JPH11219680A - Converged ion beam machining method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To satisfy required high machining position accuracy by setting a machining region with a sample display image capable of changing the image magnifying power ratio between the X direction and Y direction of the image on a display. SOLUTION: A sample is guided into the sample chamber of a focused ion beam(FIB) device, and an image of a low-magnifying power scanning ion microscope(SIM) is acquired. A sample stage is moved and rotated so that four sides of rectangular patterns (a), (b), (d), (e) are made nearly parallel with the X-axis and Y-axis of the SIM image, a SIM image is acquired again, then the drift of the rotation angle θis corrected. The magnifying power ratio (m) between the X direction and Y direction is kept at 1, the image magnifying power is increased four times, a SIM image with m=4 is acquired, then the pattern (d) is obtained. Even when a machining region setting mark and a machining region are separated and they are not included in the visual field of the high-magnifying power SIM image, the magnifying power in the Y direction only is reduced to 1/4 times, thus the opposite end sections of the patterns (b), (d), (e) are included in the same visual field, and the set position accuracy in the X direction is made four times the accuracy in the Y direction in the machining region (f).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention utilizes a focused ion beam (focused ion beam: FIB for short), which is effective for section processing and fine processing of fine devices and mechanical parts.
The present invention relates to an IB processing method and an IB processing method.

[0002]

2. Description of the Related Art The prior art of the FIB processing method is disclosed in
No. 15981. It shows a cross-sectional processing example of a device by FIB. The procedure is shown below.

(1) A two-dimensional (x-y) scan is performed by a FIB on a region including a place on the device surface where a cross-sectional processing is desired, and discharged ions (secondary electrons and the like) are detected therefrom to scan ions on a sample surface. An image by a microscope (Scanning Ion Microscope: SIM for short) is acquired, recorded, and displayed on a display.

(2) A processing area is set based on the SIM image. For example, in the case of processing a cross section at the position of a contact hole of a semiconductor device, the contact hole is a processing region setting mark, and a rectangular rough processing region is manually set based on the mark.

(3) A large current FIB having a current of 2 to 5 μA
Is selected and FIB scanning is performed in the above area to perform rough processing. The cross section is formed on the side wall of the rectangular hole.

(4) A small current FIB having a current of about 400 pA
Is selected, and a SIM image is acquired again.

(5) Using the SIM image as a base, an elongated rectangular finishing area for finishing is set so that a cross section is formed at a desired position.

(6) FIB scanning is performed in the above area to finish the cross section.

In setting the processing area, it is necessary that the processing area setting mark (contact hole in the above processing example) and the processing area are included in the SIM image having the same visual field. When the processing area setting mark and the processing area are close to each other, the above-described conventional technique is effective.

[0010] However, when the mark and the processing area are separated from each other, the above-described conventional technique may not be applicable. This is because, if the image magnification is reduced so as to be in the same SIM image field of view, the processing position setting accuracy in the XY directions is also reduced, and the processing specifications may not be satisfied. In the conventional method,
Raises or lowers the SIM image magnification, but the XY magnification ratio is always 1.
It is.

[0011]

As an example of a processing object in which the processing area setting mark and the processing area do not enter the SIM image visual field of high magnification (where the magnification ratio m in the X and Y directions is 1), a recording / reproducing separation type is used. There is narrow track processing of magnetic head.
This is described in JP-A-5-143929. The recording / reproducing separation type magnetic head (see FIG. 3) has a recording head section 19 using an inductive element and a reproducing head section 10 using a magnetoresistive effect (abbreviated as MR) element, thereby increasing the recording density. In this case, it is necessary to manufacture a recording track having a track width of about 1 μm, which is aligned with the reproducing track (magnetic sensing part) A of the reproducing head unit 10 with high positional accuracy.

The current manufacturing process is a photolithography process, and the recording track is wide, for example, 10 μm. When the width of the recording track is reduced to about 1 μm by FIB processing, it is necessary to process the recording track in accordance with the position of the magnetic sensing part A of the reproducing head unit 10. Here, rectangle 2 in the figure
Reference numerals 2 and 23 are processing areas.

However, the magnetic poles 20, 21 of the recording head are
Normally, the distance from the reproduction track A is about 10 μm
If the IM image magnification is too high, both recording and reproduction tracks cannot be observed in the same field of view. Therefore, the magnification has the highest limit. On the other hand, in the MR head, the thickness of the lead conductors 14 and 15 is usually reduced to about 100 nm, and the ends 14a and 15a of the lead conductors 14 and 15 are tapered, so that the position of the reproduction track A can be accurately determined. In order to grasp the above, the magnification of the SIM image must be considerably increased.

