JP5039633B2 - Electron microscope and astigmatism evaluation method - Google Patents

Description

  The present invention relates to evaluation and correction of astigmatism in an electron microscope.

  Adjustment of astigmatism correction in an electron microscope requires skill. It takes a considerable amount of time for an unfamiliar operator to complete the adjustment. In order to reduce such labor, various astigmatism correction automated methods have been devised.

  For example, as described in Patent Document 1, there is a method of correcting astigmatism by performing Fourier transform on a captured electron microscope image and deriving the amount of astigmatism from the Fourier transform pattern. Also, as described in Patent Document 2, the irradiation beam is tilted in different directions, and each time the tilt direction is changed, a plurality of electron microscope images are recorded, and a plurality of relative transitions of the images are measured. There is a technique to do.

Japanese Patent Laid-Open No. 7-220669 JP-A-5-62628

  However, in Patent Document 1, since the Fourier transform pattern of the electron microscope image is used, the magnification of the electron microscope to be photographed is relatively high, and the sample must be an amorphous sample having no directivity that can be regarded as a weak phase object. It is difficult to get a pattern. That is, a sample must be selected to correct astigmatism, and it is difficult to apply to a low magnification.

  Patent Document 2 is a technique that can be applied even to a crystalline sample or a low magnification. However, each time the beam is tilted in different directions and the tilt direction is changed, imaging must be performed a plurality of times. For this reason, the amount of irradiation beam with respect to the sample increases, and the total imaging time is required. It is not preferable from the viewpoint of sample damage and processing time.

  An object of the present invention relates to the automatic correction of astigmatism without depending on the measurement magnification and the sample, and by reducing the irradiation beam as much as possible to reduce the sample damage.

  The present invention acquires an image with an electron microscope (first image), and generates an image that is moved by a minute amount (for example, about 0.01 to 10 pixels) set by a magnification or a sample based on the first image ( Second image), the astigmatism of the first image using the cross-correlation value obtained by performing the cross-correlation process between the first image and the second image and changing the moving direction of the minute amount as a variable. It relates to deriving the direction and amount of aberration.

  According to the present invention, basically, only one image can be captured with an electron microscope, so that sample damage can be reduced and processing can be performed in a short time. Furthermore, since the astigmatism amount and direction can be derived from the acquired image, astigmatism can be corrected without depending on the imaging magnification and the sample, that is, without being restricted by the observation conditions.

  In this embodiment, a photographing means for acquiring an electron microscope image, a means for storing the obtained electron microscope image as first image data, a means for generating an image moved by a minute amount based on the image data, The means for storing the image as the second image data, the cross-correlation process between the first image and the second image, and the cross-correlation value obtained by changing the moving direction of the minute amount as a variable An electron microscope is disclosed that includes a calculation means for deriving the direction and amount of astigmatism of the first image.

  Also, an astigmatism evaluation method in an electron microscope is a method of storing an electron microscope image acquired by a photographing unit as first image data, and generating a second amount generated by moving a minute amount based on the image data. The first image is stored as image data, the cross-correlation process between the first image and the second image is performed by the calculation means, and the cross-correlation value obtained by changing the moving direction of the minute amount is used as a variable. Discloses a method for deriving the direction and amount of astigmatism of the lens.

  In addition, the present invention discloses that astigmatism is corrected by controlling the photographing means based on the derived direction and amount of astigmatism.

  Moreover, it discloses that the minute amount corresponds to the magnification of an electron microscope.

  Also disclosed is an input means for setting the minute amount.

  Also disclosed is a display means for displaying the derived direction and amount of astigmatism.

  Also disclosed is a display means capable of displaying correction values in the X and Y directions when correcting astigmatism.

  Further, it is disclosed that the direction and the amount of astigmatism are expressed in a polar coordinate system, the relational expression is approximated by an elliptic curve, and the ratio between the major axis and the minor axis and the axial direction are calculated.

  Further, it is disclosed that a relational expression between the direction and amount of astigmatism and correction values in the X and Y directions is obtained in advance, and correction values in the X and Y directions are derived from the relational expressions.

  Hereinafter, the above and other novel features and effects will be described with reference to the drawings. The drawings are used for understanding the invention and do not reduce the scope of rights.

