Method of delayering an area of interest in a sample
Patent Information
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- TESCAN GRP AS
- Filing Date
- 2024-08-29
- Publication Date
- 2026-07-08
Smart Images
Figure CZ2024050054_06032025_PF_FP_ABST
Abstract
Description
[0001] Method of delayering an area of interest in a sample
[0002] Field of Art
[0003] The present invention relates to a method of processing a sample using a device with a focused beam of charged particles.
[0004] Background Art
[0005] Modern semiconductor industry continues to develop new products involving ever smaller structures. Smaller structures, however, also require increasingly precise and delicate machining and handling. Devices using a stream of particles, in particular electrons or photons, are increasingly being used for machining and observing very small structures. Such devices include, in particular, a scanning electron microscope (SEM), a cluster ion source (CIS) device or a focused ion beam (FIB) device. In particular, devices using ions can be a very effective tool for sample processing. Due to the relatively high momentum with which their particles hit the surface of the sample, the particle energy may be used to eject atoms or molecules from the surface of the sample and thereby remove layers of material from the surface.
[0006] One of the methods of removing a layer of material is the so-called delayering. Delayering is a technique in which the sample surface is milled with an ion beam in such a way that the sample material is milled in parallel layers. The method is widely used, for example, in semiconductor industry, where individual layers of semiconductor devices or other layered samples are gradually exposed by delayering. The method can also be used for non-layered materials with the aim of exposing a large flat surface that was inside the volume of a sample before the start of milling, for subsequent examination of this surface.
[0007] However, when machining semiconductor devices, problems may be encountered that complicate the machining or disproportionately increase the time required for such machining. One of the problems encountered in machining is the removal of a hard metal layer which often is on the surface of semiconductor devices. Therefore, currently used methods often choose an approach in which the hard metal layer is first removed by another technique, for example mechanically, and only then the semiconductor sample is inserted into a device working with a charged particle beam where the machining of finer structures continues. The disadvantages of this procedure are mainly high demands on equipment, more complex handling of the sample, when it is necessary to move the sample from one device to another, as well as the inaccuracy of mechanical methods, which can cause irregularities on the surface of the sample, which can complicate its further observation or processing.
[0008] The surface can also be delayered using an ion beam. Most methods using ions in delayering are based on the ability of the ion to eject particles uniformly from the area of interest and thereby achieve the desired flatness. The main disadvantage of this method is that the different particles on the surface usually have different sputtering speeds, due to which unevenness is created between places with different sputtering speeds when scanning the ion beam over the sample. One of the methods for limiting this effect is the use of sputtering with ions while simultaneously injecting a working gas. This procedure is described, for example, in US10886139B2. The gas is used in order to unify the sputtering speeds of the individual materials and thereby achieve a uniform sputtering. The disadvantage of this method is that there is no universal working gas suitable for all types of sputtered material. Also due to the fact that this method of delayering takes place at an angle between the optical axis of the device and the delayered plane that is usually greater than 2°, in order to reach deeper layers it is necessary to sputter a large amount of material from around the area of interest, which causes an increase in time required as well as a risk of contamination by redeposition of the sputtered material from around the area of interest.
[0009] Another way to reduce the amount of unevenness during delayering is to sputter the sample from multiple directions or rotate the sample during sputtering. The rotation of the sample in relation to the ion beam is very often used in so-called ion polishers using a wide or diverging beam, where different sputtering speeds are compensated by changing the position of the sample. Changing the direction of sputtering is also used in the raster method of sputtering, but the disadvantages of this method are caused by the beam usually striking the sample surface at an angle greater than 2°.
[0010] Yet another way to achieve a smooth surface on the sample is to gently polish the sample with a particle beam. This method is mainly used for fine polishing of lamellae for transmission electron microscopy, and is therefore not a sample delayering as such. The lamellae also have very small dimensions, usually around 5- 10 pm in the direction of sputtering and are very thin. With this method, great emphasis is placed on not damaging the sample, therefore the energies used are very low, and thus the process is also very slow and not suitable for delayering large areas. The accelerating beam voltages used are usually up to 10 keV. In some cases, in order to protect the face of the sample from unwanted sputtering, deposition of a thin protective layer on the sample is used, but these layers are usually no thicker than 1 -2 pm.
[0011] If the working accelerating voltage of the beam is increased to values of, for example, 20 keV or more and at working currents of 1 nA or more, with larger sizes of the sputtered area, for example in the order of tens of pm, or longer exposure of the sample to the beam, very uneven sputtering and thus undesirable curvature of the surface of the sample occur. This unwanted curvature arises as a result of the particle beam striking the sample being non-ideal and not focused to a single point in the focal plane, but rather creating a region having non-zero area, called the beam spot. The reason for this is that the particles in the beam do not always have the same energy and do not always fly along a trajectory coincident with the optical axis of the column. If the particle motion vector were divided into a component in the direction of the optical axis and a second component perpendicular to the optical axis, then it is the component perpendicular to the optical axis directed towards the surface of the sample that would be responsible for the curvature. This curvature further increases with the hardness of the sample, i.e. with the decreasing sputtering speed of the material of which the sample is composed, as a result of the fact that the beam has to act on the sample for a longer time, and therefore the perpendicular component of the motion vector of the individual particles acts on the surface of the sample for a longer time. This curvature causes complications, for example, when observing a delayered area, as it can cause image distortions during observation, or in the case of a sample containing multiple layers on top of each other, interference and imaging of multiple layers at once may occur and the desired goal of revealing all the structures of a single layer in the area of interest is not achieved.