For the above reasons, it is extremely difficult for the conventional FIB processing method to perform the processing of narrowing the width of the recording track by accurately grasping the position of the reproduction track in the recording / reproducing separation type magnetic head. This difficulty becomes more remarkable as the recording density of the reproducing head is increased. For example, when the reproduction track width is reduced to about 1 μm, about 0.2 μm is obtained.
Although it is necessary to process the recording track with a positioning accuracy of 1 μm, this positioning accuracy can hardly be obtained by the conventional FIB processing method.

The above-mentioned Japanese Patent Application Laid-Open No. 5-143929 discloses that in the above-mentioned separated recording / reproducing magnetic head, apart from a reproduction track which is a mark for setting a processing area, the same field of view is shown from the side of the medium sliding surface indicating the position. The invention is characterized in that another mark observable in a SIM image is provided in a reproducing head portion having a processing area. According to the present invention, another mark is newly added. For example, when such another mark is formed in a manufacturing process, there is a disadvantage that the process becomes more complicated than before and causes a cost increase.

Another published patent related to this technique is disclosed in Japanese Patent Laid-Open No. 5-198282. If the processing area is large,
In a SIM image with a low magnification in order to include the sample surface image including the processing region in one SIM image visual field, the processing region cannot be positioned with an accuracy higher than the SIM image resolution as described above.

This patent discloses an invention in which a plurality of SIM images having different magnifications are displayed on a display in association with each other, and the setting position accuracy is enhanced by using a high-magnification SIM image for setting a processing area. It is. According to the present invention, the processing area can be set with high positional accuracy, but since a plurality of SIM images having different magnifications are displayed on the display in association with each other, the control software becomes large-scale and the display response is slow. is there.

An object of the present invention is to provide a processing area setting mark and a processing area required to have a high processing position accuracy in one of the X and Y directions.
An object of the present invention is to provide a FIB processing method that satisfies a required high processing position accuracy even when the image does not fall within the IM image visual field (however, the magnification ratio m in the X and Y directions is 1). However, the setting of the processing area in accordance with the mark is a visual setting on the SIM image, and the processing response of the SIM image is made as fast as possible to withstand practical use.

[0019]

In order to increase the processing position accuracy, it is necessary to increase the number of pixels (pixels) constituting an image as much as possible, but the image processing time becomes long. Therefore, a means for solving the above problem is to limit the number of pixels of the SIM image to a certain extent, and to increase the number of pixels in the SIM image including a plurality of corner points that define a processing area that can be approximated by a processing mark or a rectangle. The purpose is to increase the SIM image magnification in the direction where the processing position accuracy is required as high as possible. There are two specific means.
The method is referred to as “variable XY magnification ratio method” and “SIM image division method”.

First, in the first "XY magnification ratio variable method" of the SIM image, the scanning direction of the beam is rotated so that the processing mark and the processing area are arranged in the X or Y direction of the SIM image, and on the display. The image display magnification in the X and Y directions of the image is independently changed, so that the processing mark and the processing area have the same field of view.
This is to fill the inside of the IM image frame. As a result, the set position accuracy of the processing region in the X or Y direction with an increased image magnification can be improved.

In the second "SIM image dividing method", an image frame of a SIM image serving as a setting base of a processing area is divided into two or four images having the same magnification, and these divided frames are formed with processing marks or rectangles. A high-magnification SIM image including a part of a plurality of corner points defining a processing region that can be approximated is displayed, and the processing region is set. The divided image can be scrolled in the X and Y directions, so that all of a plurality of corner points that define a processing area that can be approximated by a processing mark or a rectangle can be seen, and the processing area can be positioned with high accuracy.

That is, the processing area is set using the SIM image displayed on the display. For example, this SIM image is calibrated at 512 × 512 pixels, and the FIB diameter d is 0.1 μm.
m, FIB XY scanning range L on the sample is 51.2 μm
Assuming a rectangle of 51.2 μm, the size ΔP of one pixel on the sample corresponds to exactly 0.1 μm square. It is assumed that the set position accuracy ΔS of the processing area is determined by the larger one of d and ΔP for simplification. If the XY scanning magnification is the same as in the prior art, dx and d which are the X and Y direction components of d
Both y are equal.

In the "SIM image magnification XY independent variable method" of the first means of the present invention, the magnification only in the Y direction is, for example, 1 /
4, Ly = 409.6 μm, ΔPy = 0.4 μ
m, and ΔSy = 0.4 μm. Where ΔSx
Can be maintained at 0.1 μm. In this way, even if the processing mark and the processing area are separated by a maximum of 409.6 μm in the Y direction, processing can be performed while maintaining ΔSx = 0.1 μm. However, ΔSy = 0.4 μm. In the conventional method, in order to obtain the same XY magnification, ΔSx = ΔSy = 0.4 μm
And big.

