JP4007305B2 - Method for evaluating resolution of electron microscope and method for adjusting electron microscope - Google Patents

Description

  The present invention relates to a sample for quantitative evaluation of resolution, resolution, etc. of an electron microscope, a resolution evaluation method and adjustment method of the electron microscope for evaluating resolution, resolution, etc. using the sample, an electron microscope, and an electron microscope thereof. The present invention relates to the semiconductor manufacturing method used.

  Conventionally, the resolution of an electron microscope is evaluated by the distance between two points of a sample that can be visually confirmed, and a sample in which gold particles are vapor-deposited on carbon as in JP-A-5-45265 is observed and separated. The resolution is the minimum gap between two gold particles. Moreover, the resolution evaluation of the electron microscope is evaluated by the degree of blur of the particle boundary.

JP-A-5-45265

In the conventional resolution evaluation, when the image captured by the electron microscope captures gold particles, the size and shape of the gold particles vary, resulting in individual differences in measurement, and the same sample cannot be reproduced. Therefore, accurate quantitative evaluation is not desired. Also, there is no method for accurately measuring the gap between gold particles being observed with an electron microscope. The resolution of the electron microscope varies depending on the device. For this reason, in the case of semiconductor manufacturing etc., the observation images between these electron microscopes differ in the manufacture of semiconductors in which the dimensions of each part of the semiconductor are measured using a plurality of electron microscopes. However, there is a problem that the manufacturing cannot be performed based on a centralized and constant control value.

  An object of the present invention is to provide a method for evaluating the resolution of an electron microscope and a method for adjusting an electron microscope that solve the above-described problems.

  In order to achieve the above object, according to the present invention, in a method for evaluating the resolution of an electron microscope, a plurality of materials having different generation efficiency of secondary charged particles such as secondary electrons, reflected electrons or transmitted electrons are alternately arranged on the surface. The formed sample having a repetitive pattern is irradiated with an electron beam, and secondary charged particles generated from the surface of the sample by the irradiation of the electron beam are detected to obtain a secondary charged particle image on the surface of the sample. The waveform data of the next charged particle image is subjected to frequency analysis processing using Fourier transform to perform quantitative evaluation of the resolution of the electron microscope.

  In order to achieve the above object, in the present invention, in the method of adjusting an electron microscope, materials having different generation efficiency of secondary charged particles such as secondary electrons, reflected electrons, or transmitted electrons are alternately arranged on the surface. A sample having a repetitive pattern is irradiated with an electron beam from an electron microscope, and secondary charged particles generated from the surface of the sample by the irradiation of the electron beam are detected to obtain a secondary charged particle image on the surface of the sample. The waveform data of the secondary charged particle image is subjected to frequency analysis processing using Fourier transform to evaluate the resolution of the electron microscope, and according to the evaluated resolution, the optical system, vacuum system, or electron gun of the electron microscope Was adjusted.

According to the present invention, the resolution and resolution of the SEM are evaluated by using a multilayer thin film sample whose dimensions are known in advance, and quantitatively evaluating the imaging results of the SEM by means such as frequency analysis. It becomes possible to grasp changes accurately. In particular, in a manufacturing process using a plurality of SEMs such as semiconductor inspection, individual differences between SEMs can be reduced and inspection accuracy can be improved.

  In addition, if the performance of the SEM is measured using a sample prepared according to the present invention, the performance is not affected by the size of the gold particles, and the performance is always less than that in the case of using a gold-deposited sample on conventional carbon. Quantitatively stable performance measurement is possible. In addition, the resolution can be automatically measured by measuring the bulge amount of the SEM image.

  Embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing a schematic configuration of a scanning electron microscope (hereinafter referred to as SEM) which is an embodiment of an electron microscope that measures and adjusts resolution according to the present invention. The electron optical system includes an electron gun 13, a deflector 12, an electromagnetic lens 14, and the like. Then, the electron beam 8 emitted from the electron gun 13 with a desired acceleration voltage is focused by the electromagnetic lens 14, and the surface on the sample 1 mounted on the XYZ stage 17 by the deflector 12 or the like is in any order. It is comprised so that it may scan with. Further, secondary electrons or reflected electrons generated on the surface of the sample 1 by the irradiation of the electron beam are detected by the detector 15 and input to the image input device 16 as image data. The sample 1 that is the object to be inspected can be moved in all three-dimensional directions by the XYZ stage 17. Electron beam irradiation and image input synchronized with the stage movement are possible, and the movement of the stage is controlled by the control computer 20. Although FIG. 1 shows an example of SEM using a secondary electron image, the means for observing the sample may use an image of reflected electrons, transmitted electrons, etc. in addition to the secondary electrons. Those that detect transmitted electrons are referred to as STEM.

A sample 21 for evaluating the resolution of the SEM will be described. Sample 21 includes, for example, Ta film and Si film, W film and C film, Ru film and C film, Mo film and B4C film, W film and S film.
A thin film of a material with a large difference in atomic number such as an i-film is formed on a single crystal silicon substrate by a known thin film forming technique such as plating, CVD (Chemical Vaper Deposition), PVD (Physical Vaper Deposition), etc. The substrates are alternately stacked to form a multilayer, and the substrate on which the multilayer thin film is formed is cleaved and the cross-sectional surface 22 is used. When this is imaged, for example, the generation efficiency of secondary electrons from the Ta film, W film, Ru film, and Mo film is high due to the difference in the generation efficiency of secondary electrons, and from the white, Si film, C film, and B4C film. The secondary electron generation efficiency is low, and the image has a remarkable black contrast.

In order to obtain a cross-sectional structure of a multilayer thin film using these materials, there are methods such as using a cleavage plane (a crystal mineral is easily cracked in a certain direction to form a smooth surface, that is, a cleavage plane). However, when the unevenness of the cross-sectional surface 22 is too large compared to the resolution to be evaluated, for example, abrasive particles of 100 nm or less are buried in a tin plate, and the cross-section 22 is polished by the tin plate. A process such as making the unevenness of the surface to be several nm or less is performed.

  In order to examine the resolution of the SEM, for example, a sample 21a shown in FIG. 2 is used. For example, the Ta layer 23a and the Si layer 23b having a thickness of about 100 nm, which is sufficiently larger than the diameter of the focused spot of the electron beam 8, are formed on the cross-sectional surface 22 and focused by a desired acceleration voltage from the electron gun 13 of the SEM. When the electron beam spot is scanned, the energy distribution of the electron beam and the intensity 32a of the secondary electrons detected by the detector 15 and input to the image input device 16 are schematically shown in FIG. The arithmetic processing unit 19 obtains a differential signal 33 by, for example, differentiating the intensity signal 32a of the secondary electrons input from the image input device 16 to obtain a differential signal 33, and the spot of the electron beam from the differential signal 33 is changed to the Ta layer 23a and Si layer. It is possible to know a position 35 s where scanning of the boundary surface 24 with the layer 23 b starts and a position 35 e where scanning ends. And the arithmetic processing unit 19 can obtain | require the spot diameter which is the shape of an electron beam by taking the difference of these two positions 35s and 35e. Further, the arithmetic processing unit 19 may use the half width φ1 of the curve 33 drawn by the differentiation of the output change of the secondary electrons as the electron beam diameter.

  The electron beam diameter can be regarded as equivalent to the resolution, whereby the SEM performance can be quantitatively evaluated. In this case, the noise component of the signal is reduced and the electron beam diameter is measured with good reproducibility by approximating the secondary electron intensity 32a by B-spline approximation or the differential signal 33 by normal distribution function approximation or the like. Is possible. In this case, for example, the two layers do not necessarily have to be formed by alternately laminating the Si layer 23b and the Ta layer 23a on the Si substrate by a film forming process, and a plurality of substances having a remarkable difference in the generation efficiency of secondary electrons are formed. You may laminate | stack by crimping | bonding, adhesion | attachment, etc.