  Embodiments of an electron microscope and an astigmatism correction method will be described.

  In this embodiment, one image taken with an electron microscope is used as a first image, and after copying the first image, an image shifted by a small pixel in an arbitrary direction is used as a second image. A cross-correlation function between the image and the second image is derived. The correlation function value derived by changing the minute pixel shift direction (angle) by a known amount is stored for each angle, and the angle and the value are expressed in a polar coordinate system. The astigmatism of the electron microscope is automatically adjusted by calculating the difference and direction.

  Since the present embodiment uses an image shift of one pixel or less, the following equation is used for the image shift.

  Equation (1) represents the phase spectrum of the movement vector d. Here, k is a coordinate in the frequency space region. The image shift method will be described below.

  The image can be considered as a two-dimensional wave with the density value as an amplitude. Therefore, the image is Fourier transformed to obtain the amplitude spectrum and the phase spectrum. After that, the phase of the movement vector d is added to each frequency in the phase spectrum, and inverse Fourier transform is performed, whereby the image can be shifted by the movement vector d. Since this method shifts the image by changing the phase, the image can be shifted by one pixel or less.

  Next, in this embodiment, the value of the cross-correlation function between the first image and the second image is calculated and stored for each angle, and the value and the angle are expressed in a polar coordinate system, which is elliptically approximated The astigmatism of the electron microscope is automatically adjusted by calculating the difference between the major axis and the minor axis and the direction.

  Before describing the automatic correction of astigmatism, the image resolution and the value of the cross-correlation function described above will be described.

The autocorrelation function R _ACF of the image I can be expressed as follows.

Here, Ft represents a Fourier transform, and Ft ⁻¹ represents an inverse Fourier transform. * Represents conjugate. As shown in the equation (2), since the autocorrelation function reflects the frequency characteristic of the image I, the wider the band, the higher the frequency component is reflected in the autocorrelation function given by the inverse transformation. The correlation function itself also has a high frequency component. That is, it can be seen that the peak portion is sharp and the diameter is small.

  Further, the Fourier transform of the image I is represented by the following equation, where P (r) is a point spread function (PSF).

Here, it is assumed that O (r) is an ideal sample that is uniform over a sufficiently wide band. Since Ft [O (r)] has a spectrum in the entire frequency band and can be approximated as all 1, I (r) is approximately P (r), and the autocorrelation function R _ACF (r) can be expressed as follows.

  From equation (4), it can be seen that the autocorrelation function becomes an autocorrelation function of a point spread function that determines the resolution, and the resolution can be described from the autocorrelation function. That is, it can be said that the sharper the peak portion of the autocorrelation function and the smaller the diameter, the better the resolution.

  The peak portion of the cross-correlation function is calculated from the cross-correlation function of the first image and the second image. Since the second image is an image obtained by shifting the first image, the first image The cross-correlation function of the second image is an autocorrelation function with a shifted peak position. Here, when the shift amount of the second image is less than one pixel, the peak position is not shifted on the digital image. However, since the shift is analog, the value of the autocorrelation function does not become 1, and the value slightly decreases. The amount of change in this value represents the sensitivity of the peak portion of the autocorrelation function. A schematic diagram is shown in FIG.

  Since the resolution can be described by the sharpness of the peak portion of the autocorrelation function, if the shift amount for deriving the second image is the same, the amount of change in the peak portion of the cross-correlation function is the same if the resolution is the same. Become. That is, it can be seen that when the resolutions are different, the amount of change in the peak portion of the cross-correlation function differs if the shift amount for deriving the second image is the same. Also, when the resolution is good, that is, when the autocorrelation function of the point spread function is steep, the change of the peak portion of the cross correlation function becomes larger, and when the resolution is bad, that is, the autocorrelation function of the point spread function is When it is gentle, there is less change. That is, it can be said that the peak portion of the correlation function relatively measures the sensitivity of the peak portion of the autocorrelation function of the point spread function.

  As described above, the image resolution can be relatively measured by measuring the peak portion of the cross-correlation function.

  With reference to FIG. 1, an outline of an electron microscope and an astigmatism automatic correction method in this embodiment will be described.

