FIG. 2 is an explanatory view of an electron beam optical system showing the principle of the prior art.

FIG. 3 is an explanatory diagram of an electron beam optical system showing the principle of the prior art.

FIG. 4 is an explanatory diagram of the principle of the present invention.

FIG. 5 is a flowchart of processing according to an embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Electron gun, 2 ... Electron beam, 3 ... Irradiation optical system, 4 ... Reference film, 5 ... Reduction optical system, 6 ... Magnetic head, 7 ... Magnification optical system, 8 ... Detector, 9 ... Observation control device, 10 ... Image analysis device, 11 ... Magnetic field distribution information storage device.

Claims (4)

[Claims] 1. A transmission electron microscope comprising an electron gun, a reference film stage, a reference film imaging optical system, an observation magnetic field distribution source installation device, a reference film image magnifying optical system, an imaging device, and an image analysis device. The optical system for imaging a reference film is a transmission electron microscope image of the reference film set on the stage for the reference film, and a region where a magnetic field emitted from a magnetic field distribution source installed in the device for installing a magnetic field distribution source to be observed is distributed (magnetic field distribution region). ), And the reference film image magnifying optical system converts the image of the reference film formed in the magnetic field distribution region into a desired defocus amount (including a defocus amount of 0) by the imaging device. The image analysis apparatus has a function of forming a distortion (δ _x ) at each point (x, y) in the confocal image appearing in the defocused image of the reference film obtained by the above functions. , Δ _y ) to each point (x, y ) Is the point (Re (x + iy + α (δ _x −
iδ _y )), Im (x + iy + α (δ _x −iδ _y ))) (where Re (z) and Im (z) are symbols representing the real part and imaginary part of complex number z, respectively, and α is an arbitrary complex number ) And a function to determine whether the generated image is a conformal mapping of the above-mentioned focus image, and, as a result, when this is a conformal mapping Has a function of obtaining the magnetic field distribution from the distortion (δ _x , δ _y ) and, if this is not a conformal mapping, a plurality of defocused images between the images of each point in the right focus image. A transmission electron microscope characterized by having a function of tracking the movement of a moving object and determining the magnetic field distribution from the inclination of the locus indicating the movement with respect to the optical axis. 2. The transmission electron microscope according to claim 1, wherein each point (x, y) in the focus image is a point (Re (x + i)
y + α (δ _x −iδ _y )), Im (x + iy + α (δ _x −i
δ _y ))) corresponds to each point (x,
y) is an image corresponding to the point (x + δ _x , y−δ _y ), that is, α = 1, or an image corresponding to the point (x−δ _x , y + δ _y ), ie, α = −1. A transmission electron microscope characterized by the above-mentioned. 3. The transmission electron microscope according to claim 2, wherein the reference film has a shape in which minute regions having different transmittance from the surroundings are arranged in a grid. 4. A transmission electron microscope according to claim 3, wherein said reference film has an opening in a minute area where the transmittance of said reference film is different from that of the surrounding area.