Method for performing focusing in a particle-optical device with the aid of astigmatism in the particle beam
Abstract
In a particle-optical device such as an electron microscope it is advantageous to perform the process of focusing automatically. In accordance with the invention, the electron beam which is to be focused is deliberately made astigmatic to a certain degree. With this astigmatic beam, two images of a specimen are made at two different settings of the objective, after which in each of the images the direction of the astigmatic smearing is determined - for example, with the aid of a two-dimensional Fourier transform (FFT). The directions of the astigmatic smearing are perpendicular to each other if, in the transition from a first setting of the objective to a second one, the point of optimum focus is passed. Through a process of interpolation (which process may be iterative) between these two settings, the point of optimum focus can now be determined. Possible anisotropy in the specimen itself can be eliminated by making two images at both of the settings of the objective and by subtracting the FFTs thereof from each other. <IMAGE>

Term
Term ended
Expired 29 August 2023, 3.1 years ago.
- Priority and filed
- Granted
- Expired
- Today
7 claims: 3 independent, 4 dependent
- 1Conclusies 5 1 Werkwijze voor het focusseren van een bundel van elektrisch geladen deel^'es in een deeltjes-optisch toestel met een afbeeldende objectieflens, omvattende de stappen van:(a) het uit de genoemde bundel vormen van een eerste astigmatische bundel met een bijbehorende astigmatische richting bij een eerste instelling van de 10 objectieflens;(b) het met de eerste astigmatische bundel maken van een afbeelding van een preparaat in het deeltjes-optisch toestel;(c) het bepalen van de richting van de astigmatische versmering (blurring) in de in stap (b) gemaakte afbeelding;15 (d) het geven van een andere instelling aan de objectieflens;(e) het uit de eerstgenoemde bundel vormen van een tweede astigmatische bundel met een bijbehorende astigmatische richting bij de andere instelling van de objectieflens;(f) het met de tweede astigmatische bundel maken van een afbeelding van 20 het preparaat;(g) het bepalen van de richting van de astigmatische versmering in de in stap (f) gemaakte afbeelding;(h) het vergelijken van de richting van de astigmatische versmering in de in stap (b) gemaakte afbeelding met die in de in stap (f) gemaakte afbeelding;25 (i) het bij gelijkheid van laatstgenoemde richtingen herhalen van de stappen (d) Lm. (h), en het bij ongelijkheid van laatstgenoemde richtingen uitvoeren van een interpolatieproces tussen de eerste instelling van de objectieflens en de laatstveikregen instelling van de objectieflens voor het bepalen van die instelling van de objectieflens waarbij de bundel van elektrisch geladen deeltjes zijn optimale focus bereikt.
- 22 Werkwijze volgens conclusie 1, waarin het bepalen van de richting van de astigmatische versmering in een afbeelding plaats vindt door het bepalen van de spectrale energie-inhoud van een eerste spectraal gebied van de betreffende afbeelding met een eerste richting en van de spectrale energie-inhoud van een tweede spectraal gebied van die afbeelding met een tweede richting dwars op de eerste richting, en het bepalen van tenminste het teken van het verschil van de spectrale energie-inhoud van 5 het eerste gebied en van de spectrale energie-inhoud van het tweede gebied.
- 33 Werkwijze volgens conclusie 2, waarin het vergelijken van de richting van de astigmatische versmering in de in stap (b) gemaakte afbeelding met die in de in stap (f) gemaakte afbeelding plaats vindt door het vergelijken van de tekens van het verschil 10 van de spectrale energie-inhoud van het eerste gebied en van de spectrale energieinhoud van het tweede gebied van elk der afbeeldingen.