[0012] The curvature of the sample surface most typically has the shape of an S- shaped curve and is described, for example, in US10832889B2.
[0013] US10832889B2 uses a combination of sectional sputtering using a diverging beam, a movable aperture and tilting the sample to eliminate this undesirable effect, so as to achieve a straight surface in the largest possible desired area. The working voltages used in the recited patent are around 6 keV, but in this case the relatively low working voltage is replaced by a large beam cross-section. Furthermore, with this delayering method, there is no rastering by the particle beam, the so-called point by point, on the sample. This solution is quite demanding for setting the microscope parameters during the method and constant correction of the optical elements of the microscope to achieve an optimal result.
[0014] Disclosure of the Invention
[0015] The above-mentioned drawbacks are overcome by the present invention, which is a method of delayering an area of interest in a sample using a focused converging particle beam.
[0016] The most prominent complication in achieving the goal of exposing all structures of a single layer in the area of interest at once is caused by a first curvature at the leading edge of the sample and a second curvature appearing further down the sputtered plane of the sample, the region between these curvatures being largely linear. Therefore, it is necessary to provide such a method of delayering and such a combination of beam parameters that eliminates the curvature at the leading edge of the sample and at the same time extends the relatively straight central part of the delayered area as much as possible so that in the area of interest the straightness of the sample surface has a lower value than the thickness of the area of interest. To achieve the desired results, the present invention proposes the method of delayering the surface of the area of interest described herein below.
[0017] The method of delayering the surface of an area of interest in a sample by a charged particle beam distributed along the optical axis of a column of a device generating the charged particle beam, wherein said device further includes means for setting the values of the working voltage and current and means for focusing said beam to the focal plane, means for injecting a working gas to working chamber, means for forming a protective layer on the sample and a sample stage for receiving and positioning the sample located in the working chamber of the device, the sample stage being configured to adjust the position of the sample relative to the optical axis of the column, wherein the said method of delayering the area of interest comprises the steps of: a) selecting a value of tolerance of straightness of a straight section in the horizontal profile of the surface to be delayered of the area of interest in the direction of the optical axis, b) setting the values of the working voltage and current of the particle beam so that with the selected value of tolerance of the straightness of the straight section in the horizontal profile of the surface to be delayered, the length of the said section, between the first point of the section and the last point of the section, is equal to or greater than the length of the area of interest in the direction of the optical axis, c) determining of the thickness of the protective layer so that for the set working voltage and current and for the selected straightness tolerance value, the said first point is located inside the protective layer, d) forming the protective layer adjacent to the area of interest, the protective layer having the thickness determined in step c), e) placing the surface to be delayered of the area of interest in a position substantially parallel to the optical axis of the column, while the distance between the surface to be delayered of the area of interest and the optical axis in the focal plane is less than half the diameter of the said charged particle beam in the focal plane, f) scanning the sample with said charged particle beam in a plane substantially parallel to the delayered surface of the area of interest, wherein the said charged particle beam produces a section in the horizontal profile of the area of interest of the sample, parallel to the optical axis, meeting the condition of straightness at the selected tolerance value.
[0018] The main advantage of this method of delayering consists in determining the optimal thickness of the protective layer at the working current and voltage set for delayering of the area of interest so that after delayering the sample, a surface in the area of interest is obtained that meets the conditions of straightness of the straight section in the horizontal profile of the delayered surface at the selected tolerance value. In order for the delayering method to bring the expected advantages of the invention, which mainly consists in delayering large areas of interest above 400 pm2, higher working currents above 10 keV and voltages above 1 nA are usually used, which requires the use of a thicker protective layer. The use of a thicker protective layer also has the advantage that it is easier to set the focal plane of the beam up to the level of the protective layer, i.e. further from the sample towards the column, thereby improving the profile of the beam hitting the sample, while the ability of the beam to sputter the sample in the plane perpendicular to the direction of the optical axis does not essentially change with the shift of the focal plane towards the column, but the ability of the beam to sputter the sample in the plane parallel to the optical axis is reduced, which can be used to extend the straight section in the horizontal profile or reduce the tolerance value of the straightness of the straight section in the horizontal profile of the delayered surface.
[0019] The optimal value of the thickness of the protective layer for performing the method can be determined, for example, from a database containing the position of the first point and the length of the straight section at the selected tolerance value of the straightness of the straight section in the horizontal profile for the set values of the working voltage and current of the particle beam.