  As shown in FIG. 1, the electron microscope according to this embodiment forms an image of an irradiation lens system 1 for irradiating a sample 4 with an electron beam, an astigmatism correction coil 3, and an electron beam transmitted through the sample 4. An image detection unit 5 that detects an image magnified by the lens system 10, a computer 6 that performs various arithmetic control processes, an arithmetic device 9 inside the computer, a storage device 11 that stores data, a computer 6, and a microprocessor 2. Communication interfaces 14a and 14b that communicate with each other, a microprocessor 2 that transmits a control signal via the bus 15, a DAC 12 that performs digital-analog conversion on a signal output from the microprocessor 2, and a signal output from the DAC 12 A power supply 13 for amplification and output to the astigmatism correction coil 3, an input device 7 for inputting parameters, and an image are output. And an output device 8 for.

  The computer 6 duplicates the first image detected by the image detection unit 5 at an appropriately set focal position, changes the phase of the duplicate image, and shifts the pixel by a known amount Δd in a certain direction θ. A second image is created and the value of the cross-correlation function between the first image and the second image is stored along with an angle θ in a certain direction. Similarly, the shift direction is further changed by the known angle Δθ, and the process is repeated until the shift direction θ reaches 180 °. An arithmetic unit 9 that approximates an ellipse from the value of the cross-correlation function thus obtained and the angle in the shift direction, calculates the difference between the major axis and the minor axis of the ellipse, and derives the astigmatism correction amount. Is equipped. Further, the result calculated by the arithmetic unit 9 is sent as correction amount information to the microprocessor 2 from the communication interfaces 14a and 14b via the bus 15. Next, the value of the correction amount is input from the microprocessor 2 to the DAC 12 via the bus 15, amplified by the power supply 13, and then output to the astigmatism correction coil 3. Astigmatism is automatically corrected by the above steps, and an image in which astigmatism is corrected can be obtained from the image detector 5.

  The above content will be described in detail below.

  In this embodiment, the direction and amount of astigmatism are derived from an image, and the astigmatism is automatically corrected using the result. This will be described below with reference to the flowchart of FIG.

First, the astigmatism correction current values I _x and I _y of the astigmatism correction coil 3 are reset (step 101). Subsequently, using the input device 7, the image shift amount Δd and the image shift direction (angle) θ, the astigmatism correction step angle Δθ, which are parameters necessary for the present embodiment, and Flag = false for calibration processing are used. Is set (step 102). At this time, Δd is an amount set by the magnification and the sample (for example, about 0.01 to 10 pixels). Here, when the astigmatism correction step angle is increased, the processing speed can be improved, but the correction accuracy is lowered. When the astigmatism correction step angle is reduced, the correction accuracy is improved but the processing speed is reduced. have.

Subsequently, in step 103, the image obtained by the image detection unit 5 is defined as a first image F ₁ (x, y), and each parameter (image shift amount: Δd pixel, image A second image G ₁ (x, y) that has undergone image shift by phase change with reference to shift direction: θ and astigmatism correction step angle: Δθ) is derived by the arithmetic unit 9 and stored in the storage unit 11. (Step 104).

In step 105, the value R _spscm of the cross-correlation function between the first image F ₁ (x, y) and the second image G ₁ (x, y) is calculated by the arithmetic unit 9, and stored for each θ in step 106. Store in device 11. At this time, the second image G ₁ (x, y) is deleted. Next, a determination of θ ≧ 180 ° is performed (step 107). If this determination is not satisfied, θ is set to θ = θ + Δθ (step 108), the image shift amount Δd and the image shift direction θ are set again, and a new one is set. A second image G ₁ (x, y) is derived (step 104).

Subsequently, the cross-correlation function value R _spscm between the first image F ₁ (x, y) and the second image G ₁ (x, y) is calculated for each θ by the arithmetic unit 9 and stored in the storage unit 11. . This is repeated until the condition of θ ≧ 180 ° is satisfied (steps 104 to 108). If the determination is satisfied, the cross-correlation function values R _spscm and θ stored for each θ in the storage device 11 are expressed in a polar coordinate system, and elliptic approximation is performed (step 109). As an example, an ellipse approximated in FIG. 4 is shown.