- 44 Werkwijze volgens conclusie 1, waarin (j) bij de eerste instelling van de objectieflens een eerste verdere 15 astigmatische bundel met een bekende astigmatische richting dwars op de astigmatische richting van de eerste astigmatische bundel wordt gevormd, en waarin met de eerste astigmatische bundel een eerste afbeelding en met de eerste verdere astigmatische bundel een eerste verdere afbeelding van het preparaat gemaakt wordt;(k) bij de tweede instelling van de objectieflens een tweede verdere 20 astigmatische bundel met een bekende astigmatische richting dwars op de astigmatische richting van de tweede astigmatische bundel wordt gevormd, en waarin met de tweede astigmatische bundel een tweede afbeelding en met de tweede verdere astigmatische bundel een tweede verdere afbeelding van het preparaat gemaakt wordt;(l) in de vier aldus gemaakte afbeeldingen de richting van de astigmatische 25 versmering bepaald wordt door het bepalen van de spectrale energie-inhoud van een eerste spectraal gebied van de betreffende afbeelding met een eerste richting en van de spectrale energie-inhoud van een tweede spectraal gebied van die afbeelding met een tweede richting dwars op de eerste richting, waarbij in elke afbeelding tenminste het teken van het verschil van de spectrale energie-inhoud van het eerste gebied en van de 30 spectrale energie-inhoud van het tweede gebied bepaald wordt 1 0241 92 1θ
- 55 Werkwijze volgens één der conclusies 2 tm. 4 waarin het uitvoeren van het interpolatieproces tussen de eerste instelling van de objectieflens en de laatstverkregen instelling van de objectieflens plaats vindt doordat (m) in elk der bijbehorende afbeeldingen de grootte van het verschil van de spectrale energie-inhoud van het eerste gebied en van de spectrale energie-inhoud van het tweede gebied wordt bepaald (n) dat door interpolatie tussen deze waarden een schatting wordt verkregen van de waarde van die instelling van de objectieflens waarbij de bundel van elektrisch geladen deeltjes zijn optimale focus bereikt (o) dat bij de aldus geschatte instelling opnieuw een afbeelding wordt gemaakt waarop stap (m) wordt toegepast, waarna het bij deze nieuwe instelling verkregen stelsel van waarden dient als nieuwe waarden waarmee opnieuw interpolatie wordt uitgevoerd (p) welke laatstgenoemde stap wordt herhaald totdat het verschil tussen twee opvolgende waarden van de instelling van de objectieflens kleiner dan een vooraf voorgeschreven waarde is, waarna de laatst verkregen waarde geldt als instelling van de objectieflens waarbij de bundel van elektrisch geladen deeltjes zijn optimale focus bereikt.
- 66 Werkwijze volgens conclusie 5 waarin bij de afbeelding die in stap (o) wordt gemaakt tevens een andere waarde van het astigmatisme wordt ingesteld.
- 77 Werkwijze volgens één der conclusies 2 tm. 6 waarin de beeldbewerkingen in digitale vorm plaats vinden, en waarin voor het bepalen van de spectrale energieinhoud van de genoemde spectrale gebieden een subframe van de betreffende afbeelding wordt gevormd. 1/5
Independent claims7
84 paragraphs in 1 section, as filed
<img file="NL1024192C2_D0001.tif" />
Patent Center
Netherlands © 1024192 © C PATENT<sup>20</sup> © Patent application: 1024192 © Int.CI.<sup>7</sup> ff 'H01J37 / 21, E01D15 / 24 © Submitted: 29.08.2003
<td>© Registered: 01.03.2005 (© Date: 01.03.2005 © Published: 02.05.2005 IE 2005/05</td><td>© Patent holder (s): FEI Company of Hillsboro, Oregon, United States of America (US). © Inventors): Willem Hendrik Maes in Lommel (BE) Robertas Johannes Michael van Vucht at Best Hendrikus Petrus Maria Sterken at Deurne</td>
<td></td><td>© Authorized representative: H. Bakker at 5651 GG Eindhoven.</td>
© Method for focusing in a particle-optical device using astigmatism in the particle beam.
© In a particle optical device such as an electron microscope, it is advantageous to have the focusing process proceed automatically. According to the invention, the electron beam to be focused is deliberately made astigmatic to a certain extent. With the astigmatic beam, images of a preparation are made at two different lens settings, after which the direction of the astigmatic smearing is determined in each of the images, eg using a two-dimensional Fourier transform (FFT). The directions of the astigmatic smear are perpendicular to each other when the point of optimal focus has passed at the transition from the first lens setting to the second. The point of optimal focus can now be determined by an (optionally iterative) interpolation process between these two settings. Any anisotropy in the preparation itself can be eliminated by taking two images at either lens setting and subtracting the FFTs from each other.
NL C1024192
<img file="NL1024192C2_D0002.tif" />
The contents of this patent correspond to the original filed description with claim (s) and any drawings.
The Netherlands Patent Office is the Office for Industrial Property, an agency of the Ministry of Economic Affairs
A method of focusing in a particle optical device using astigmatism in the particle beam.
The invention relates to a method for focusing a beam of electrically charged particles in a particle optical device with an imaging objective lens.
Such a method is known from a publication in SCANNING, Vol. 19, (1997) pp. 553-563, entitled "A Robust Focusing and Astigmatism Correction Method for the Scanning Electron Microscope". This article describes a method of focusing a beam of electrically charged particles, wherein the electrically charged particles are electrons. The particle optical device with an imaging objective lens in which this method is performed is a scanning electron microscope (SEM). In the method described there, at two different settings of the imaging objective lens, an image of a preparation is made in the particle-optical device, after which the spectral energy content of each of the images is determined depending on the spatial frequency occurring in that image . The latter process is performed with a so-called Fast Fourier Transform (FFT).
Before making the two images, the nominal strength of the objective lens is first determined, ie the strength at which the electron beam is approximately focused on the specimen. Subsequently, a deviation from this nominal setting is made in such a way that a setting of “overfocus” is created, and a deviation from the nominal setting is also made in such a way that a setting of “underfocus” is created. Imaging objective lens are thus composed of an “over focus” image and an “under focus image. For the automatic focusing of the electron beam, the total spectral energy content of the two images is determined, as well as the difference between the spectral energy content of the “over focus” image and that of the “under focus” image. The ratio R of this difference to the total spectral energy content gives a measure of the defocusing of the electron beam. If R is positive, the “over focus” image is sharper than the “under focus” image and so the focal length should be reduced; if R is negative, the “under focus” image is sharper than the “over focus” image, so the focal length must be increased.
