JP2002075262A - Image processing application equipment - Google Patents

Abstract

(57) [Summary] [Object] To acquire image data many times for focusing in a microscope apparatus. Once the focus has been adjusted, the focus state changes with the aging of the apparatus state during use. The present invention provides an apparatus that can automatically perform focus correction in such a case with a minimum number of image acquisitions. An apparatus stores in advance a focus parameter dependency of an image deterioration amount. First, a reference image signal of a sample is acquired in a just-focused state, and a parameter dependence characteristic of a correlation length is calculated and stored from the reference image signal and the parameter dependence of the deterioration amount. At the time of focus correction, two candidate values of the parameter shift amount are estimated from the correlation length obtained from the deteriorated image signal and the parameter dependence of the correlation length, and the focus correction is completed by changing the parameter based on the estimated value. I do.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a microscope apparatus having an image display function, and more particularly to a sample observation apparatus or a sample processing apparatus using a charged particle beam, and in particular, automatically corrects an image focus based on an image signal of the sample. The present invention relates to an image processing application device.

[0002]

2. Description of the Related Art In a microscope apparatus, a method used for automatically adjusting a focus is to change a parameter for changing a focus state on a trial basis, and to set an index of complexity of an image acquired under each parameter value. This is a method of calculating the amount of integration of the absolute value of the differential value, and determining the value of the parameter that maximizes the complexity index. For example, in a scanning electron microscope, a method is used in which a voltage or a current applied to an objective lens is changed as this parameter to select a value that maximizes the sum of absolute values of smoothed differential values of an image. ("Electron / Ion Beam Handbook (3rd Edition)"
Nikkan Kogyo Shimbun, 1998, pp. 894-896).
In this method, the amount of defocus cannot be estimated from the complexity index of the image. Therefore, it is necessary to change the parameters to various values on a trial basis, acquire an image each time, and evaluate the complexity index.

Therefore, it is necessary to read a large number of times (several tens to one hundred times) of images before finding a focused state. The image is acquired many times by so-called trial and error, and focus adjustment is repeated. Therefore, it is necessary to perform image reading many times until finding a parameter value that is in focus. In the meantime, if the state of the sample surface hardly changes (the image complexity index is an index for evaluating the focus state), it takes time, but it can be said to be a useful method.

[0004]

However, in a microscope apparatus which obtains a sample image by irradiating a charged particle beam or the like, there are many cases where the sample surface is gradually damaged by repeatedly irradiating the beam. In particular, when acquiring a scanned ion image of a sample with a focused ion beam apparatus having a function of processing the sample, the surface of the sample is scraped by the act of reading the image itself, and the surface state changes. Since the index of the complexity of the read image depends not only on the focus state but also on the state of the sample surface, the complexity index of the image is not an evaluation index reflecting only the focus state, and correct focusing is performed. You will not be able to do it.

In the scanning electron microscope apparatus, if the electron beam is continuously applied to the sample, contamination of the surface of the sample may occur, and the quality of an image may be deteriorated such as an increase in the width of a line pattern. Contamination itself is usually not desirable, and if an image is continuously acquired under conditions of electron beams that cause significant contamination, it also affects the index of the complexity of the image, making it difficult to properly focus. Become.

According to the present invention, in a microscope apparatus, after the observation of a sample image or the processing of a sample is started once the focus is adjusted, if the focus state changes due to the aging of the apparatus state during use, a minimum image is obtained. It is an object of the present invention to provide an image processing application device that can automatically perform focus correction by the number of acquisitions.

[0007]

In order to reduce the number of times of acquiring an image necessary for focus correction, it is possible to estimate how much the parameter for determining the focus of the objective lens from the image is deviated from the just focus state. It is necessary to obtain important information.

[0008] The blurred image is formed mainly by information on the surface morphology of the sample itself and information on the spread of the beam hitting the sample. Defocusing is related to the latter beam spread. Therefore, if the surface morphology of the sample is known and the manner in which the beam spreads due to the shift of the parameter for determining the focus is known, the shift amount of the parameter can be estimated from the actual blurred image.