[0020] The optimal value of the thickness of the protective layer can be determined, for example, by first selecting from the database the records including the set values of the working voltage and current of the particle beam. From the selected records, those records are selected that have the same or greater value of the length of the straight section at the selected tolerance value than the value of the length of the area of interest. The thus selected records are sorted according to the position of the first point, i.e. the distance of the first point from the edge of the sample from the smallest distance to the largest. The position of the first point of the sample corresponding to the smallest distance of the first point from the edge of the sample is also the distance corresponding to the minimum thickness of the protective layer.
[0021] The database can be created in advance by applying the selected values of working voltage and current, for example, to a test sample that is different from the examined sample or directly to the examined sample so that the area of interest is not damaged during the creation of the database.
[0022] Alternatively, this database can be created for a specific device in advance and contain tabulated values of the thickness of the protective layer and the length of the straight section at various values of working voltages and currents, which will allow the user to very quickly select the correct conditions for delayering the area of interest while respecting the selected straightness tolerance value of the straight section in the horizontal profile of the delayered surface of the sample. The tabulated values of the thickness of the protective layer in this case correspond to the smallest distance of the first point from the edge of the sample using the same values of the working voltage and current of the particle beam applied to the sample without the protective layer.
[0023] The obtained information about the position of the first point and the length of the straight section at the selected tolerance value of the straightness of the straight section in the horizontal profile for the set values of the working voltage and current of the particle beam can be used as training data for an artificial neural network, which is trained using this data to determine the minimum thickness of the protective layer for unknown combinations of working current and voltage values and the selected straightness tolerance value of the straight section in the horizontal profile. This neural network is trained so that, based on at least one value of the distance of the first point and the length of the straight section at the selected tolerance value and with the known working voltage and current values of the particle beam of a specific device, it can estimate the values of the distance of the first point and the length of the straight section at the selected tolerance value, for different combinations of particle beam working voltage and current values of the same or similar device.
[0024] The advantage of using artificial neural networks lies in the fact that it allows to determine the optimal parameters of the thickness of the protective layer even for the values of working voltage and current of the particle beam, which were not yet tabulated in the database for given lengths of the straight section and given straightness tolerance values of the straight section in the horizontal profile of the delayered surface.
[0025] In some cases, there may be multiple combinations of protective layer thickness, working current and voltage that meet the straightness tolerance value of a straight section in the horizontal profile for a given area of interest size in the direction of sputtering. In such a case, the output of a search in the database or the output of the trained artificial neural network, after entering the parameters of the size of the area of interest in the sputtering direction, can be a set of combinations of the values of the thickness of the protective layer, the values of the set working voltage and current of the particle beam, and the selected tolerance value of the straightness of the straight section in the horizontal profile, sorted by working voltage and current values of the particle beam. As the size of the working voltage and current values of the particle beam increases, so does the sputtering speed of the sample, and it is possible to rank the results according to the predicted speed of delayering of the area of interest.
[0026] The advantage in this case is mainly the minimization of the time needed to carry out the delayering of the area of interest.
[0027] The protective layer that is applied to the sample before the start of delayering also has a great influence on the speed of the entire process. There are many types of layers that are used in combination with charged particle beam material removal techniques. The reason for applying and using these layers is usually to protect the sample from damage by charged beam particles. Materials commonly used for forming protective layers can be, for example, organic materials, often consisting of polymers or layers of carbon. Another type of protective layers can be metal layers that include metal atoms, for example platinum, gold, aluminum, or tungsten. Alternatively, these layers can be made of metal oxides such as aluminum oxide or silicon dioxide. There are also multiple ways of depositing these layers on the sample surface, such as vapor deposition (CVD) or physical vapor deposition (PVD), or high-energy particle beam deposition, such as ion, electron or laser beam deposition. For carrying out the method according to the invention, it is particularly advantageous when the speed of sputtering of the protective layer is not higher than the speed of sputtering of the sample which is protected by the protective layer.
[0028] A charged particle beam device for carrying out the method of the invention comprises a particle beam distributed along the optical axis of a column of a charged particle beam generating device using a charged particle source, said device further comprising means for setting the values of the working voltage and current, such as, for example, control elements for setting heating voltage or electrodes for setting extraction voltage, and means for focusing said beam to the focal plane, such as, for example, electromagnetic lenses, stigmators, potential electrodes, apertures and other commonly used components for particle optics, wherein the device further comprises means for injecting a working gas to working chamber, such as, for example, inlet pipe, valve, controller, gas reservoir and nozzle, which can be used to supply the gas near the sample, which is transformed by irradiation with a charged particle beam into a deposit forming protective layers on the sample, and further a sample stage for receiving and positioning the sample located in the working chamber of the device, the sample stage being configured to adjust the position of the sample relative to the optical axis of the column.
[0029] The device for carrying out the method for delayering of the area of interest may further include a second columne having a second optical axis and generating a second beam of charged particles distributed along a second optical axis, positioned such that the first optical axis and the second optical axis intersect within the working chamber. The sample stage for the sample, which is part of the device, allows the sample to be placed in close proximity to the intersection of the first and second optical axes.