In addition to this method of focusing the electron beam, a method of minimizing astigmatism of the electron beam is described. In the latter method, the spectral energy content of a number of sectors of the image is determined for each of the two images and it is decided from the difference of the respective spectral energy contents in which direction the astigmatism must be increased or decreased in order to finally obtain a obtain a bundle that is practically free from astigmatism.
Both the method for focusing the electron beam and the method for minimizing the astigmatism of the electron beam are described in particular in the aforementioned article on page 558 from equation (2) to page 559 for the section entitled "Implementation" . It will be clear that for the automatic focusing of the electron beam use is only made of the ratio of the spectral energy contents of the two images and not of the degree of astigmatism of the electron beam. In other words, in this known method it is possible to carry out the focusing method without astigmatism being present in the beam to be focused.
The object of the invention is to provide an alternative method for focusing a beam of electrically charged particles in a particle-optical device with an imaging objective lens. In the most general form of this method according to the invention, the following steps are carried out:
(a) forming from said beam (ie, the beam of electrically charged particles to be focused in the particle optical device) a first astigmatic beam with an associated astigmatic direction at a first adjustment of the objective lens;
(b) imaging the preparation in the particle optical device with the first astigmatic beam;
(c) determining the direction of the astigmatic blurring in the image made in step (b);
(d) giving another setting to the objective lens;
(e) forming from the former beam a second astigmatic beam with an associated astigmatic direction at the other objective lens setting;
(f) making an image of the preparation with the second astigmatic beam;
(g) determining the direction of the astigmatic smearing in the image made in step (f);
(h) comparing the direction of the astigmatic smearing in the image made in step (b) with that in the image made in step (f);
(i) if the latter directions are the same, repeat steps (d) to. (h), and if the latter directions are uneven, perform an interpolation process between the first objective lens setting and the last objective lens setting to determine that objective lens setting in which the beam of electrically charged particles reaches its optimal focus.
In this method according to the invention, the beam to be focused is intentionally made astigmatic and the astigmatism in this beam is used to focus the beam. Use is made of the per se known insight that the direction of the smear caused by the astigmatism in the image changes direction 90 ° when the focus of the beam changes from top focus to bottom focus or vice versa. To this end, a first image of the composition is made with a known direction of the astigmatism applied in the beam, and the smearing in this image is determined due to that astigmatism (steps a through c). Then, with another adjustment of the objective lens, a second image of the preparation is made, also with a known direction of the astigmatism applied in this beam, and the smearing in this second image is determined due to that astigmatism (steps dt / mg). It is easy if the direction of astigmatism in the first image is the same as in the second image, but this is not necessary. It is now assumed that the direction of astigmatism in the beam is the same in both cases. If, when comparing the direction of the astigmatic feathering in the first image and that in the second image, it appears that these directions are not equal, then the conclusion must be that when taking one image, under focus was set and when making of the other image, top focus was set. The optimal focus will therefore lie somewhere between these two settings, and the setting where this can be done can now be determined by means of an interpolation process between the two settings of the objective lens. If, when comparing the direction of the astigmatic smearing in the first image and that in the second image, it appears that these directions are equal, then the conclusion must be that when making both images only sub focus or only top focus was present. A new setting should now be made to the objective lens until the directions of both smearings are uneven, after which the above-mentioned interpolation process can be performed. An advantage of the method according to the invention lies in the fact that a zero-crossing is sought with this method (n.1. in the interpolation between the first objective lens setting and the last objective lens setting to determine that objective lens setting in which the beam of electrically charged particles reaches its optimum focus), which generally results in a faster convergence of the method to be applied algorithm then offers the usual methods for automatic focusing in a particle-optical device, where the algorithms used look for a minimum or maximum. As is known, the course of a curve in the vicinity of an extreme value is relatively flat, while that in the vicinity of a zero crossing is much less flat.
In a preferred embodiment of the invention, the direction of the astigmatic smearing in an image is determined by determining the spectral energy content of a first spectral region of the relevant image with a first direction and the spectral energy content of a second spectral region of that image with a second direction transverse to the first direction, and determining at least the sign of the difference of the spectral energy content of the first region and of the spectral energy content of the second region. In this embodiment, in a two-dimensional graphical representation of the spectral energy content, a sector is selected (a first spectral region of the relevant image with a first direction) and the energy content is determined therein; the same is done with a sector that is transverse (preferably perpendicular to it) (a second spectral region of that image with a second direction transverse to the first direction). With a certain direction of the astigmatic smearing in the image, one spectral energy content will differ from the other and from the sign of this difference the direction of smearing can now be determined.