Specifically, means for storing a reference image signal obtained from a sample or reference data calculated from the signal, and means for calculating an autocorrelation function or an autocorrelation length of a deteriorated image signal obtained from the sample A focus correction calculation unit comprising: the calculated autocorrelation function or autocorrelation distance; and estimation calculation means for calculating a deviation estimation value of the parameter value from a reference image signal or reference data collation. Means for setting and changing the parameter value based on the estimated value obtained, wherein the focus state is automatically corrected by changing the parameter value.

[0010] The focus correction calculation unit may include means for storing the parameter dependence of an image deterioration amount dependent on a focus shift, and a reference image signal obtained from the sample and the parameter dependence of the deterioration amount. Means for calculating the parameter dependence of the autocorrelation length of the image; means for calculating the autocorrelation length of the defocused degraded image signal obtained from the sample; and the autocorrelation length of the degraded image signal and the reference image. Means for comparing two parameter-dependent characteristics of the signal autocorrelation length to calculate two estimated candidate values of the parameter deviation amount, and first changing means for selecting any one of the estimated candidate values and changing the parameter value Means for calculating the autocorrelation length of the trial correction image read based on the parameter values; and the autocorrelation length of the degraded image signal and the autocorrelation of the trial correction image. Means for determining whether or not the candidate value of the shift amount of the selected parameter is appropriate for focus correction, and second changing means for changing the parameter to another candidate value when the candidate value is not suitable for focus correction; There is a feature in having.

Further, the focus correction calculation unit has means for storing a function representing a characteristic that an image signal changes according to the parameter, and a parameter shift amount between the reference image signal obtained from the sample and the stored function. To estimate. The change of the parameter for changing the focus state must be the objective lens voltage or the objective lens excitation current of the charged particle beam device. Changing the parameter that changes the focus state involves changing the height of the sample stage in the charged particle beam device. It is characterized in that the reference image signal and the deteriorated image signal are used as an image signal of a mark for detecting a sample position in the ion beam processing apparatus.

In addition, the focus parameter dependence of the image deterioration amount is an image signal acquired when the intensity of the objective lens fluctuates during the axis adjustment of the charged particle beam device.
It is characterized by having means for designating an arbitrary area on the display screen and means for storing the designated area as registration data.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described with reference to FIG. Ion beam 2 emitted from ion gun 1
Is squeezed by the electrostatic objective lens 5 and irradiates the sample 6. The lens voltage of the objective lens 5 is set by the power supply 23 through the parameter setting circuit 10. The deflection signal generated by the deflection signal generator 13 is a signal that can change the scanning range and the scanning position on the sample by the device control unit 17, and operates the electrostatic deflector 4 through the deflection amplifier 9 to 2 is scanned on the sample 6 (3 is an image shift deflector). Electrons such as secondary electrons or reflected electrons generated by the ion beam 2 incident on the sample 6 are detected by the detector 11 to become electric signals.

The electric signal is converted from an analog signal to a digital signal by the A / D converter 14 and stored in the image memory 15. The image signal of the image memory 15 is sent to the focus correction calculation unit 18 and the device control unit 17.
The electric signal emitted by the detector 11 is also sent to the display device 16 and displayed as an image. The parameter setting circuit 10 drives the objective lens with a designated focus parameter (objective lens voltage) based on signals from the device control unit 17 and the focus correction calculation unit 18.

First, a case will be described in which the amount of defocus is estimated from the autocorrelation function or autocorrelation length and the reference image signal or reference data and the parameter value is changed.

The focus correction calculating section 18 is constructed in principle as shown in FIG. Reference image signal f (x)
Means 22 for storing data reflecting the surface morphology of the sample obtained from step (b), and calculating an autocorrelation function (or autocorrelation length) reflecting the state of deterioration of the image from the blurred deteriorated image signal g (x). Means 24. Means 26 is provided for calculating an estimated value of the parameter deviation based on the relationship between the known parameter deviation and the beam spreading method from these signals. Therefore, the focus correction calculation unit 18 has a means 28 for changing the focus parameter based on the estimated parameter shift amount, and outputs to the parameter setting circuit 10 so that the focus state is corrected to the just-focus state. Will be. Although FIG. 2 shows a case where the reference image signal is stored in the device control unit 17, a method of storing f (x) in the image memory 15 may of course be used (hereinafter, FIGS. 3 and 6). 7 is also the same. Both f (x) and g (x) are read from the image memory 15).