[0030] The advantage of the second optical column is mainly its location in the working chamber and the possibility to observe the delayered area of interest using this column before, during or after the delayering without the need to change the position of the sample stage. This cannot be done with the first column, since the area of interest is not in its field of view during the delayering process, or is shaded away by the protective layer.
[0031] Thanks to the position of the second column in combination with a sufficient thickness of the protective layer, it is possible, after the step of forming a protective layer adjacent to the area of interest and before the step of placing the surface-to-be- delayered of the area of interest into a position substantially parallel to the optical axis of the column, while the distance between the surface-to-be-delayered of the area of interest and the optical axis in the focal plane is less than half the diameter of said beam of charged particles in the focal plane, to add a step of forming a mark e.g. fiducial mark in the protective layer in a plane substantially parallel to the surface-to- be-delayered, the mark being in the field of view of the second column when the first charged particle beam scans the sample in a plane substantially parallel to the delayered surface. To implement this step, after forming the protective layer, it is necessary to first place the region of the protective layer where the mark is to be created in the field of view of the charged particle beam and form this mark in the protective layer with the beam of charged particles. After the mark is made, the sample is placed in a position where the optical axis of the particle beam is substantially parallel to the delayered surface, in which position the mark is not visible in the field of view of the charged particle beam intended for sample delayering. An easily recognizable, in some cases machine-recognizable, structure can be used as the mark, which enables easier navigation on the sample and its surroundings, or predictable deformations of the mark during delayering may be used to monitor the delayering process.
[0032] Furthermore, it is possible to use the mark to correctly set the position of the optical axis of the first column relative to the sample and the area of interest. Due to the fact that the first optical axis and the second optical axis intersect, it is possible to set the foci of the first and second beams to the intersection of the two axes. Subsequently, it is possible to position the sample so that the mark is located in the focus of the second beam. The next step is to use rotation to adjust the position of the sample so that the surface of the area of interest is parallel to the optical axis of the first column and thus the marker is shaded by the protective layer and is not visible to the first beam. Thanks to this, it is possible to easily achieve that the focal plane of the beam is located between the sample and the column and passes through the protective layer and is perpendicular to the plane of the surface of the area of interest.
[0033] Examples of carrying out the invention
[0034] Figure 1 shows a device 1. for carrying out the method according to the invention with a first column 2 generating a charged particle beam. Charged particles can be, for example, ions, especially ions of metals or noble gases, but also ion clusters, or the particles can be electrons. Furthermore, the device 1. is equipped with a working chamber 3, inside which there is an adjustable sample stage 4, on which a sample 5 is placed. The sample stage 4 can be set to at least one position in which the first optical axis 6 is essentially parallel to the surface of the area of interest 7 on the surface or inside the sample 5. The adjustment can be made, for example, by rotating the sample stage 4 around the first axis 21 of rotation, which is parallel to the plane in which the first optical axis 6 is located. The sample 5 can be, for example, a semiconductor device with layers of integrated circuits, for example a chip or a battery, or other samples, including biological samples, for which it is important to achieve a relatively large and flat surface after delayering.
[0035] Furthermore, the device 1_ includes means for injecting a working gas, in particular a reservoir 9 of gas precursor, a pipe 10 for guiding the precursor into the working chamber 3, and a nozzle 11 opening into the working chamber near the sample 5. The working gas can include molecules containing, for example, Pt, W, or carbon, which form a solid deposit on the sample after being irradiated with the charged particle beam.
[0036] The device can further be equipped with a second column 12 having a second optical axis 13. The second column 12 can generate a beam of particles for observing or machining the area of interest 7 or checking the process of delayering.
[0037] Other components of the first column 2 are the first source of particles 14, optical means 15 for guiding and shaping and focusing the first beam, for example lenses, stigmators, potential electrodes, apertures 17 and other commonly used components for particle optics, which allow to reach and work at the working voltage of at least 10 keV, but for delayering of large areas, voltages of at least 20 keV or 30 keV with a working current of at least 1 nA and higher are more suitable for time reasons, for example 2.5; 10; 20; 30; optionally up to 100 nA, deflecting means 16 enabling deflection of the beam from the first optical axis 6, and aperture 17, participating in setting the shape and current value of the first beam 26. Given that the first beam serves mainly to remove material from the sample by ejecting the sample particles from its surface, it is advisable to choose particles with high momentum. Particularly suitable for this purpose are ions of noble gases, for example Ar, Xe, or metals, for example Ga, or optionally ion clusters. However, other elements known from the literature can also be used. Other components in the second column 12 are a second source of particles 18, a second optical means 19 for guiding and shaping and focusing the second beam, for example lenses, stigmators, potential electrodes, apertures and other commonly available components for particle optics, a second deflection means 20 enabling deflection of the beam of the second beam from the optical axis 13. The particles generated by the second column 12 can be, for example, electrons or ions of light elements, because they are not usually used for treating the sample, but for observing or analysis of the sample. In special cases, the second particle source 8 could be replaced by a light source, for example a laser, for observation and analysis.