  A method when the sample 21b shown in FIG. 4 is used to measure the resolution of the SEM will be described. This sample 21b has a cross section formed by superposing a thin Ta layer 23a of, for example, about 1 nm between Si layers 23b sufficiently larger than the electron beam diameter, for example. , When the electron beam spot focused and emitted by a desired acceleration voltage is scanned, the electron beam energy distribution and the secondary electron intensity 32b detected by the detector 15 and input to the image input device 16 are schematically illustrated. As shown in FIG. Then, the processor 19 subtracts the value of the background (intensity of secondary electrons obtained from the Si layer 23b) from the curve drawn by the secondary electron output at this time, and the curve shows the layer with the electron beam Ta. It is possible to know the position 35s where scanning starts and the position 35e where scanning ends. The diameter of the electron beam can be calculated by further subtracting the Ta film thickness from the difference between the two positions 35s and 35e. In this case as well, the full width at half maximum φ2 may be calculated in the same manner as when using the sample 21a shown in FIG. 2, and then the Ta film thickness may be subtracted to obtain the electron beam diameter. In this case, the thickness of the thin Ta layer is changed to, for example, 0.5 nm, 2 nm, and 3 nm, and the arithmetic processing unit 19 determines the energy distribution of the electron beam alone from the state of change of the generation intensity of secondary electrons according to the result. It can be obtained with high accuracy.

  A method when the sample 21c shown in FIG. 6 is used to measure the resolution and resolution of the SEM will be described. As shown in FIG. 6, when using a sample 21c in which a white thin slit pattern is drawn on a black background, when scanning with a scanning electron microscope, diffraction is observed when two pinholes are observed in a general optical system. A signal similar to the light intensity distribution with the image is obtained. Optically, the limit of resolution by Rayleigh is defined as shown in Iwanami Shoten's “Wave Optics” p359-377, etc. The minimum value at the center of such a signal is 74% or less of the maximum value In this case, the two pinhole images are considered to be separated.

  In order to apply this definition to SEM, it appears that a white thin slit pattern (striped pattern; for example, the black background is Si layer and the white slit is Ta) as shown in FIG. A sample 21c is prepared. For example, when the slit interval is the same as the resolution desired to be set and the SEM having a resolution of 1 nm is to be adjusted, the film thickness is controlled to 1 nm to prepare the sample 21c. Then, the arithmetic processing unit 19 obtains the background (Si) from the maximum value of the secondary electron intensity change signal 32c obtained when the sample 21c is imaged by SEM and the local minimum value as shown in FIG. When the intensity of secondary electrons obtained from the layer 23b) is subtracted and the ratio between the minimum value and the maximum value is taken, for example, when it is 74%, the resolution or resolution limit can be defined as 1 nm.

If the sample having such a configuration is used, the resolution to be evaluated can be changed by changing the interval between the slits. If each SEM having a different resolution or resolution limit is quantified with samples having different slit intervals as shown in FIG. 13 by using the above-described method, the arithmetic processing unit 19 has a certain fixed value. The resolution and resolution limit can be quantitatively evaluated using only the sample with the slit interval. That is, the arithmetic processing unit 19 associates the quantified resolution and resolution limit with function approximation or the like as shown in FIG. 12 and creates a performance evaluation chart, so that the slit interval is the same, for example, 1 nm. By measuring the minimum and maximum values on the sample, the resolution and resolution limit can be quantitatively evaluated.

  A method in the case where a sample 21d as shown in FIG. 8 is imaged and frequency analysis is performed using Fourier transform in order to measure the SEM resolution will be described.