On the other hand, in the "SIM image segmentation method" of the second means, a plurality of corner points defining a machining area which can be approximated by machining marks or rectangles are divided into two or four SIM segments.
In order to capture an image, the image magnification can be set higher than that of a SIM image in which all of a plurality of corner points that define a processing mark and a processing area are captured.

That is, the processing position can be set with high accuracy. First, "SIM image magnification XY independent variable method" of the first means
In the same manner as described above, before setting the processing area, the rotational deviation correction is performed when displaying the already acquired and recorded SIM image data so that the processing mark and the corner point of the processing area are aligned in the X or Y direction of the SIM image. It is important to put on.

As a result, the processing mark and the corner point of the processing area can be searched only by the X or Y scroll of the XY scroll of each divided SIM image, and the usability is greatly improved.

[0027]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is an explanatory diagram of a FIB processing method using the “XY magnification ratio variable method” of a SIM image of the first means. FIG.
(A) shows the sample surface of the device, and there are rectangular patterns a, b, d and e and a circular pattern c on the surface. The four sides of the rectangular pattern are parallel to each other. The processing area is a rectangular pattern f that cuts out opposite ends of the patterns d and e.

The two sides at both ends of the processing pattern f are parallel two sides rising from the two sides at both ends of the distant rectangular pattern b.
This is a processing example required to accurately match straight lines (broken lines) j and k. Specifically, the degree of coincidence is a processing example that is obtained with four times higher accuracy than the orthogonal direction.
Hereinafter, the processing procedure will be described step by step. However,
The SIM image frame on the display has XY coordinates, and the image display uses image data of 512 × 512 pixels.

Acquisition of SIM Image Data A sample is introduced into the sample chamber of the FIB apparatus, and a low-magnification SIM image is acquired. The sample stage is moved and rotated so that the four sides of the rectangular patterns a, b, d and e are approximately parallel to the XY axes of the SIM image on the display, and the SIM image acquired again is shown in FIG. . However, the SIM image is obtained by FIB scanning corresponding to 4096 × 4096 pixels, and the SIM image signal is stored in the image memory 1 (4096).
× 4096). In addition, these image data are
A SIM image is displayed by compressing or thinning to 12 × 512.

Rotation Deviation Correction The four sides of the rectangular patterns a, b, d and e are still slightly deviated by the rotation angle θ from the XY axes of the SIM image on the display, and are corrected. The relationship between the xy coordinates on the sample surface and the XY coordinates (the coordinates of the SIM image) on the display is as follows, and the correction is performed using this conversion formula.

[0031]

X = x cos θ−y sin θ (1a) Y = x sin θ + y cos θ (1b) or

[0032]

X = X cos θ + Y sin θ (2a) y = −X sin θ + Y cos θ (2b) In the display, x is calculated from the XY coordinate value corresponding to the pixel having 512 × 512 pixels by using Expression (2).
The y-converted coordinate value is calculated (θ is known), and then the SIM image signal of the coordinate value is calculated from the SIM image signal data in the image memory 1 by interpolation or extrapolation, and the result is stored in the image memory 2. And displays the image on a display.

Display of SIM Image (m = arbitrary value) and Setting of Processing Area When the image magnification is increased four times while maintaining the magnification ratio m in the XY directions at 1, the SIM image shown in FIG. Only the pattern b is included in the field of view, and the processing area f cannot be set. Therefore, FIG.
FIG. 1 (d) is obtained by changing from (c) to a SIM image with m = 4.

As a result, the magnification in only the Y direction is reduced to 1/4, and the pattern b and the opposite ends of the patterns d and e are in the same field of view, so that the processing pattern f can be set. In this case, the set position accuracy in the X direction can be four times higher than the Y direction. Expression for arbitrary m value
The conversion equations corresponding to (1) and (2) are given by the following equations (3) and (4), respectively.

[0035]

X = xcosθ−ysinθ (3a) Y = (xsinθ + ycosθ) / m (3b) or

[0036]

X = Xcosθ + mYsinθ (4a) y = −Xsinθ + mYcosθ (4b) In the display, the number of pixels is 512 as in the case of
An xy-converted coordinate value is calculated from the XY coordinate value corresponding to this pixel of × 512 using equation (4) (m and θ are known), and then a SIM image signal of the coordinate value is stored in the SIM of the image memory 1. The image signal data is calculated by interpolation or extrapolation, the result is stored in the image memory 2, and the image is displayed on a display. However, m and θ are large and xy
When the converted coordinate value greatly deviates from the xy coordinate area of the image memory 1, the number of pixels is 512 × by the xy converted coordinate value again.
It is necessary to acquire an image by performing 512 FIB scans and store the result in the image memories 1 and 2.