[0038] Furthermore, the device is equipped with at least one detector 8 of signal particles, wherein this detector can be located temporarily or permanently in the working chamber 5 or in one of the columns 2, 12. The movable sample stage 4 can be further adjustable to at least one second position in which the normal of the area of interest 7 is parallel to the second optical axis 13. This position is most suitable for observing the delayered region without image distortions or for creating a mark 29 on the sample 5 or in a protective layer 22.
[0039] Still further, the sample stage 4 can be equipped with such a positioning mechanism that allows the area of interest 7 to be rotated around the second axis 24 of rotation, which is parallel to the normal of the area of interest 7, or the second axis 24 of rotation can be perpendicular to the first optical axis 6, thereby eliminating the so-called curtaining effect, which usually appears on samples during sputtering with a focused ion beam from only one direction.
[0040] Figure 2 shows a schematic cross-section of the sample 5 before the start of delayering. The sample 5 contains an area of interest 7 that includes structures 36 to be prepared for observation or further analysis by delayering. The area of interest is delineated on two sides by planes 32 defining the tolerance limits of the surface straightness. In the section of the sample 5, these planes 32 represent straight lines. The planes 32 define the straightness tolerance limits of the delayered area 7 of the sample. In the case of the cross section of the sample 5, where the planes are represented by straight lines, these are the straightness tolerance limits in the sectional view of the delayered region. The planes 32 delineate the boundaries between which the entire surface of the area of interest 7 should be located after delayering. Thus, the lines or planes 32 defining the straightness or flatness tolerance define that after the sample 5 has been delayered, the surface of the area of interest 7 or its sectional view should lie between the lines or planes 32.
[0041] In Figures 3A to 3D, sectional views showing the result of delayering the sample 5 at various thicknesses of the protective layer 22 are shown. The surface of the sample was delayered using the first beam 26 distributed along the first optical axis 6 substantially parallel to the planes 32 defining the tolerance limits of the surface straightness. It is not always technically possible to set the position of the planes 32 parallel to the area of interest 7 and the first optical axis 6 so that they are completely mutually parallel. In such a case, we consider an essentially parallel arrangement such that the angle of incidence between the first optical axis 6 and the plane 32 is not greater than 2°. In cases of devices allowing very precise adjustment, the value of the angle of incidence may be less than 0.5°. The first beam 26 is generally converging and having a focal plane 25 outside the column 2. In the focal plane 25, the first beam 26 has an approximately circular cross-section with an approximately Gaussian distribution of particles in the beam. The first beam 26 is made up of charged particles, for example ions or electrons, which move inside the cone of the beam in the region between the aperture 17 and the sample 5 along trajectories towards the focal plane 25. Due to the essentially Gaussian distribution of the angular and energy distribution of the particles in the beam, during delayering on the surface of the sample 5 in the direction of the optical axis 6 an S-shaped profile corresponding to the distribution of particles is formed. Figure 3A shows the situation where the sample 5 was delayered without the use of a protective layer 22. The horizontal profile 33 has a distinct S-shape and a relatively straight central part. The first point 34 and the last point 37 are then determined in the horizontal profile. The points 34, 37 are determined so that, within the selected tolerance of straightness in the horizontal profile 33, they delimit the longest part of the horizontal profile 33 meeting this tolerance. Determining the distance of the first point 34 and the last point 37 can be done by taking a cross- sectional image of the delayered area, as shown, for example, in Figures 3A to 3D. Two parallels representing planes 32, whose distance corresponds to the required tolerance, are subsequently adjusted to the image. The parallels representing the planes 32 are then gradually moved along the image in a direction perpendicular to the direction of the optical axis 6, for example from the upper edge towards the lower edge of the image, and in the selected positions of the parallels the distance of the first point 34 from the edge of the sample and the last point 37 from the first point 34 is recorded. These values are then plotted into a graph. An example of such a graph is shown in Figures 4A and 4B. Figure 4A shows the distance measurements for sample 1 sequentially at 9 positions. The figure shows that the largest distance between the first point 34 and the last point 37 is in position 7. In this position, the distance of the first point 34 from the edge of the sample is 15 pm and the distance of the second point 37 from the first point 34 is 90 pm. Figure 4B shows the distance measurements for sample 2 sequentially at 9 positions. The figure shows that the largest distance between the first point 34 and the last point 37 is in position 6. In this position, the distance of the first point 34 from the edge of the sample is 40 pm and the distance of the last point 37 from the first point 34 is 100 pm. At the same time, it is clear from the graph in Figure 4B that if the distance of the last point 37 from the first point 34 of 60 pm, as determined in point 8, would be sufficient for the needs of the analysis at the chosen level of straightness, then at this value the distance of the first point 34 is 20 pm. Therefore, when determining the thickness of the protective layer, a much thinner protective layer would be sufficient than if a straight section length of 100 pm were required, as measured in point 6.