Further, the focus correction calculation section 18 has a function of performing the calculation shown in FIG. 3, and the handling of the input signal received from the image memory 15 is distinguished by the signal received from the apparatus control section 17. That is, the image shift deflector 3 is operated through the DC amplifier 8 by a signal from the device control unit 17, and the range scanned by the ion beam 2 can be shifted. Also, by sending a signal 60 from the device control unit 17 to the stage drive circuit 12 and moving the sample stage 7, the range over which the ion beam 2 scans the sample can be changed.

When processing the sample 6 using the ion beam 2, the operator first displays the display device 1 before starting the processing.
The voltage of the objective lens 5 is adjusted to an optimum value through the device control unit 17 while looking at 6. In this state, a specific area other than the area to be processed on the sample is set as the field of view for focus correction, and the reference image signal f (x) is obtained.
Here, an example of each signal is shown in FIG. 4A shows the reference image signal f (x), and FIG. 4B shows Cf (ξ) / Cf.
4 (c) shows the degradation function h (x), FIG. 4 (d) shows the degradation image signal g (x), and FIG. 4 (e) shows Cg
It is an example of the (ξ) / Cg (0) signal.

At this time, as shown in FIG. 3, the focus correction calculating section 18 stores the previously stored FIG.
The autocorrelation length as shown in FIG. 5B is obtained from the signal of the storage means 31 of the objective lens voltage dependency (σ−ΔV) of the image deterioration amount σ as shown in FIG. 5A and the reference image signal f (x). Means 3 for calculating and storing the objective lens voltage dependence characteristic (d−ΔV)
2 is provided.

After the processing of the sample by the ion beam is started, if the focus state changes due to the aging of the apparatus state, normal processing cannot be performed. In this case, it is necessary to correct the focus. When a certain time has elapsed, the degraded image signal g (x) is read from the image memory 15 to the focus correction calculation unit 18 by a signal from the device control unit 17. In the focus correction calculating section 18, first, as shown in FIG.
3, the autocorrelation length d [g] is calculated. And FIG.
The dependence of the autocorrelation length on the objective lens voltage as shown in (c) and d [g] are checked, and the estimated candidate values ΔV1 and ΔV2 are calculated by the objective lens voltage shift amount calculating means 34. One of the values ΔV1 is temporarily adopted, and the signal for setting the objective lens voltage to (V0−ΔV) obtained by subtracting ΔV1 from the value V0 when the deteriorated image signal g (x) is obtained is the first setting. The value changing means 35 causes the focus correction calculation unit 18 to output the parameter setting circuit 10 for the objective lens.
And the objective lens voltage of the objective lens 5 becomes (V0−
ΔV1). When the change of the objective lens voltage is completed, the trial correction image signal g1 (x) is read from the image memory 15 into the focus correction calculation unit 18 by the signal from the device control unit 17.

As shown in FIG. 3, the autocorrelation length of g1 (x), that is, d [g1], is calculated by the autocorrelation length calculation means 36 in the focus correction calculation unit 18. And d [g1]
Is smaller than d [g], the determination means 37 determines whether the candidate value ΔV1 of the provisionally adopted objective lens voltage correction amount is appropriate. If inappropriate (d [g
1]> d [g]), the objective lens voltage is set to (V0−
The setting is changed by means 38 for changing the setting to ΔV2). The change signal is sent from the focus correction operation unit 18 to the parameter setting circuit 10, and the objective lens voltage of the objective lens is changed to (v0-ΔV2). With this, the focus correction is completed, and the processing on the processing area is restarted by the signal from the device control unit 17.

The dependence of the autocorrelation length obtained from the reference image signal f (x) on the objective lens voltage is stored as it is, and if the focus correction is performed after a certain processing time has elapsed again in the same sample, the deterioration will occur. If the processing after reading the image signal g (x) is repeated, the correction is completed. That is, the number of times of acquisition of an image necessary for one focus correction is two times, one for acquiring the deteriorated image signal and one for acquiring the trial correction image signal.

