The present invention relates to a charged particle beam apparatus using a charged particle beam such as an electron beam or ion beam. More particularly, it relates to a charged particle beam apparatus and a dimension measuring method which are preferable for measuring the dimension of a specified domain with a high accuracy even if height differences exist within the field-of-view.
In charged particle beam apparatuses representative of which is a scanning electron microscope, a narrowly converged charged particle beam is scanned on a sample, thereby acquiring desired information (e.g., sample image) from the sample. In the charged particle beam apparatuses like this, implementation of the high resolution has been progressing year by year. Of these charged particle beam apparatuses, in electron beam apparatuses in particular, diffraction phenomenon of the electrons is conspicuous and dominant. As a result, a decrease in the focal depth in accompaniment with the high-resolution implementation is unavoidable in principle. Meanwhile, under the circumstances like this, it is now required to perform a high-reliability dimension measurement with a higher accuracy. However, in a situation in particular where a plurality of measurement positions accompany height differences with respect to the electron beam, it becomes difficult to perform the high-reliability dimension measurement.
Conventionally, in order to automatically perform the dimension measurement, the following method has been generally used: Namely, a condition for allowing the best focus to be achieved is found out from the entire contrast within the field-of-view including a dimension-measuring domain. Then, after setting the focus, the dimension of a predetermined domain is measured from a SEM image newly acquired based on this focusing condition. Also, as a technique for acquiring a deep focal-depth SEM image of a sample which accompanies asperities or height differences, a technique of acquiring a plurality of SEM images with different focuses has been disclosed in JP-A-2002-75263. In JP-A-2002-75263, the method has been disclosed which allows the deep focal-depth SEM image to be acquired by extracting best-focused image domains from these SEM images respectively and superimposing these image domains into the one piece of SEM image. Also, in JP-A-11-264726, the following method has been disclosed: Namely, the plurality of SEM images with the different focuses are acquired, and the dimension measurements are performed in length-measuring domains in the respective SEM images. Then, a measurement value whose variation in the measurement results acquired for the focus variations becomes the smallest is assumed as the dimension's true value.
Each of the above-described conventional techniques has the following problems: There exists a problem that, if an automatic focus adjustment is performed in a state where height differences exceeding the focal depth of an electron-optics system exist within the field-of-view, the entire focusing condition is biased toward the focus of a densely-structured domain within the field-of-view. Accordingly, it turns out that, if the dimension-measuring domain exists outside the densely-structured domain, the length measurement is performed under a condition which deviates from the appropriate focusing condition. This situation results in a decrease in the accuracy and reliability of the length measurement value. Also, influences by magnetic hysteresis of lenses make unavoidable the inconsistency between the optimum focusing condition found out by focus search and the actual optimum focusing condition. This inconsistency causes a certain extent of error to occur in the automatic focus adjustment. Consequently, in accompaniment with the decrease in the focal depth, this focusing error becomes one cause for the reliability decrease in the length measurement value.
In the technique disclosed in JP-A-2002-75263, the highest image-sharpness-degree domains detected from the respective SEM images are combined thereby to form the one piece of SEM image. At this processing step, however, there exists a possibility that the image superimposing may fail because of influences by noise. Accordingly, in some cases, it is difficult to apply this technique to the high-reliability dimension measurement. Also, in the technique disclosed in JP-A-11-264726, no consideration has been given to the focal depth of the electron-optics system. This situation requires that, in order to perform the secure dimension measurements, the large number of SEM images be acquired by varying the focus over steps which are minute more than necessary. However, when measuring a beam-damage sensitive substance such as organic material, it is desirable to complete the measurement with the smallest possible beam irradiation amount. This is because the material which is easily subject to the electron-beam damage will shrink when exposed to the large amount of beam irradiation, and because measuring the accurate dimension becomes impossible. Consequently, in the method disclosed in JP-A-11-264726, the problem of the beam damage becomes conspicuous depending on configuration materials of the sample.