[0042] Figure 3B and 3C show the situation where the sample 5 was delayered under the same conditions, especially the same working current and voltage, as in the case of Example 3A. In contrast to the situation in Figure 3A, in this case a thin protective layer 22 was deposited on the sample. As in the example in Figure 3A, here also the horizontal profile 33, indicated by a dashed line, contains the first point 34 and the last point 37, which are determined as so that, within the selected tolerance of straightness in the horizontal profile 33, they delimit the longest part of the horizontal profile 33 meeting this tolerance. In these particular cases, two different sections can be found in the horizontal profile, the first one marked in Figure 3B and the second one in Figure 3C, having the same length. In the case where it is possible to identify two different first points 34 in the same interval in the horizontal profile 33, it would be appropriate to consider whether the optical axis 6 is in the optimal position relative to the area of interest 7, and possibly adjust the position of the optical axis so that in the horizontal profile 33 in the area delimited by the planes 32 exists exactly one first point 34 and one last point 37.
[0043] Figure 3D shows the situation where sample 5 was delayered under the same conditions, especially the same working current and voltage as in Example 3A. In contrast to the situation in Figure 3A, in this case a thick protective layer 22 was deposited on the sample. The thickness of the protective layer 22 was chosen so that the first point 34 was not located in the area of the sample 5, but in the area of the protective layer. Since the particle distribution in the beam does not change substantially on the same device at the same working current and voltage, the thickness of the protective layer meeting such parameters is essentially the same as the distance between the edge of the sample and the first point in Figure 3A. The easiest way to determine the thickness of the protective layer is to first delayer the sample 5 in a test area that is not the area of interest 7 but has a similar material composition, and determine the first point 34 in the horizontal profile 33 of the test area at the specified tolerance level. Determine the distance between the edge of the sample 5 and the first point 34 in the horizontal profile 33 of the test area. This distance corresponds to the minimum thickness of the protective layer needed for performing the method. Distance values for different working currents and voltages and different levels of straightness tolerance in the horizontal profile 33 can be stored in a database from which they can be retrieved if necessary without the need to perform repeated delayering in test areaa. The database can also contain other data, for example the composition and structure of the sample.
[0044] Another piece of information that the database can contain is the distance between the last point 37 and the first point 34 in the test area, wherein this distance determines the maximum distance between the last point 37 and the first point 34 in the horizontal profile 33 meeting the selected interval of straightness or flatness. In order for the sample surface to be considered straight, the flatness tolerance value or the straightness tolerance value in the horizontal profile 33 must be equal to or less than 100 nm. The size of the area of interest 7, measured in the direction of the first optical axis 6, should not be greater than the distance between the first point 34 and the last point 37 in the horizontal profile 33 in the test area without the applied protective layer 22. In practice, it can be found that the shape of the beam 26 in combination with the protective layer 22 also changes the distribution of particles in that part of the beam 26 that hits the surface of the sample 5 in the area of interest 7, and usually, in the resulting horizontal profile with the same straightness tolerance, a larger distance between the first point 34 and last point 37 than in the experimental area can be observed.
[0045] The distance of the last point 37 is therefore influenced by the thickness of the protective layer 22 and the beam profile. Assuming that all the necessary beam parameters can be reliably determined, the distances of the last points 37 for various conditions could be calculated or modeled. In practice, however, it is very difficult to determine with sufficient accuracy all the parameters necessary for such a calculation, therefore a solution is proposed where the obtained values of the distance of the last point 37 from real measurements together with the known parameters in particular of the working current and voltage, the focal length and the aperture 17 size serve as data for training an artificial neural network, which after entering known parameters will provide an estimate of the distance of the last point 37 in the horizontal profile 33. Figure 5 shows a schematic cross-section of the sample during delayering of a layered semiconductor sample 5 with a deposited protective layer 22 having thickness t. The sample 5 consists of individual layers that can have the same or different height h. Hard top metal layer is usually found on the upper side of the semiconductor sample followed by alternating metal layers and dielectric layers. The sample 5 is delayered by the first beam 26 in a position where the focal plane 25 is shifted to the front edge of the protective layer 22 and is substantially perpendicular to the area of interest 7. The first optical axis 6 and the area of interest 7 are less than half a diameter of the beam spot of the first beam 26 in the focal plane 25 apart. Thanks to the deposition of the protective layer 22, in this case the focal plane 25 can be moved closer to the first particle source 14, up to the level of the protective layer 22. Moving the focal plane 25 even closer to the first particle source 14, outside of the sample level 5 or outside the protective layer 22, would be very difficult, since the empty space does not include any reference points that could facilitate the focusing of the beam 26. Furthermore, the thickness (t) of the protective layer and the position of the focal plane 25 together with the shape of the beam 26 determine the angle of incidence 28, below which particles can fall into the area of interest 7. As the value of the angle of incidence 28 decreases, the ability of the falling particles to sputter the material also decreases and thus unwanted sputtering in the area of interest 7 is also reduced. This allows to achieve a better straightness of the surface of the sample 5, i.e. smaller distance between the planes 32a and 32b which determine the tolerance limits for the desired straightness value. The minimum thickness of the protective layer 22 on the tested device at working voltage and current values of 30 keV and 2.5 pA for Ga ions, and for straightness of 100 nm was determined to be 8 pm, which corresponds to the distance of the first point 34 from the front side of the protective layer 22.