In the case where, for example, the sample width is small and a sample edge is included in the displayed image, or when there is a remarkable level difference on the sample surface, the entire image includes portions having different proper focus states. Therefore, it is necessary for the operator to appropriately specify an image signal area to be used for focus correction on the display screen.

FIG. 8 shows an example of a screen displaying a narrow sample. Both ends 82 of the thinned sample are visible in the display screen, and a processing pattern 84 is registered for further thinning the sample for observation with a transmission electron microscope. At this time, if an image signal used for focus correction is obtained from a region including a region outside the sample end 82, proper focus correction cannot be performed. Therefore, the apparatus is provided with a function for registering a focus correction position on the display screen, and an operator designates a sample surface portion on the display screen and pre-registers a mark image for sample position detection as in the focus correction position 86. If the reference image signal, the deteriorated image signal, and the trial correction image signal are all acquired from the registered position, appropriate correction can be performed. The registration range of the focus correction position may be a planar area in the display screen, or may be a linear segment.
It may be a dimensional area. It is also possible to designate a plurality of surface regions or line segment regions and to perform focus correction using signals of all the portions.

Next, a specific operation of the image processing application apparatus will be described with reference to FIGS. 3, 4, and 5. FIG. For example, one-dimensional data is extracted from the image in the horizontal direction (x direction), and a data function is defined as follows. f (x); reference image signal (data acquired by just focus) h (x; σ): deterioration function (function representing relative image deterioration due to focus shift) g (x); deteriorated image signal (focus is Image signal when it is shifted)

Here, h (x; σ) can be approximated by a normal distribution function with a standard deviation σ, and the degradation amount σ is a function that depends on a deviation amount ΔV (for example, an objective lens voltage deviation) of a parameter that determines a focus state. As shown in FIG.
It becomes minimum when it is 0. At this time, g (x) is approximately f
It can be expressed by the convolution integral of (x) and h (x; σ).

G (x) = f (x) * h (x; σ) = ∫f
(Α) h (x−α; σ) dα Examples of the reference image signal f (x), the deterioration function h (x), and the deteriorated image signal g (x) are described above with reference to FIGS. ),
(E) respectively.

The autocorrelation function (autocovariance function) Cg of g (x) is a variation Δg (x) from the average value of g (x).
That is, Cg (ξ) = <Δg (x) Δg (x + ξ)> using Δg (x) = g (x) − <g (x)>. This is an even function having a maximum peak at the origin ξ = 0. Then, d where Cg (d) = Cg (0) / 2 is called a correlation length. Since the correlation length d is determined when the function g (x) is determined, d is considered as a functional of g (x),
d = d [g (x)].

FIGS. 4B and 4E show f (x) respectively.
And autocorrelation functions of g (x) and correlation lengths d [f] and d [g] are shown in the figure. Deterioration function h
The correlation length increases with the deterioration of the image as represented by the convolution integral of (x). The correlation length increases as the deterioration amount σ increases.

FIG. 3 shows an example of the focus correction calculation.
As a premise of the focus correction, first, the relationship between the image deterioration amount σ and the parameter shift amount ΔV as shown in FIG. 5A is checked in advance, and this is stored (31) in the apparatus. Then, before starting the processing, the operator adjusts the processing beam to the just-focused state and acquires the reference image signal f (x). Then, according to d = d [f (x) * h (x; σ (ΔV))], the ΔV dependency d (ΔV) of the correlation length d under the reference image is shown in FIG. 5B in advance. Can be calculated as follows. When focus correction is performed after a predetermined time has elapsed, the deteriorated image signal g (x) is obtained from the same portion of the sample.
To get. The correlation length d [g (x)] is obtained, and FIG.
By comparing with the previously obtained d (ΔV) as in (c) (34 in FIG. 3), ΔV candidate values ΔV1 and ΔV2 are obtained.

With respect to the set value V0 of the parameter at the time when the signal g (x) was obtained, the set value is corrected to V0-ΔV1, and in this state, the trial correction image signal g1 (x) is obtained. After calculating the correlation length d [g1], the two are compared. If d [g1 (x)] <d [g (x)], it is determined that the parameter shift amount is correct at ΔV1, and the focus correction is completed. On the other hand, both are compared. If d [g1 (x)]> d [g (x)], it is determined that the correct deviation of the initial parameter is not ΔV1 but ΔV2, and the parameter is set to (V0−
ΔV2). This completes the focus correction.