Also, the variations in the dimension measurement value depending on the focus variations vary depending on a threshold value at the time of edge detection. The reason for this will be explained referring to
It is an object of the present invention to provide a charged particle beam apparatus and a dimension measuring method which allow the high-reliability dimension measurement value to be acquired under the optimum focusing condition in a dimension-measuring domain even if the height differences exist within one and the same field-of-view, or even if the focal depth is very shallow and the focusing accuracy is insufficient.
In order to accomplish the above-described object, in the present invention, a plurality of SEM images are acquired while varying the focus with a variation width which is substantially equal to value of the focal depth of the electron-optics system. Next, the image sharpness degrees of partial domains including a dimension-measuring domain of the plurality of SEM images acquired are evaluated. Moreover, a SEM image for which the image sharpness degree of the partial domain is the highest is selected, then performing the dimension measurement using this SEM image. Incidentally, if a plurality of measurement locations exist within one and the same field-of-view, corresponding partial domains are set on each measurement-location basis, then evaluating the image sharpness degrees of the images. Also, there are provided in advance an input unit for setting a range in which the beam is to be converged, and a unit for calculating the value of the focal depth of the electron-optics system. This makes it possible to acquire multi-focus SEM images in an appropriate image number.
This makes it possible to grab the images with the minimum beam irradiation amount with respect to a beam-damage sensitive sample as well. Furthermore, without resetting the focusing, the evaluation of the image sharpness degrees and the dimension measurement are performed using the plurality of SEM images themselves which are acquired under the different focusing conditions. This makes it possible to avoid the focusing error which occurs when the focusing is newly reset, thereby allowing an enhancement in the accuracy and reliability of the dimension measurement.
According to the present invention, using a high-resolution SEM with a shallow focal-depth, it becomes possible to measure the dimension of a specified domain with a high reliability and accuracy even if some extent of height differences exist within the field-of-view.
Other objects, features and advantages of the invention will become apparent from the following description of the embodiments of the invention taken in conjunction with the accompanying drawings.
Hereinafter, referring to the drawings, the explanation will be given below concerning embodiments of the present invention.
The primary electron beam 4 is scanned on the sample 10 in a two-dimensional manner by a scanning coil 9 controlled by a scanning-coil control power supply 24. A secondary signal 12 such as secondary electrons, which are generated from the sample 10 by irradiation of the primary electron beam 4, travels to a region over the objective lens 7. After that, the secondary signal 12 is separated from the primary electrons by an orthogonal-electromagnetic-field generation device 11 for separating the secondary signal, then being detected by a secondary-signal detector 13. The secondary signal 12 detected by the secondary-signal detector 13 is amplified by a signal amplifier 14. After that, the secondary signal amplified is transferred to an image memory 25, then being displayed on an image display device 26 as a sample image.
A sample stage 15 is capable of displacing the sample 10 in at least two directions (i.e., X direction and Y direction) within the surface perpendicular to the primary electron beam 4. An input device 42 permits specification of image grabbing conditions (e.g., scanning speed and acceleration voltage), image output, and image storage into a storage device 41. Also, the input device 42 permits specification of a length-measuring domain as well.
This scanning electron microscope also includes a focal-point shift amount determination unit 51, a SEM-image continuous acquisition unit 52, and an image processing unit 53. The image processing unit 53 includes an image-sharpness-degree evaluation unit 54 and a length-measurement unit 55.
Based on image forming conditions of the scanning electron microscope, the focal-point shift amount determination unit 51 determines the focal depth by calculation or the like, thereby calculating focal-point shift amounts among a plurality of SEM images to be acquired. The SEM-image continuous acquisition unit 52 acquires and stores the series of SEM images whose focuses are shifted by the focal-point shift amounts determined by the focal-point shift amount determination unit 51. With respect to the series of SEM images acquired, the image-sharpness-degree evaluation unit 54 of the image processing unit 53 evaluates the image sharpness degrees of partial domains including a length-measuring domain specified by the input device 42, thereby determining a SEM image whose image sharpness degree is the highest. With respect to the SEM image whose image sharpness degree has been judged to be the highest, the length-measurement unit 55 performs the dimension measurement of the specified pattern by using the already-known method as were illustrated in
Incidentally, the focal-point shift amount determination unit 51, the SEM-image continuous acquisition unit 52, and the image processing unit 53 may be provided outside the computer 40, or may be implemented by pieces of software which operate on the computer 40.