In a further embodiment of the invention, the comparison of the direction of the astigmatic smearing in the image made in Step (b) with that in the image made in step (f) takes place by comparing the signs of the spectral difference energy content of the first region and of the spectral energy content of the second region of each of the images. In this embodiment, the process mentioned in the previous section is applied to each of the two images to determine the direction of the astigmatic lubrication. From the sign of the difference between these two directions it can be determined whether the astigmatic smearings are in the same direction or are perpendicular to each other.
In another preferred embodiment of the invention, the following steps are performed:
(j) at the first adjustment of the objective lens, a first further astigmatic beam with a known astigmatic direction is formed transverse to the astigmatic direction of the first astigmatic beam, and a first image and the first further astigmatic beam are formed with the first astigmatic beam made a first further image of the preparation;
(k) at the second adjustment of the objective lens, a second further astigmatic beam with a known astigmatic direction is formed transversely to the astigmatic direction of the second astigmatic beam, and a second image is formed with the second astigmatic beam and with the second further astigmatic beam made a second further image of the preparation;
(l) in the four images thus made, the direction of the astigmatic smearing is determined by determining the spectral energy content of a first spectral region of the respective image with a first direction and the spectral energy content of a second spectral area of that image with a second direction transverse to the first direction, wherein in each image at least the sign of the difference of the spectral energy content of the first region and of the spectral energy content of the second region is determined
This embodiment of the invention is particularly advantageous for compositions in which the structure has a high degree of directional preference, such as e.g.
is the case with integrated circuits. At the first adjustment of the objective lens, two images are now made with transverse to each other, preferably mutually perpendicular astigmatism. The same is done with the second adjustment of the objective lens. For each of the four images thus made, in (the two-dimensional graphic representation of) the spectral energy content, two different sectors are again chosen (preferably with mutually perpendicular axes) and of the spectral energy contents of each of those the difference is determined for two sectors, here referred to as the sector difference. This difference is a measure of the size and direction of the anisotropy contained in the accompanying illustration. This anisotropy has thus contributed both the direction of orientation of the specimen and the astigmatism of the beam.
The above sector difference is determined for each of the four images mentioned in this embodiment. In both images taken at the same setting of the objective lens, the contribution of the structures in the specimen with strong directional preference is the same in each of the sector differences;
if one determines the difference between these two sector differences (the final difference), this contribution will disappear. As a result, the effect of the astigmatic smearing is almost exclusively retained in the final verachil. This applies to both the final difference of the images taken at the first setting of the objective lens and the final difference of the images taken at the second setting of the objective lens.
The two final differences thus formed can now be compared with each other and on the basis of this comparison it can again be determined whether or not the image taken on the initial adjustment of the objective lens shows a direction of the astigmatic smearing which is equal to that of the second setting of the objective lens captured image. The aforementioned continuation of the method can then be carried out, viz repeating steps (k) and (1) if the directions of the astigmatic smearing are equal, and the
Ί if these directions are uneven, perform an interpolation process between the initial setting of the objective lens and the last setting of the objective lens to determine that adjustment of the objective lens in which the beam of electrically charged particles reaches its optimal focus
In yet another embodiment of the invention, the interpolation process between the initial adjustment of the objective lens and the final adjustment of the objective lens takes place in that (m) in each of the accompanying images the magnitude of the difference of the spectral energy content of the first region and the spectral energy content of the second region is determined (n) by interpolation between these values an estimate is obtained of the value of that adjustment of the objective lens in which the beam of electrically charged particles reaches its optimum focus (o) at the setting thus estimated, an image is again made to which step (m) is applied, after which the system of values obtained at this new setting serves as new values with which interpolation is again performed (p), the latter step being repeated until the difference between two successive values of the objective lens setting is less than a predetermined one, after which the the last value obtained counts as adjustment of the objective lens, whereby the beam of electrically charged particles reaches its optimal focus
By determining the difference between the spectral energy contents of the two regions in each of the images, a measure of the size in which the objective lens setting deviates from the optimum focus for each of those images is obtained. It is then known that the optimum focus in the vicinity of an interpolation point will lie between those two points mentioned; By performing a new interpolation with this estimate, a better approximation of the optimal focus is obtained, which process is continued until the optimal focus is sufficiently approached.
In a further embodiment of the invention, the image taken in step (o) also sets a different value of astigmatism. These measures allow one to use the shape of the elliptical cross-section of the astigmatic beam (the “ellipticity”) to increase sensitivity of the algorithm result. This sensitivity is preferably maximized with the least effort of the algorithm. The influence of the spectral content of the image can be increased by adjusting the ratio of the long and short axes (the “ellipticity”) to the extent to which the algorithm has approached the optimal focus
In yet another embodiment of the invention, the image processing takes place in digital form, and a subframe of the relevant image is formed to determine the spectral energy content of the said spectral regions. In this way, the computational effort of the algorithm is significantly reduced, allowing faster focusing of the beam to be achieved. The formation of a subframe can take place by selecting an area in the image with the correct information content and with this subarea performing the further operations of the method according to the invention.