  As shown in FIG. 8, screen data obtained by imaging a multilayer thin film material with a striped observation image obtained at equal intervals with an SEM is Fourier-transformed in one or two dimensions in the arithmetic processing unit 19. FIG. 9 is a graph showing the relationship between the frequency of the stripe pattern and the signal intensity for the power spectrum obtained by the Fourier transform. The origin at which the direct current component is maximum indicates that the frequency is zero.

  The frequency indicates that the frequency increases as it moves left and right around the origin. The arithmetic processing unit 19 can quantitatively evaluate the resolution from the peak signal intensity at the frequency of the stripe pattern. The arithmetic processing unit 19 can quantitatively evaluate the resolution from the ratio or difference between the signal intensity of the DC component and the signal intensity of the peak of the stripe pattern. That is, as the signal intensity at the peak of the stripe pattern increases, the resolution improves as it approaches the signal intensity of the DC component.

  Further, in order to measure the SEM resolution, a sample 21d as shown in FIG. 8 can be imaged to obtain a CTF (Contrast Transfer Function). In this case, when the maximum luminance value of the striped pattern portion of the imaged screen data is maxL and the minimum luminance value is minL, the resolution is quantitatively evaluated with a value of (maxL−minL) / (maxL + minL) × 100 (%). it can.

  In the case of the sample shown in FIG. 8, when the magnification of the electron optical system and the acceleration voltage are constant, when the SEM image is observed as shown in FIGS. 10 and 11, the white portion image appears to expand with a constant size. . For example, the magnification and acceleration voltage of the electron optical system of the SEM are as follows when the white pattern portion appears to expand to 14 nm in the SEM image when a striped sample with an interval of 10 nm is observed as shown in FIG. When a striped sample with an interval of 20 nm is observed, the white portion appears to expand to 24 nm as shown in FIG. Since the magnitude of the expansion is the same in this way, the magnitude of the expansion can be defined as the resolution, for example, 4 nm in the above example.

  FIG. 13 shows an example of resolution and a sample for evaluating the resolution. By arranging the fringe patterns shown in FIGS. 2, 4, 6, and 8 on one sample, the resolution and resolution can be evaluated by a plurality of methods without exchanging the samples.

  FIG. 14 shows an embodiment of a stage or sample holder 24 to which the sample 1 and the sample 21 for resolution evaluation are attached. For example, when used for semiconductor inspection, a square stage or sample holder 24 has a space in the four corners for mounting a circular wafer. For example, a sample holder 41 as shown in FIG. The sample holder 41 is made energizable with the sample 21 sandwiched therebetween. Although the sample 21 usually cannot be used again due to electrification or surface contamination once observed with an SEM, charging and surface contamination are removed when the sample is heated to, for example, 1000 ° C. or more, and the number of times the sample is replaced. Is reduced or unnecessary.

The structure of the sample 51 for evaluation for decomposition is shown in FIG. The W film 53 and the C film 54 are repeatedly laminated on the Si wafer 52 with a thickness of the C film 54 having a dimension of 6 nm and a thickness of the W film 53 of about 3 to 10 nm between the intervals A55. Next, the C film 54 is repeatedly laminated with a thickness of 5 nm between the intervals B56 and the W film 53 is approximately 3 to 10 nm, and then the C film 54 is increased between the intervals C57. With the dimensions of 4 nm, the W film 53 is repeatedly laminated with a thickness of about 3 to 10 nm. Obviously, the thickness of the W film 53 is not necessarily about 3 to 10 nm. Then, when the sample 51 is observed with an SEM, the SEM image is displayed on the display means 30 so that it looks like an SEM image 58 as shown in FIG. When this SEM image 58 is seen, the W film 3 swells and the C film 4 which is a dark part cannot be observed in the portion of the interval 57 although the C film 54 is originally laminated by 4 nm. This means that the SEM of 5 nm can be observed separately, but the 4 nm cannot be separated, that is, it has a resolution of 5 nm.