Since the reacquisition time of the SIM image is almost proportional to the number of pixels, the initial number of pixels is 4096 × 40.
It can be obtained in about 1/64 times shorter time than in the case of 96. In the future, when the m and θ are slightly changed, the SIM image display of the memory 1 having 512 × 512 pixels is required for the SIM image display.
What is necessary is just to re-convert and display using the image signal data.
There is no need to rescan B. This can significantly reduce sample damage due to FIB scanning.

By setting the processing area in the above order, the position in the X-axis direction of the processing area f far from the processing mark pattern b can be set to be four times more accurate than the Y-axis direction, and the FIB processing can be performed with high accuracy. I was able to.

Next, an embodiment of the "SIM image division method" of the second means will be described with reference to FIG. 2A shows the maximum SIM image frame and the entire SIM image using the same sample as that of FIG. 1A, and FIGS. 2B and 2C show the two- and four-part image frames and the divided SIM image, respectively. It is. Here, the image magnification of the divided image is set to four times that of the whole image. In the two-divided image frame shown in FIG. 2B, the lowermost X scroll q moves the two upper and lower SIM images in conjunction with each other. Similarly, upper and lower two Y scrolls r ₁ and r at the right end in the four-divided image frame of FIG.
Numeral ₂ is for moving the left and right SIM images in association with each other.

Also, by dragging the boundaries of these divided image frames with a mouse, the boundaries can also be moved vertically and horizontally. In the present embodiment, the first whole SIM image is recorded in the image memory with the number of pixels of 4096 × 4096, and the image data of the 512 × 512 area of the specific portion in the 4096 × 4096 area of the image memory is stored in one of the four divided image frames. Displayed in one image frame.

If the SIM image magnification is eight times that of the whole SIM image, a 16 × SIM image is displayed by displaying image data of 256 × 256 area in the same frame. If the processing area f and the processing mark pattern b are not parallel to the Y-axis, the SI from the image memory is
When the image is displayed on the M image frame, the image is displayed after performing a rotation deviation correction.

Thus, both end positions in the X direction of the processing area can be set with high accuracy. However, when this correction rotation angle is large and a SIM image is reacquired, it is necessary to perform FIB scanning corresponding to the number of pixels of 4096 × 4096 as in the first case, and the “XY magnification” of the SIM image of the first means is required. This is 64 times more than the number of pixels 512 × 512 in the case of reacquisition by the “variable ratio method”, and the time is about 64 times longer.

FIG. 4 is a configuration diagram of the FIB column of the present invention. The ions emitted from the liquid metal ion source 100 are focused on the FIB by the condenser lens 101 and the objective lens 106.
The light is irradiated onto the sample 91. The beam acceleration voltage is 30 kV. Variable aperture 10 between lenses
2, Aligner / Stigmer 103, Blanker 10
4, a deflector 105 is arranged.

Variable aperture 102, blanker 10
An aperture driving unit 102a, a blanking amplifier 104a, and a deflection control unit 105a are respectively attached to the deflector 4 and the deflector 105. The sample 91 is on the stage 10
8 to move. The stages are x, y, z, tilt,
It is controlled in five directions of rotation.

Secondary electrons generated from the sample 91 by the FIB irradiation are detected by the secondary electron detector 109.
The secondary electron signal from the secondary electron detector 109 is A / D-converted and taken into the image memory of the computer 110 in synchronization with the deflection control of the FIB, thereby displaying the display (CR).
T) A SIM image can be displayed on 110a.

By specifying a processing area with this SIM image as a base and performing FIB scanning while limiting only that part, local processing at a desired position can be performed. In particular, the rotational displacement correction important in the present invention, the independent setting of the beam scanning magnification in the x and y directions on the sample with respect to the FIB scanning reference signal, and the SI based on the set magnification ratio in the X and Y directions of the SIM image.
The computer 110 controls the M image display and the like, and the deflection controller 105a sets the deflection voltage in consideration of the rotational displacement correction and the beam scanning magnification.

A gas source 1 for forming an FIB-induced deposition film
Gas containing metal generated from 07 (for example, W (CO) ₆ gas) is guided to the vicinity of the sample irradiation part of FIB 1 by gas nozzle 92. The gas source 107 is controlled by a gas source control unit 107a. The computer 110 drives the aperture of the variable aperture 102 via the system bus 111, turns on / off the FIB irradiation by the blanker 104, scans the FIB deflection by the deflector 105, detects the signal from the secondary electron detector 109, moves the stage 108, and Each control such as gas supply is performed.