The invention will be described in more detail with reference to the Figures, in which like reference numerals indicate corresponding elements. Thereby shows
Figure 1: a schematic representation of an astigmatic electron beam;
Figure 2a: a graphical representation of an energy spectrum of a non-isotropic preparation depicted with a non-astigmatic beam;
Figure 2b: a graphical representation of an energy spectrum of the preparation of Figure 2a depicted with an astigmatic beam with a first direction of astigmatism;
Figure 2c: a graphical representation of an energy spectrum of the preparation of Figure 2a depicted with an astigmatic beam with an astigmatism direction perpendicular to that of Figure 2b;
Figures 3a and 3b: a representation of two masks to be applied to the energy spectra of the images made with the astigmatic beams for determining the astigmatic smearing;
Figure 4: A graphical representation of the progression of a quantity that made the difference in spectral energy content of two in different astigmatism
1024132 displays images, as a function of the excitation of the beam focusing lens;
Figure 5: A graphical representation illustrating the course of the iteration algorithm for determining the setting of the beam focusing lens at the optimal focus.
Figure 1 schematically shows an astigmatic electron beam 2. In this beam, the direction of the electrons is from top to bottom. The nominal focus in this beam is located at the cross-section 4, which cross-section is cyclic in this location. Above and below the nominal focus, the beam has an elliptical cross section, as shown by ellipses 6 and 8. The long (short) axis of ellipse 6 is perpendicular to the corresponding axis of ellipse 8. Conversely, it can be said that if of two elliptical cross-sections in an electron beam the corresponding axes are perpendicular to each other, the nominal focus of this beam must be between these cross-sections, and if of two such elliptical cross-sections the corresponding axes are mutually parallel, these cross sections should be on the same side of the nominal focus. The present invention makes use of this insight. Between said elliptical cross-section 6 (resp. 8) and nominal focus 4 is a place where the cross-section of the beam is in the form of a line 10 (resp. 12) With the astigmatic beam shown, it is now possible to expose a preparation in an electron microscope and thereby make images in the usual manner in such a microscope, with the proviso that for the purposes of the invention the beam is intentionally made astigmatic and that the method according to the invention is generally started by exposing the preparation with an astigmatic cross section of the electron beam. The result of such an exposure when creating an image is that every pixel in the image is not round (i.e. approximately point-shaped) but stretched in the direction of the long axis of the ellipse, so approximately in the form of a stripe, so that a “fresh measurement” in the direction of the long axis occurs in the image. to a method known per se for spectral analysis of the spatial frequencies in the image, such as the so-called Fast Fourier Transform (FFT). As is known, the FFT of an image represents the distribution of the spectral energy (the energy spectrum) as it occurs in the image as a function of the spatial frequency, so with a two-dimensional image as a function of the spatial frequencies in both the x- as the y direction. A graphical representation of the result of this FFT applied to an electron microscopic image is shown in Figures 2a, 2b and 2c.
In Figures 2a, 2b and 2c, Figure 2a is a graphical representation of an energy spectrum of a non-isotropic preparation depicted with a non-astigmatic beam, Figure 2b is a graphical representation of an energy spectrum of the preparation according to Figure 2a, depicted with a astigmatic beam with a first direction of astigmatism, and Figure 2c is a graphical representation of an energy spectrum of the preparation according to Figure 2a, depicted with an astigmatic beam with an astigmatism direction perpendicular to that of Figure 2b. Although it is possible to graphically represent the energy spectrum in the form of a gray distribution in the two-dimensional plane, in Figure 2, for the sake of clarity of the representation, a representation in the form of lines of equal energy density is preferred. In Figures 2a-2c the position of the lines of equal energy density is chosen such that the center of the figures (i.e. the origin of the xy-axis cross) corresponds to the spatial frequency value zero.
In Figure 2a, the energy spectrum is based on a non-astigmatic beam image. Here one would have an energy spectrum with circular lines of equal relative energy 18-1 tm. 18-5, generally 18-i, expect because such a beam cannot exhibit smearing in an astigmatic direction. The fact that the lines 18-i mentioned are not circles after all is caused by anisotropy in the preparation, ie that the preparation itself exhibits spatial frequencies different from those in other directions. In the case of Figure 2a, the preparation has more high spatial frequencies in the direction of dashed line 14 than in direction of dashed line 16; this can be caused, for example, by the fact that the preparation contains many elongated details, the length of which is in the direction of dashed line 16.