Next, referring to
In this processing, the field-of-view including a dimension-measuring domain and an initial value of the focusing condition are set. Positioning of the dimension-measuring domain is performed by the matching with the coordinates or template registered in advance. Also, the initial value of the focusing condition is determined as follows: Focus search is performed once with the image sharpness degree of the entire image. Then, a value acquired by defocusing its search result by a predetermined value determined in advance is defined as the initial value. The initial value of the focusing condition can also be determined as follows: The focus measurement is performed on different stage-coordinate basis, or on different field-of-view basis, thereby creating in advance a focus map which makes the stage coordinates or the field-of-views related with the focuses. Then, based on the focus map, the focus value (i.e., initial value) is determined on each stage-coordinate basis, or on each field-of-view basis.
In this processing, the variation width of the objective-lens current corresponding to the focal depth of the electron-optics system is determined, then acquiring multi-focus SEM images in a predetermined number. The value of the focal depth of a SEM image varies depending on various types of factors such as pixel size of the grabbed image, the acceleration voltage, and the resolution. Accordingly, the value of the focal depth can be defined from the defocus characteristic resulting from taking these factors into consideration (i.e., relationship indicating blurring amount of the image in response to the focus variation).
When observation magnification is low, the focal depth fd of a piece of scanning image with the focusing condition fixed is represented by the following expression [1]:
fd=A1×(dpix/M)×R×√{square root over (Vacc)} [1]
Here, A1 denotes a constant, dpix denotes the pixel size, M denotes the observation magnification, R denotes beam resolution (resolution determined by the beam diameter), and Vacc denotes the acceleration voltage.
If the observation magnification becomes higher, image resolution of the scanning image turns out to be limited by the beam resolution R. As a result, the focal depth fd at this time is represented by the following expression [2]:
fd=A2×R2×√{square root over (Vacc)}/√{square root over ((1+0.73×(Ip/B0)×1014))} [2]
Here, A2 denotes a constant, Ip denotes a probe current, and B0 denotes luminance of the electron gun converted into per-V basis. In the case of an electric-field emission electron source where the luminance B0 is exceedingly high, the term (Ip/B0) within the expression [2] becomes exceedingly small. As a result, the focal depth fd in the high-magnification domain can be represented as the following expression [3] from the practical standpoint:
fd=A2×R2×√{square root over (Vacc)} [3]
Incidentally, in the expression [1] to the expression [3], the beam resolution R can be represented by the following expression [4]: Accordingly, the beam resolution R in the expressions [1] to [3] can be represented in a manner of being replaced by the second term or third term of the expression [4]. Incidentally, λ denotes electron wavelength, and a denotes convergence angle (half angle) of the primary beam.
R=0.61λ/α=0.75/(α×√{square root over (Vacc)}) [4]
When acquiring a plurality of SEM images with different focuses, the focal-point shift amounts among the plurality of images are made equal to or somewhat smaller than the values represented by the expressions [1] to [3]. This makes it possible to obtain the maximum focal-depth enlargement effect in the minimum image number.
Based on the calculations in the expressions [1] to [3], the focal-point shift amount determination unit 51 calculates optimum focal-point shift amounts from the image forming conditions such as the acceleration voltage, electron-source luminance, probe current, pixel number, magnification, and beam resolution. The focal-point shift amount determination unit 51 is also capable of describing these calculation results onto a table in advance, and determining the focal-point shift amount corresponding to an image forming condition by making reference to the table.