According to the conventional auto focus method,
For one focus correction, it was necessary to obtain several tens to 100 times of images while changing focus parameters. However, according to the above means according to the present invention, the number of image signal readings for focus correction is:
Only two times of acquisition of the deteriorated image and acquisition of the trial correction image are required. Therefore, focus correction can be performed even on a microscope device or a sample having a property that the surface of the sample is damaged or deteriorated by repeatedly acquiring images. Since the focus correction can be performed with an extremely small number of image acquisitions as described above, there is an effect that contamination generated in a visual field region used for focus correction of the scanning electron microscope can be reduced.

The focus correction can also be performed by using a deterioration function instead of the deterioration amount and an autocorrelation function instead of the autocorrelation length, and comparing the functions instead of the correlation length. In this case, if it is evaluated that the deterioration function is strictly different between the underfocus state and the overfocus state, one candidate parameter value can be selected from the deteriorated image signal, and it is not necessary to acquire a trial correction image signal.

Further, since the deterioration amount can be calculated from the reference image signal and the deteriorated image signal, focus correction can be performed by comparing the deterioration amount with the reference data instead of the autocorrelation length.

FIG. 6 shows a second embodiment of the present invention. In this embodiment, an ion beam 2 is used. Beam 2
Is scanned on the sample, and an image of secondary electrons or reflected electrons is acquired as an image. In this case, the height of the sample stage is used as a parameter for changing the focus state. In this case, the focus parameter deviation amount estimation candidate value calculated by the focus correction calculation unit 18 is a sample stage height deviation amount estimation candidate value. Estimated candidate values ΔV1 and ΔV2 of the shift amount of the sample stage height from the deteriorated image signal are calculated by the focus correction calculation unit 18.
, And the value of ΔV1 is sent to the stage drive unit 12 to obtain a trial correction image signal (this is different from the embodiment of FIG. 1). By determining the autocorrelation length of the trial correction image signal, the shift amount of the stage height is determined. The height shift of the determined shift amount is controlled by a signal 62 from the device control unit 17. In this way, the amount of shift in the stage height can be returned to the just-focus state by the two image acquisitions with respect to the shift amount.

The third embodiment shown in FIG. 7 is an example in which an image processing application device is applied to a scanning electron microscope.
Basically, the difference from the embodiment of FIG. 1 is the difference between the ion gun (1) and the electron gun (19). Here, the exciting current value of the objective lens 5 is used as the focus parameter V. When the degraded image signal is taken into the focus correction calculation section 18, the focus correction calculation section 18 calculates excitation current deviation amount estimation candidate values ΔV1 and ΔV2. The trial correction image is obtained by changing the exciting current by ΔV1, and the autocorrelation length of the self image is determined to determine the amount of change in the exciting current. In this way, the excitation current can be returned to the just-focused state by two image acquisitions.

If the characteristics of the electron source change with time, it is necessary to re-examine the image deterioration characteristics as shown in FIG. A function of automatically acquiring a plurality of images using the exciting current as a parameter when the exciting current of the lens 5 is moved,
If there is a function of calculating the parameter dependence of the amount of deterioration of the image from the image data, the characteristics of the amount of deterioration can be easily checked again. By automatically registering the calculated image deterioration characteristics, the characteristics of the deterioration amount can be updated at any time, and in this case, the focus correction can be always performed according to the state of the apparatus.

[0038]

According to the present invention, the automatic focus correction can be completed in two image acquisitions without acquiring a large number of images as in a normal autofocus method. For this reason, the sample surface is not damaged due to repeated image acquisition.

[Brief description of the drawings]

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

FIG. 2 shows a principle diagram of focus correction according to the present invention.

FIG. 3 shows an example of a focus correction calculation according to the present invention.

FIG. 4 is an example of a reference image signal, a degradation function, a degradation image signal, and an autocorrelation function.