In Figure 2b, the energy spectrum is based on an image of the same preparation as in Fig. 2a but exposed with an astigmatic beam. The lines of equal relative energy 18-i hereby exhibit measurement in a direction determined by the combination of the anisotropy in the composition and the astigmatism in the beam. If the depicted preparation had the same direction as in the image of FIG. 2a it can be assumed that the direction of the long axis of the elliptical cross section of the beam with which the specimen was exposed is the same as that of the longitudinal direction of the elongated details in the specimen
In Figure 2c, the energy spectrum is based on an image of the same preparation as in Figures 2a and 2b but exposed with an astigmatic beam whose direction of astigmatism is perpendicular to that of Fig. 2b. The lines of equal relative energy 18-i hereby show a measurement in a direction which, as in FIG. 2b is determined by the combination of the anisotropy in the composition and the astigmatism in the beam. Since the high frequencies in FIG. 2c now mainly occur in the direction of dashed line 14, i.e. transverse to the direction of line 16, it may be assumed that the direction of the long axis of the elliptical section with which the specimen was exposed was perpendicular to that of the longitudinal direction of the elongated details in the preparation and that the influence of this beam ellipticity was significantly greater than that of the prepaxate anisotropy.
Figure 3 shows two masks to be applied to the energy spectra of the images taken with the astigmatic beams to determine the astigmatic fresh measurement. As already noted above, the FFT of an image represents the distribution of the spectral energy as it occurs in the image as a function of the spatial frequency, so with a two-dimensional image as a function of the spatial frequencies in both the x and the y direction . For the description of the masking process, it is assumed that said energy spectra are graphically represented in the form of a gray distribution in the two-dimensional plane and not, as in Figure 2, in the form of lines of equal relative energy density. With respect to Figure 3, the location of the gray distribution in the two-dimensional plane is chosen such that the center of the figure of the gray distribution corresponds to the spatial frequency value zero, which center coincides with the center of Figure 3.
One can imagine the application of these masks in this way that the figure of the masking is fully covered with the energy spectrum of the image whose astigmatic smear must be determined. When these masks are used, the area of the energy spectrum represented as two-dimensional gray distribution covered by the screened areas 20a, 22a, 20b, 22b of the masking does not contribute to the spectral energy content to be determined, while the area of the energy spectrum, represented as a two-dimensional gray distribution, which is covered by the non-rasterized regions 24a, 26a, 24b, 26b of the masking, fully contributes to the spectral energy content to be determined. Therefore, by applying the masking according to Figure 3a, the spectral energy content is determined in a direction mainly represented by arrow
28a, and using the masking of Figure 3b, therefore, the spectral energy content is determined in a direction substantially represented by arrow 28b. Now when an image of a preparation has been made with an astigmatic beam with a given astigmatic direction (i.e. the direction of the long axis of the elliptical beam cross-section at the location of the specimen), then relatively few high spatial frequencies will occur in this astigmatic direction and in a direction perpendicular to relatively many. When the astigmatic direction is in the direction of arrow 28a (Figure 3a), a relatively low spectral energy content of this image masked according to Figure 3a will be observed and a relatively high spectral energy spectrum of the same image shown in Figure 3b. spectral energy content.
In the simplest embodiment of the method according to the invention it is assumed that the preparation per se is completely isotropic, ie that there is no directional preference for the spatial frequencies in the preparation. In this situation, two images of the specimen are taken. When taking the first image, the specimen is exposed with a beam obtained with an initial excitation of the focusing lens (the objective lens) with a certain amount of astigmatism applied in the illumination beam Then, when creating the second image, a different actuation of the objective lens is set and all other settings, in particular the astigmatism, left unchanged. Thereafter, an energy spectrum is produced from both images and each of these energy spectra is subjected to two masks, one according to Figure 3a and one according to Figure 3b. To now the direction
1024132 to determine the astigmatic smearing, a quantity V is defined which gives an indication of the direction in which the highest spatial frequencies occur in the energy spectrum of an image. For the situation described above, this quantity V has the following form:
(Σϊ - Σ<sup>ρ</sup><j „tf __ region 24a, 26o region 24b, 26b <sub>z</sub>,<sub>s</sub><sup>5</sup> Σϊ <sup>ω</sup> total j
In the above expression, the first summation extends over the areas 24a and 26a of Figure 3a and the second summation extends over the areas 24b and 26b of Figure 3b; furthermore, Pi is the spectral power of a frequency range i, and the summation extends over all the frequencies i of practical importance in the areas of the masking mentioned by the sum signs. The summation shown in the denominator of expression (1) represents the total spectral power in the respective image. The quantity V is now determined for each of the two images. If the astigmatic direction in the first image is equal to the direction of arrow 28a (Figure 3), the region 24a, 26a will have a lower spectral energy content than the region 24b, 26b; the sign of V will then have a first wind, eg positive. If the astigmatic direction in the second image is the same as in the first image, then the sign of the quantity V will be the same as in the first image. It is then known that when the power of the objective lens is changed, the optimum focus of the beam has not passed the specimen. In that case a further image has to be made, which must be repeated until the sign of V shows a different value. It is then known that the astigmatic direction of the last two images were perpendicular to each other and that the optimal focus between the last two settings of the objective lens must lie. The value of the objective lens actuation is now determined by means of an interpolation process to be described later, with the optimum focus being on the specimen. This achieves the desired focusing of the electron beam.