In step 110, from the ellipse calculated by the arithmetic unit 9, the long-axis cross-correlation function value L ₁ , the short-axis cross-correlation function value S _1, and their difference (L ₁ −S ₁ ) and direction ( An angle) θ _L is calculated by the arithmetic unit 9 and stored in the storage unit 11.

Next, in step 111, the calibration processing flag is determined. When the calibration processing flag Flag = false, the following calibration processing is performed. First, astigmatism correction currents I _AX and I _AY are set in the astigmatism correction coil 3 so as to have magnetic forces F _SX and F _SY of an angle (θ _L + α) with a known intensity as shown in FIG. Steps 112 and 113) and Flag = true for the calibration processing flag (Step 114). At this time, the angle α is an angle formed by the coordinates of the astigmatism correction coil 3 and the coordinates of the image obtained by the image detection unit 5. The magnetic forces F _SX and F _SY are derived by synthesizing the magnetic forces F _xx , F _xy , F _yx , and F _yy generated by the astigmatism correction coil 3.

Subsequently, in step 115, the image shift direction θ is set to θ = 0 °, and the third image F ₂ (x, y) is obtained again by the image detection unit 5 and stored in the storage device 11 (step 103). ).

Next, with reference to the image shift amount Δd and the image shift direction θ, a fourth image G ₂ (x, y) that has undergone image shift due to phase change is derived (step 104). Thereafter, the cross-correlation function value R _spscm between the third image F ₂ (x, y) and the fourth image is calculated by the calculation device 9 and stored in the storage device 11 for each θ (steps 105 and 106). At this time, the fourth image G ₂ (x, y) is erased.

Next, it is determined that θ ≧ 180 ° (step 107). If this determination is not satisfied, θ is set to θ = θ + Δθ (step 108), the image shift amount Δd and the image shift direction θ are set again, and the fourth image G ₂ (x, y) is derived (step). 104). Subsequently, the cross-correlation function value R _spscm between the third image F ₂ (x, y) and the fourth image G ₂ (x, y) is calculated by the arithmetic unit 9 and stored in the storage unit 11 for each θ. (Steps 105 and 106). This is repeated until the condition of θ ≧ 180 ° is satisfied (steps 104 to 108). When the above determination is satisfied, the cross-correlation function values R _spscm and θ stored for each θ in the storage device 11 are expressed in a polar coordinate system, and elliptic approximation is performed.

From the ellipse calculated by the arithmetic unit 9, the long-axis cross-correlation function value L ₂ , the short-axis cross-correlation function value S ₂ and the difference (L ₂ −S ₂ ) are calculated and stored in the storage device 11. And remember.

として、演算装置９にて算出された非点収差を補正する為に必要な非点収差補正電流値Ｉ_x，Ｉ_yを以下の式より算出する（ステップ１１６）。 The calibration value for the amount of astigmatism change for known currents I _AX and I _AY

Astigmatism correction current values I _x and I _y necessary for correcting the astigmatism calculated by the arithmetic unit 9 are calculated from the following equations (step 116).

The astigmatism correction current values I _x and I _y calculated by the arithmetic unit 9 are set as the current values of the astigmatism correction coil 3 to correct astigmatism (step 117).

In the above description, the calibration values for deriving the astigmatism correction current values I _x and I _y are calculated and used, but the calibration values vary depending on the observation conditions (magnification, acceleration voltage, etc.) of the electron microscope. The calibration value may be measured in advance under each condition and stored in the storage device 11. In this case, the steps for calculating the calibration value (steps 103 to 115) can be omitted, and the processing time for automatic astigmatism correction can be greatly shortened.

  In the above description, the transmission electron microscope has been described as an example. However, since the present embodiment derives the direction and amount of astigmatism from the difference in resolution of the captured image, it is not limited to the transmission electron microscope. The present invention can also be applied to a scanning electron microscope (SEM), a scanning transmission electron microscope (STEM), and a focused ion beam processing observation apparatus (FIB).