[0046] To facilitate focusing, a recognition mark 29 can be formed on the side of the protective layer 22 parallel to the area of interest 7, which facilitates the recognition of the protective layer 22 and its focusing and setting of the focal plane 25 to the correct position using the second beam of particles, which is distributed along the second optical axis 13.
[0047] The start of the deformation of the mark 29 caused by the first beam 26 during the delayering, which can be observed by the second column 12, can be used to determine the moment when the first beam 26 reached the area of interest 7 during the delayering or to monitor the progress of the gradual delayering of the sample 5 using the gradual deformations of the mark 29.
[0048] In Figure 6, there is an example of an embodiment wherein an area of interest 7 of a sample 5 with a protective layer 22 is delayered. The protective layer 22 has a sufficient thickness so that the first point 34 is located at the level of the protective layer 22. The area of interest is the area that is in the section view shown in figure 6 by lines 7a, 7b and 7c. Within the area of interest are structures 36 to be delineated for observation or analysis as described herein. The planes 32a and 32b define in the cross-section the straightness tolerance limits that must be observed for the correct delayering of the structures 36. To implement the method, the distance between the planes 32a and 32b should be smaller than the distance between the regions 7a and 7b delimiting the area of interest 7 and should be located between these regions. The last point 37 is positioned so that its distance from the front edge of the sample 5 is greater than the distance of the end of the area of interest from the edge of the sample 5 in the same direction. This should ensure that all structures 36 in the area of interest 7 will be exposed or not removed after delayering and can be further analyzed, observed or processed.
[0049] Figure 7 shows the dependence of the distance of the first point 34 and the last point 37 at working currents within the range of 2.5 to 30 nA for straightness levels of 40, 60 and 100 nm and Ga ions at an accelerating voltage of 30 keV. Distances were measured in the test area without the protective layer applied. The minimum thickness of the protective layer 22 for Ga ions at an accelerating voltage of 30 keV corresponds to the distance of the first point from the edge of the sample 5, or from the edge of the sample 5 including the protective layer 22. The distance between the first point 34 and the last point 37 corresponds to the maximum achievable length of a straight section at the required straightness value of the horizontal profile 33. For example, for a working voltage of 2.5 nA the minimum thickness is determined to be 8 pm, for 10 nA the minimum thickness is 13 pm, and for example, for 30 nA the minimum thickness of the protective layer is 18 pm. The smallest length of the straight section or delayering depth to meet large area conditions is at least 20 pm, therefore, taking into account the required straightnesses, the maximum usable current for a 40 nm straightness is about 5 nA, for a 60 nm straightness about 15 nA, and for a 100 nm straightness about 30 nA. Figure 8 shows a similar dependence as in Figure 6, with the difference that the graph was created for Xe ions and a current range of 10 to 100 nA. For example, for a working voltage of 10 nA the minimum thickness is determined to be 13 pm, for 30 nA the minimum thickness is 18 pm, and for example, at 100 nA the minimum thickness of the protective layer is 45 pm. The length of the section between the first point 34 and the last point 37 was 20 pm for the tested device. This value is met by all currents in the specified range and at all specified straightnesses.
[0050] Fig. 9 shows the beginning of the delayering of the sample 5 using the first beam 26. A protective layer 22 is deposited on the layered sample 5, into which a mark 29 can be etched, visible in Fig. 10. The first optical axis 6 is located in parallel to the layer 30, which includes the area of interest 7. The first beam 26 is directed at the sample 5. If the means for deflecting the first beam 16 do not allow such a large deflection of the first beam 26 from the first optical axis 6, it is necessary to compensate for this deficiency by shifting or tilting the sample stage 4 on which the sample is placed 5, which can temporarily cause non-parallelism of the area of interest 7 and the optical axis 6.
[0051] In the following step, the irradiation of the sample 5 with the particle beam 26 and the removal of material from the sample 5 is started. In doing so, it is advantageous to proceed from the upper edge of the sample 5 to the layer 30 including the area of interest 7, in the direction of the arrow 31 .
[0052] Figure 10 shows the sample 5 after completion of delayering using the first beam 26. Part of the sample, up to the layer 30 containing the area of interest 7, has been removed and delayered up to the level of the area of interest 7. The sample is now ready for observation using, for example, the second column 12, for depositing material or for sputtering specific structures 36. In this case, the second optical axis 13 of the second column 12 is directed at the protective layer 22 or at the sample 5 and is oriented in such a way as to allow reading the mark 29 in the protective layer 22. The second optical axis 13 and the first the optical axis 6 intersect in close proximity to the sample 5, approximately in the focal plane 28 of the first column 2. This can be used for precise adjustment of the focal plane 28 to the level of the protective layer 22.