In another embodiment of the method of the invention, it is believed that the composition is non-isotropic, ie that there is a directional preference for the spatial frequencies in the composition. In this situation, four images of the preparation are taken. When the objective lens is first energized, a first image is made in which a certain amount of astigmatism is applied in the illuminating beam, so that this first image takes place with a first astigmatic beam. Then, with the same excitation of the objective lens, a first further image (i.e. the second image of the total of four images) is made, in which another astigmatism is applied in the illuminating beam such that a first further astigmatic beam is created with a known astigmatic direction which is transverse indicates the astigmatic direction of the first astigmatic beam. After taking these two images, two more images are taken, namely a second actuation of the objective lens creates another image (i.e. the third image of the total of four images) with a certain amount of astigmatism applied in the illuminating beam so that the latter image takes place with a second astigmatic beam. Then, with the same excitation of the objective lens, a second further image (i.e. the fourth image of the total of four images) is made, in which another astigmatism is applied in the illuminating beam such that a second further astigmatic beam is created with a known astigmatic direction which is transverse is on the astigmatic direction. An energy spectrum is now made from each of the four images, and each of these four energy spectra is now masked twice, once with the mask of Figure 3a and once with the mask of Figure 3b. This means that of each of the four images, the total spectral energy content is determined from an area voided by areas 24a and 26a together and from an area formed by areas 24b and 26b together. For the situation discussed here that the preparation to be imaged is non-isotropic, a quantity V can again be defined which gives an indication of the direction in which the highest spatial frequencies occur in the energy spectrum of an image. For the situation described above, this quantity V has the following form:
10241 y '
<img file="NL1024192C2_D0003.tif" />
region 24α, 26α region 24b, 26b
<img file="NL1024192C2_D0004.tif" />
total
<img file="NL1024192C2_D0005.tif" />
region 24a, 28a region 24 a, 26 bz image I \
<img file="NL1024192C2_D0006.tif" />
total
Image 2 (2)
This expression (2) is applied once to both images taken with the first actuation of the objective lens and then also once to both images with the second actuation of the objective lens. The expression (2), the left term, provided with the index “image 1”, refers to the image taken with the first astigmatic beam (ie the first of the four images), and has the right term, provided with the index “image 2”, referring to the image taken with the first further astigmatic beam (ie the second image out of a total of four images). The first summation in the counter of the left term is now obtained by masking the first image with the mask according to Figure 3a and the second summation in the counter of the left term is obtained by masking the first image with the mask according to Figure 3b . The sum in the denominator of the left term represents the total spectral energy in the first image. The left term of expression (2) in this way represents the degree of astigmatic fresh measurement, which astigmatic fresh measurement consists of two components, viz. A contribution from the anisotropy of the preparation and a contribution from the astigmatism applied in the electron beam. The right term of expression (2) is composed in an analogous manner, the contribution from the anisotropy of the preparation being of course equal to that of the left term. Because according to expression (2) the difference between these two terms is taken, this contribution in expression (2) is omitted. Thus, the value of the quantity V is obtained from the two images taken with one objective excitation, and in the manner described above, the sign of this quantity represents only the direction of the measurement of the measurement in the image resulting from the astigmatism in the electron beam.
The above-described method as applied to the images taken at the first actuation of the objective lens can also be applied to the images taken at the second actuation of the objective lens. The sign of the quantity V determined thereby also represents
10241 92.
only the direction of smearing in the image resulting from the astigmatism in the electron beam. Now when the transition from the first to the second actuation of the objective lens has reached the optimum focus of the beam on the other side of the specimen (i.e. transition from bottom focus to top focus or vice versa), this will result from the reversal of the sign of the quantity V. Conversely, if the sign of V does not change, it can be concluded that the optimal focus setting has not been passed. In that case, further pairs of images must be made, which must be repeated until the sign of V shows a different value, after which the optimum value of the objective lens energization is determined via the aforementioned polarization process.