Schematic which shows the structure of the electron microscope in an Example. 9 is a flowchart for explaining an astigmatism automatic correction procedure using a cross-correlation function value between a captured image and an image obtained by shifting the captured image by a minute pixel in the transmission electron microscope of the present example. The schematic diagram explaining the sensitivity measurement method of an autocorrelation function from the value of an autocorrelation function. The schematic diagram which carried out the ellipse approximation of the value of a cross correlation function, and the image shift direction. The schematic diagram explaining the method of applying arbitrary magnetic fields to arbitrary directions.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Irradiation lens system 2 Microprocessor 3 Astigmatism correction coil 4 Sample 5 Image detection part 6 Computer 7 Input device 8 Output device 9 Arithmetic device 10 Imaging lens system 11 Storage device 12 DAC (digital-analog converter)
13 Power supply 14a, 14b Communication interface 15 Bus

Claims (12)

An irradiation lens system for irradiating the sample with an electron beam;
An astigmatism corrector;
An image detector for detecting an electron beam transmitted through the sample;
A computer that executes arithmetic control processing;
In an electron microscope equipped with
The astigmatism based on the shift direction dependency information of the correlation value between the first image acquired under the predetermined conditions of the irradiation lens system and the second image obtained by shifting the first image by a predetermined distance. Determining the astigmatism correction direction and correction amount of the electron beam by the corrector;
An electron microscope. The electron microscope according to claim 1,
The computer changes a correlation value of the second image obtained by shifting the first image and the first image by a predetermined distance, and changes the direction of the shift,
An electron microscope characterized by acquiring information on dependency of the correlation value in the shift direction. The electron microscope according to claim 1,
The computer performs Fourier transform on the first image, adds a frequency corresponding to the shift amount of the predetermined distance to the phase spectrum of the Fourier-transformed data, and performs Fourier inverse transform to obtain the second image. An electron microscope. The electron microscope according to claim 1,
An electron microscope comprising input means for setting the predetermined distance. The electron microscope according to claim 1,
An electron microscope comprising display means for displaying a correction direction and correction amount of the electron beam. The electron microscope according to claim 1,
An electron microscope characterized in that the correlation value is expressed in a polar coordinate system, and the relationship between the shift directions of the correlation value is approximated by an elliptic curve to calculate the difference between the major axis and the minor axis and the axial direction. The electron microscope according to claim 1,
The relationship between the signal amount of the astigmatism corrector and the amount of change in astigmatism is stored in advance, and based on the relationship,
To calculate the amount of signal that corrects astigmatism obtained from information on the dependency of the correlation value on the shift direction
An electron microscope. An astigmatism correction method in an electron microscope,
The electron microscope includes an astigmatism corrector,
The astigmatism based on the shift direction dependency information of the correlation value between the first image acquired under the predetermined conditions of the irradiation lens system and the second image obtained by shifting the first image by a predetermined distance. Determining the astigmatism correction direction and correction amount of the electron beam by the corrector;
Astigmatism correction method in an electron microscope characterized by the above. The astigmatism correction method for an electron microscope according to claim 8,
By changing the direction of the shift of the correlation value of the second image obtained by shifting the first image and the first image by a predetermined distance,
Astigmatism correction method in an electron microscope characterized in that information on dependency of the correlation value in the shift direction is acquired. The astigmatism correction method for an electron microscope according to claim 8,
The first image is subjected to Fourier transform, and a frequency corresponding to the shift amount of the predetermined distance is added to the phase spectrum of the Fourier-transformed data to perform Fourier inverse transform to obtain the second image. Astigmatism correction method in an electron microscope. The astigmatism correction method for an electron microscope according to claim 8,
Astigmatism in an electron microscope characterized in that the correlation value is expressed in a polar coordinate system, and the relationship between the shift direction of the correlation value is approximated by an elliptic curve to calculate the difference between the major axis and the minor axis and the axial direction thereof. Aberration correction method. The astigmatism correction method for an electron microscope according to claim 8,
The relationship between the signal amount of the astigmatism corrector and the amount of change in astigmatism is stored in advance, and based on the relationship,
To calculate the amount of signal that corrects astigmatism obtained from information on the dependency of the correlation value on the shift direction
Astigmatism correction method in an electron microscope characterized by the above.

Images