  As another example of the method for measuring the resolution of the SEM, a case where the sample 61 shown in FIG. 17 is used without using the resolution chart as shown in FIG. 12 will be described.

The sample 61 is created as follows, for example. Ru film 63 on Si wafer 62
And the C film 54 are repeatedly laminated with the C film 64 having a thickness of 10 nm between the intervals A65 and the Ru film 63 with a thickness of about 3 to 10 nm, and then the C film 64 between the intervals B66. The thickness of the W film 63 is repeatedly laminated with a dimension of about 3 to 10 nm, and then the thickness of the C film 64 is 5 nm between the intervals C67. The film thickness is repeatedly laminated with a dimension of about 3 to 10 nm. Obviously, the film thickness of the W film 63 is not necessarily about 3 to 10 nm.

  For example, when an SEM image of the sample 61 is obtained, if the SEM image is Fourier transformed, a power spectrum similar to that in FIG. 8 is obtained. A signal at the frequency of the striped pattern in FIG. 8, in this case, only frequency components of the interval D65, the interval E66, and the interval F67 are extracted. This means that a plurality of spots 82 appearing in the Fourier transform image 81 of FIG. 18 are extracted. The plurality of spots 82 are taken out, inverse Fourier transform is performed, and an SEM image to which a band pass filter is applied is reproduced. The brightness profile of the SEM image is shown in FIG.

The brightness amplitude of brightness decreases in the order of amplitude D, amplitude E, and amplitude F in the interval D65, interval E66, and interval F67. This is because when the interval is narrowed, the image is not completely resolved. When the interval D is completely resolved, the ratio of the amplitude D to the amplitude E and the amplitude F in FIG. 18 is taken, and plotted on a graph with the horizontal axis as the interval (nm) and the vertical axis as the amplitude ratio (%). A resolution curve or a resolution straight line as shown in FIG. 20 is obtained. When Rayleigh's optical resolution limit is applied, the value of the horizontal axis at the position where this curve or straight line falls to 26% can be defined as the resolution limit and resolution. This 26% may be changed for each model by a visual correlation experiment or the like.

  In addition to the W film / C film, there are W film / Si film, Ru film / C film, Mo film / B4C film, Ta film / Si film, etc. as a combination of materials to be laminated.

  Further, although a single crystal silicon substrate is used as a substrate on which a thin film of these materials is formed, the material of the substrate is not limited to this.

  These SEM resolution samples 51 can be used not only for performance evaluation at the time of SEM production, but also for aging evaluation during daily inspection and maintenance.

  If the above-described phenomenon is used, the processing unit 19 adds an image processing technique to automatically measure the bulge amount of the image, thereby automatically measuring the resolution, which is one of the SEM performances, and quantitatively determining the resolution. Can be evaluated. Therefore, if the sample 51 is mounted on the SEM, the apparatus itself can always manage the resolution performance of the SEM.

  The above description has been given for one dimension (X-axis direction), but it is obvious that it can be applied individually to two dimensions. It is clear that the resolution also changes when the acceleration voltage is changed.

  When the resolution cannot be satisfied, the focus condition can be adjusted by inputting the condition setting value for controlling the electron optical system 14 using the input means 31. Further, the degree of vacuum can be adjusted by inputting a condition set value for controlling the vacuum control system 32 using the input means 31. Since the condition of the electron gun 13 may be poor, it is necessary to adjust the electron gun 13 in that case.

  In addition, since history data relating to resolution is stored in the storage device connected to the arithmetic processing unit 19, it is possible to notify the user at any time by using the display means 30, for example.