Based on the set values of the focal-point shift amounts, the SEM-image continuous acquisition unit 52 varies the focus on one-image grabbing basis. Moreover, the SEM-image continuous acquisition unit 52 continuously performs the focus control and the image grabbing, then storing a series of SEM images with different focuses. The focus control at this time is capable of assuming the following various types of control modes: Namely, the focus is controlled with the present focusing condition selected as the center, the focus is controlled with the present focusing condition selected as the end point, or the focus is controlled within a focusing range set in advance.
In this processing, with respect to the series of respective SEM images with the different focusing conditions acquired in the processing at S12, the image sharpness degrees of partial domains including the length-measuring domain specified are evaluated, thereby determining a SEM image whose image sharpness degree is the highest. This processing is performed by the image-sharpness-degree evaluation unit 54. In the present embodiment, the image sharpness degrees are evaluated by the maximum contrast gradient of the specified length-measuring domain. The contrast gradient indicates variation ratio of brightness between adjacent pixels with respect to brightness distribution of an image. Namely, as an image becomes shaper, the image exhibits larger contrast gradient (i.e., larger variation ratio of brightness). This is because the shaper image accompanies steeper brightness variation at the edge portion.
The image sharpness degrees, however, can also be evaluated by various types of methods other than the maximum contrast gradient. For example, there exists a method of evaluating the image sharpness degrees by applying a spatial filter referred to as “differential filter” to the partial domains to be evaluated, and evaluating the image sharpness degrees based on statistical amounts of the pixel values of the partial domains. In this case, although, as the differential filter, filters such as the Sobel filter as primary differential filter and the Laplacian filter as secondary differential filter have been known, these spatial filters or their modified techniques are also usable. As the statistical amounts, values are used such as total values, average values, variance values, and standard deviation values of the pixel values of the entire partial domains. Then, an image for which the corresponding value becomes the maximum is assumed as the image having the maximum image sharpness degree.
Also, in the present invention, it is also possible to specify image-sharpness-degree evaluating methods which are suitable for the structures of dimension-measuring domains on each domain basis within one and the same field-of-view. The use of the input device 42 permits execution of the specifications of the dimension-measuring domains, the image-sharpness-degree evaluating methods, and the evaluating parameters. Next, referring to
Furthermore, it is also possible to add a function of judging whether or not the maximum value of the image sharpness degrees evaluated with respect to the series of the plurality of SEM images has turned out to become a relative maximum value. This function allows implementation of the detection if the optimum focusing condition for the dimension-measuring domain were not to be included in the SEM images. For example, if, as illustrated in
Incidentally, in the above-described explanation, the explanation has been given concerning the embodiment of selecting the image whose image sharpness degree has been found to be the maximum value. The present invention, however, is not limited thereto. For example, it is also preferable to store a predetermined threshold value in advance, and to select an image having the image sharpness degree which has exceeded this threshold value. Also, if an image is to be selected under a certain limited condition, it is also preferable to select an image which allows the maximum image sharpness degree to be acquired under this condition.
In this processing, with respect to the SEM image determined in the processing at S13, the dimension measurement of the predetermined domain is performed by using the already-known method. The dimension measurement is performed by the length-measurement unit 55.
If a plurality of measurement locations exist within one and the same field-of-view, the dimension measurement is performed by repeating the processing at S13 and the processing at S14. This method allows the length measurement to be executed under a focusing condition appropriate for each length-measuring domain even if height differences exist among the plurality of measurement locations. Also, this method reuses the plurality of SEM images which have been grabbed once. This characteristic makes unnecessary the execution of newly beam irradiation for the measurement, thereby making it possible to perform the high-reliability dimension measurement with the minimum beam irradiation amount.
Next, referring to
Namely, when measuring the bottom diameter of the pattern, as illustrated in
It should be further understood by those skilled in the art that although the foregoing description has been made on embodiments of the invention, the invention is not limited thereto and various changes and modifications may be made without departing from the spirit of the invention and the scope of the appended claims.
Number | Date | Country | Kind |
---|---|---|---|
2004-292772 | Oct 2004 | JP | national |