To illustrate the relationship between the quantity V according to expression (1) or (2) and the objective lens actuation, Figure 4 gives a graphical representation of a measured variation of the quantity V (in arbitrary units) as a function of the excitation of the objective lens (in arbitrary units). This gradient is included in the vicinity of the zero crossing of the curve, ie in the vicinity of the optimal focus. Each measurement point in the graph is obtained in the manner described above, thus, with each setting of the objective lens strength, the quantity V is calculated according to expression (1) or (2). Figure 4 now shows that the quantity V changes from a positive value to a negative value, whereby the quantity V thus passes the value zero; this zero crossing therefore has the setting for the optimum focus of the objective lens. To find this point in a practical situation, an iteration process is applied as explained in more detail below
The purpose of the iteration algorithm is to find successive working distances (actuations of the objective lens) that bring the optimal focus closer to the specimen. The iteration algorithm to be described here is based on linear interpolation. This algorithm is explained with reference to Figure 5, in which curve 30 shows the progression of the quantity V depending on the distance of the preparation to the optimum focus (the "focus distance"). In contrast to Figure 4, Figure 5 shows the course of curve 30 in a wide area of the zero crossing of the curve, ie the start and end points of curve 30 are located relatively far away from the optimal focus. The algorithm starts by assuming an initial excitation of the objective lens where the focus distance is represented by Fm. Depending on the sign of the quantity V determined at that first actuation (Vm), an estimate for a new focus distance Fj is made such that the new focus distance F brings the optimum focus closer to the specimen than the previous focus distance Fm. If at this new focus distance Fj, the sign of the associated quantity Vj is the same as the sign of the previous value Vm, the same procedure is repeated until the sign of V changes. After this, a linear interpolation is performed between the two measuring points (Fm, Vm) and (Fi, V;), as represented by the straight line 32 between these two measuring points. The value of the focus distance at which line 32 intersects the horizontal axis is Fb-i; at this value of the focus distance, the corresponding value of the quantity V is determined (Vh-i). Depending on the sign of Vj + i, a previous measuring point is now selected with which the interpolation continues; if the sign of the immediately preceding value of V, i.e. Vi, is the opposite of the sign of Vi + i, then the interpolation is continued with the pair of measuring points (Fi, Vi) and (Fi + i, Vi + i), otherwise with the pair of measuring points (Fm, Vm) and (Fi, Vi). This iteration process is then repeated until the difference between two consecutive defocus values drops below a predetermined value. The associated focus distance is considered optimal focus
It should be noted that it is possible to significantly reduce the computation time required for this algorithm by using a subframe of the dataset instead of the entire digital data set of an image. This can significantly reduce the computational effort of the algorithm, thereby faster focusing of the beam can be achieved The formation of the subframe can take place by selecting an area in the image with high information content, ie an area in which a lot of detail in the image can be distinguished, which means that it contains a high spectral energy content.
With this sub-region it is now possible to carry out further operations of the method according to the invention.
10241 92
11 sheets
Sheet 1 Sheet 2 Sheet 3 Sheet 4 Sheet 5 Sheet 6 Sheet 7 Sheet 8 Sheet 9 Sheet 10 Sheet 11
Every citation, both waysCites: the store holds 2 of 3
| Document | Relation | Office | Category | Cited during | Relevant claims |
|---|---|---|---|---|---|
| EP0435195A2 | Cites | European Patent Office (EPO) | A | Search report | 1 |
| EP0435195A2 | Cites | European Patent Office (EPO) | A | Search report | 1 |
| ONG K H ET AL: "ROBUST FOCUSING AND ASTIGMATISM CORRECTION METHOD FOR THE SCANNING ELECTRON MICROSCOPE", SCANNING, MAHWAH, NJ, US, vol. 19, no. 8, 1997, pages 553 - 563, XP000920708 | Non-patent | – | – | Search report | – |
| OGASAWARA M ET AL: "Automatic focusing and astigmatism correction method based on Fourier transform of scanning electron microscope images", JAPANESE JOURNAL OF APPLIED PHYSICS, PUBLICATION OFFICE JAPANESE JOURNAL OF APPLIED PHYSICS. TOKYO, JP, vol. 38, no. 2A, February 1999 (1999-02-01), pages 957 - 960, XP002167741, ISSN: 0021-4922 | Non-patent | – | – | Search report | – |
9 members in 5 offices
Priority claims2
| Document | Office | Kind | Date |
|---|---|---|---|
| 1024192 | Netherlands (Kingdom of the) | A | |
| NL20031024192 | – | – | – |
Members9
| Document | Office | Kind | |
|---|---|---|---|
| NL1024192C2This record | Netherlands (Kingdom of the) | C2 | |
| EP1511065A1 | European Patent Office (EPO) | A1 | |
| US2005045831A1 | United States of America | A1 | |
| CN1591761A | China | A | |
| JP2005079100A | Japan | A | |
| US6992289B2 | United States of America | B2 | |
| CN100585786C | China | C | |
| JP4868723B2 | Japan | B2 | |
| EP1511065B1 | European Patent Office (EPO) | B1 |
2 legal events, as the office reported them to INPADOC
Over the term
Point at a mark for the eventEvents
| Event | Code | |
|---|---|---|
| Lapsed due to non-payment of the annual feeLapsedVD1 | VD1 | |
| A search report has been drawn upPD2B | PD2B |
Numbers
- Publication, DOCDB
- 1024192
- Publication, EPODOC
- NL1024192C
- Application
- 1024192
- Application, DOCDB
- 1024192
- Application, EPODOC
- NL20031024192
Titles2
- Dutch
- Werkwijze voor het focusseren in een deeltjes-optisch toestel met behulp van astigmatisme in de deeltjesbundel.
- English
- A method of focusing in a particle optical device using astigmatism in the particle beam.
Classification
- CPC, 2
- H01J37/21
- H01J2237/216
- IPC, 2
- H01J37 153
- H01J37 21