Finally, the first method "SI
The characteristics of the “M image magnification XY independent variable method” and the second method “SIM image division method” are shown in Comparative Table 1.

[0049]

[Table 1]

[0050]

According to the present invention, even if the processing mark is far from the processing area, the magnification of the SIM image including a plurality of corner points defining the processing area which can be approximated by the processing mark or rectangle can be increased. The position of the processing area can be set with high precision, and the processing can also be performed with high precision. In addition, since the number of pixels of the SIM image is not as large as the TV level, the image processing can be performed at the TV level.

[Brief description of the drawings]

FIGS. 1 (a) and 1 (b) to 1 (d) show an F image using a XY magnification ratio variable method for a SIM image according to a first embodiment of the present invention.
FIG. 2 is a plan view of a sample surface of a device for explaining an IB processing method and a plan view showing SIM images of respective rectangular regions g, h, and m in FIG.

FIGS. 2 (a) and (b) to (c) show an entire SIM image frame and a SIM image for explaining an FIB processing method using a “SIM image division method” of a second means of the embodiment of the present invention. The top view and the top view which shows the SIM image frame and the 2 and 4 division SIM image frame.

FIG. 3 is an explanatory view of a read / write separation type magnetic head of the present invention.

FIG. 4 is a configuration diagram of an FIB column used in an example of the present invention.

[Explanation of symbols]

Reference numeral 10: reproducing head portion (A: magnetic sensing portion), 11: lower shield layer, 12: lower insulating layer, 13: MR element, 14, 15
... lead conductors, 14a, 15a ... ends of lead conductors, 1
6, 18 ... upper insulating layer, 17 ... upper shield layer, 19 ...
Recording head part, 20, 21 ... magnetic poles of recording head, 22,
23: FIB processing area, 90: FIB, 91: sample, 1
00: ion source, 101: condenser lens, 102
... variable aperture, 102a ... diaphragm drive unit, 103 ...
Aligner / Stigmer, 104 ... Blanker, 104
a: Blanking amplifier, 105: Deflector, 105
a: deflection controller, 106: objective lens, 107: gas source, 108: stage, 109: secondary electron detector, 11
0: Computer, 110a: Display (CR
T), 111 ... system bus, ae ... pattern, f ...
Processing area (rectangle), g: Corresponding area on the sample of the high-magnification SIM image in FIG. (B), h: Corresponding area on the sample of the high-magnification SIM image in FIG. (C), k, j: processing mark Pattern b
, A line connecting both ends of the X with the processing area f, θ: rotation correction angle
q, q ₁ , q ₂ ... X scroll, r, r ₁ , r ₂ .

Claims (2)

[Claims] An ion source, a focusing system for focusing ions emitted from the ion source on a sample and irradiating the sample with a beam, and applying the focused ion beam (FIB) to the sample two-dimensionally (x-
y) a deflection system for scanning, a secondary charged particle detector for detecting secondary charged particles emitted from the sample surface by irradiation of the focused ion beam, A / D conversion of a signal of the secondary charged particle detector, A display for displaying an X-Y image of the sample surface on a display by capturing the image of the sample surface in an image memory of a computer in synchronization with a scanning signal of the FIB;
FIB processing in which a processing area is designated based on a processing area setting mark on the basis of an image of the sample surface on the display as a base, and FIB scanning is performed by limiting only that area, thereby performing local processing at a desired designated position. A focused ion beam processing method, wherein a processing area is set using a sample surface image in which an image magnification ratio of an image on the display in X and Y directions can be changed. 2. An ion source, a focusing system for focusing ions emitted from the ion source on a sample and irradiating the sample with a beam, and applying the focused ion beam (FIB) two-dimensionally (x-
y) a deflection system for scanning, a secondary charged particle detector for detecting secondary charged particles emitted from the sample surface by irradiation of the focused ion beam, A / D conversion of a signal of the secondary charged particle detector, A display for displaying an image of the sample surface on an X-Y display on a display by taking in an image memory of a computer in synchronization with the deflection control of the FIB;
FIB processing in which a processing area is designated based on a processing area setting mark on the basis of an image of the sample surface on the display as a base, and FIB scanning is performed by limiting only that area, thereby performing local processing at a desired designated position. In the method, an image on the display is displayed using two or four divided images (same magnification) having a scroll function with a frame, and a processing area is set using the divided images. A focused ion beam processing method characterized by the following.