It is a schematic block diagram of a scanning electron microscope. It is a figure which shows the example of the performance evaluation sample which consists of two layers. It is explanatory drawing of the performance evaluation method by the sample of FIG. It is a figure which shows the example of the performance evaluation sample which pinched | interposed one thin film with two layers. It is explanatory drawing of the performance evaluation method by the sample of FIG. It is a figure which shows the example of the performance evaluation sample which pinched | interposed two thin films with three layers. It is explanatory drawing of the performance evaluation method by the sample of FIG. It is a figure which shows the example of the performance evaluation sample formed into a film | membrane in stripe form. It is explanatory drawing of the performance evaluation method by frequency analysis. It is explanatory drawing of the performance evaluation sample and the imaging result which were formed into the striped form at intervals of 10 nm. It is explanatory drawing of the performance-evaluation sample formed into a stripe at 20 nm space | interval, and an imaging result. It is a SEM resolution evaluation chart figure. It is a figure which shows the performance evaluation sample structural example. It is a figure which shows a sample stage. It is a figure which shows the sample holder for electric heating. It is a figure which shows the structure and SEM image of the sample which laminated | stacked the metal which has a difference in the generation | occurrence | production efficiency of the secondary electron which concerns on this invention, changing a dimension for every certain space | interval. It is a figure which shows the example of a structure of the sample which laminated | stacked the metal which has a difference in the generation | occurrence | production efficiency of the secondary electron which concerns on this invention, changing a dimension for every certain space | interval. It is the image which carried out the Fourier transform of the resolution evaluation sample. It is a figure which shows the profile of the image which carried out the inverse transformation after applying the band pass filter after Fourier-transforming a SEM image. It is a figure which shows a SEM resolution evaluation curve.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Sample, 8 ... Electron beam, 13 ... Electron gun, 14 ... Electromagnetic lens, 15 ... Secondary electron detector, 16 ... Image input device, 17 ... XYZ stage, 19 ... Arithmetic processing device, 20
... Control computer, 24 ... Sample holder

Claims (4)

A method for evaluating the resolution of an electron microscope, the width dimension in a repetitive direction formed by alternately arranging a plurality of materials having different generation efficiency of secondary charged particles of secondary electrons , reflected electrons, or transmitted electrons on the surface. secondary but which irradiated by scanning the electron beam to a sample having a known repetitive pattern, generated in response to the secondary charged particle generating efficiency of each material from the surface of the specimen by irradiating by scanning the electron beam detecting the charged particles to obtain a contrast marked secondary charged particle image of the surface of the sample caused by the different secondary charged particle generating efficiency depending on the material, 2 contrast according to the materials significant primary and performing quantitative evaluation of resolution of the electron microscope from the peak intensity signal at the frequency of the repeated pattern and the frequency analysis processing using a Fourier transform of the waveform data of the charged particle image Resolution evaluation method of child microscope. A plurality of repetitive patterns are formed on the surface of the sample, and between the plurality of patterns,
2. The resolution evaluation of an electron microscope according to claim 1, wherein the width dimension in the repeating direction of at least one of the materials having different generation efficiency of the secondary charged particles constituting the repeating pattern is different. Method. 2. The resolution evaluation method for an electron microscope according to claim 1, wherein the repetitive pattern is formed of tantalum (Ta) and silicon (Si). An electron microscope adjustment method, in which an electron microscope is applied to a sample having a repetitive pattern formed by alternately arranging materials with different generation efficiency of secondary charged particles of secondary electrons , reflected electrons, or transmitted electrons on the surface from irradiated by scanning the electron beam, by detecting the secondary charged particles generated in accordance with the secondary charged particle generating efficiency of each material from the surface of the specimen by irradiating by scanning electron line said material contrast of the surface of said sample to obtain a significant secondary charged particle image caused by the different secondary charged particle generating efficiency, contrast according to the material of the waveform data of significant secondary charged particle image Fourier according to when said by frequency analysis processing using the converted evaluate the resolution of the electron microscope from the intensity signal of the peak at the frequency of the repeated pattern, the resolution was the evaluation can not be satisfied Adjustment method for an electronic microscope and adjusting the focus condition of an optical system of a child microscope.

Images