FIG. 5 shows an image deterioration amount, a parameter shift amount dependency of a correlation length, and parameters.

FIG. 6 is a configuration diagram showing another embodiment of the microscope apparatus according to the present invention.

FIG. 7 is a configuration diagram showing another embodiment of the microscope apparatus according to the present invention.

FIG. 8 shows a display example of a registered focus correction position.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Ion gun, 2 ... Ion beam, 3 ... Image shift deflector, 4 ... Deflector, 5 ... Objective lens, 6 ...
Sample 7 Stage 8 DC amplifier 9 Deflection amplifier 10 Parameter setting circuit 11 Detector
12: stage drive circuit, 13: deflection signal generator,
14 ... A / D converter, 15 ... Image memory, 16 ...
Display device, 17: device control unit, 18: focus correction calculation unit, 19: electron gun, 20: electron beam, 21 ...
Image shift coil, 22 ... deflection coil, 23 ...
Power supply, 82: Sample end in display screen, 84: Display of registered machining pattern, 86: Display of registered focus correction position

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Hiroyuki Muto 882-Chair, Ichigo-shi, Hitachinaka-shi, Ibaraki Prefecture Within the Measuring Instruments Group of Hitachi, Ltd. 2G001 AA03 AA05 BA07 BA15 CA03 FA06 GA06 GA08 GA09 HA13 JA02 JA03 JA20 2H052 AF10 AF13 AF14 5C033 MM03 UU05 UU06

Claims (8)

[Claims] A means for displaying an image of a sample on a display screen;
An image processing application device having means for changing a focus state by changing a parameter value, wherein a means for storing a reference image signal obtained from a sample or reference data calculated from the signal, and a deteriorated image signal obtained from the sample Means for calculating an autocorrelation function or an autocorrelation length of the above, and estimation calculation for calculating a deviation amount estimation value of the parameter value from a comparison between the calculated autocorrelation function or autocorrelation distance and a reference image signal or reference data. And a means for changing a parameter value based on the calculated estimated value,
An image processing application apparatus wherein a focus state is automatically corrected by changing a value of the parameter. 2. The image processing apparatus according to claim 1, wherein the focus correction calculation unit stores the parameter dependence of an image deterioration amount depending on a focus shift, and a reference image signal obtained from a sample and the deterioration amount. Means for calculating the parameter dependency of the autocorrelation length of the sample image from the parameter dependency of the sample, means for calculating the autocorrelation length of the defocused deteriorated image signal obtained from the sample, and Means for comparing the parameter dependence characteristics of the autocorrelation length and the autocorrelation length of the reference image signal to calculate two estimated candidate values of the parameter shift amount, and selecting any one of the estimated candidate values to determine the parameter value First changing means for changing, a means for calculating the autocorrelation length of the trial correction image read based on the parameter value, an autocorrelation length of the deteriorated image signal, and Means for comparing the autocorrelation length of the corrected image to determine whether or not the candidate value of the shift amount of the selected parameter is appropriate for focus correction, and changing the parameter to another candidate value when the candidate value is not suitable for focus correction. An image processing application device comprising: a second changing unit. 3. The apparatus according to claim 1, wherein the focus correction calculation unit has a unit for storing a function representing a characteristic that an image signal changes in accordance with the parameter, and stores a reference image signal obtained from a sample and the storage. An image processing application apparatus for estimating a parameter shift amount from a function obtained. 4. An apparatus according to claim 1, wherein the change of the parameter for changing the focus state is an objective lens voltage or an objective lens excitation current of the charged particle beam device. 5. An apparatus according to claim 1, wherein the change of the parameter for changing the focus state changes the height of the sample stage in the charged particle beam apparatus. 6. An image processing apparatus according to claim 1, wherein the reference image signal and the deteriorated image signal are image signals of a mark for detecting a sample position in an ion beam processing apparatus. 7. The method according to claim 2, wherein the focus parameter dependence of the image deterioration amount is an image signal obtained when the intensity of the objective lens fluctuates during the axis adjustment of the charged particle beam device. Image processing application equipment. 8. An image processing application apparatus according to claim 1, further comprising means for designating an arbitrary area on the display screen, and means for storing the designated area as registration data.