[0053] Figure 11 shows the XZ view, so-called top view, of the area of interest 7, where the letter d represents the dimension of the area of interest in the direction of the first optical axis 6 and the letter w represents the dimension in the direction perpendicular to the first optical axis 6. The area of interest is located inside the sample 5. Adjacent to the area of interest is a protective layer 22. A cross-shaped mark 29 is made in the direction parallel to the area of interest 7 in this embodiment. The mark 29 can be used to correctly adjust the position of the sample 5 with respect to the focal plane 25 of the first beam 26. In doing so, it is possible to use the fact that the first optical axis 6 and the second optical axis 13 are located so that they intersect and the focal plane 25 passes through their intersection. The position is then adjusted by placing the mark 29 in this intersection and thereby placing the focal plane in the desired position.
[0054] Arrow 38 indicates the rastering movement of the first beam 26 in the final phase of delayering in close proximity to the area of interest 7, optionally within the area of interest 7. In the movement indicated by the arrow 38, the beam moves in a plane essentially parallel to the area of interest 7. The beam 26 can perform the same movement in the area of the sample 5 which is located above the area of interest 7 and which must be removed in order for the area of interest 7 to become accessible. In contrast to scanning in close proximity to the area of interest 7 in more distant layers, it is not necessary to observe the condition of scanning in a plane parallel to the area of interest 7, as can be seen for example in Figure 8. To avoid the curtaining effect, the sample 5 can also be rotated around the second axis 24 perpendicular to the first axis 21 during delayering. In the figure, the second axis 24 is shown as passing through the area of interest 7, but its position may vary, for example, depending on the location of the sample 5 on the sample stage 4. Preferably, the second axis 24 of rotation is positioned so that so that it is perpendicular to the first axis 21 of rotation in all rotational positions of the sample 5 around the first axis 21 of rotation. The advantage of such an arrangement is easier handling of the sample to eliminate the curtaining effect, since the first axis 21 and the second axis 24 are always in a stationary position relative to each other.
[0055] List of reference signs
[0056] 1 - Device
[0057] 2 - First column
[0058] 3 - Working chamber
[0059] 4 - Sample stage
[0060] 5 - Sample - First optical axis
[0061] 7 - Area of interest
[0062] 8 - Detector
[0063] 9 - Reservoir
[0064] 10 - Pipe
[0065] 11 - Nozzle
[0066] 12 - Second column
[0067] 13 - Second optical axis
[0068] 14 - Source of particles
[0069] 15 - Optical means
[0070] 16 - Deflection means
[0071] 17 - Aperture
[0072] 18 - Second source of particles
[0073] 19 - Second optical means
[0074] 20 - Second deflection means
[0075] 21 - First axis
[0076] 22 - Protective layer
[0077] 24 - Second axis
[0078] 25 - Focal plane
[0079] 26 - First beam
[0080] 28 - Impact angle
[0081] 29 - Mark
[0082] 30 - Layer
[0083] 31 - Arrow
[0084] 32 - Plane
[0085] 33 - Horizontal profile
[0086] 34 - First point
[0087] 36 - Structures
[0088] 37 - Last point
[0089] 38 - Arrow
Claims
CLAIMS1 . Method of delayering the surface of an area of interest in a sample by a charged particle beam distributed along the optical axis of a column of a device generating the charged particle beam, wherein the said device further includes means for setting the values of working voltage and current and means for focusing said beam to the focal plane, means for injecting a working gas to working chamber, means for forming a protective layer on the sample and a sample stage for receiving and positioning the sample located in the working chamber of the device, the sample stage being configured to adjust the position of the sample relative to the optical axis of the column, wherein the said method of delayering the area of interest comprises the steps of: a) selecting a value of tolerance of straightness of a straight section in the horizontal profile of the surface to be delayered of the area of interest in the direction of the optical axis, b) setting the values of the working voltage and current of the particle beam so that with the selected value of tolerance of the straightness of the straight section in the horizontal profile of the surface to be delayered, the length of the said section, between the first point of the section and the last point of the section, is equal to or greater than the length of the area of interest in the direction of the optical axis, c) determining of the thickness of a protective layer so that for the set working voltage and current and for the selected straightness tolerance value, the said first point is located inside the protective layer, d) forming the protective layer adjacent to the area of interest, the protective layer having the thickness determined in step c), e) placing the surface to be delayered of the area of interest in a position substantially parallel to the optical axis of the column, wherein the distance between the surface to be delayered of the area of interest and the optical axis in the focal plane is less than half the diameter of the said charged particle beam in the focal plane, f) scanning the sample with said charged particle beam in a plane substantially parallel to the surface to be delayered of the area of interest,wherein the said charged particle beam produces a section in the horizontal profile of the area of interest of the sample, parallel to the optical axis, meeting the condition of straightness at the selected tolerance value.
2. The method of delayering the area of interest in the sample according to claim 1 , where the thickness of the protective layer is determined from a database containing information about the position of the first point and the length of the straight section at the selected tolerance value of the straightness of the straight section in the horizontal profile for the set values of the working voltage and current of the particle beam, wherein the smallest distance of the first point from the edge of the sample is also the distance corresponding to the minimum thickness of the protective layer.
3. The method of delayering the area of interest in the sample according to claim 1 , wherein the thickness of the protective layer is determined by means of an artificial neural network, which is trained using training data containing information on the position of the first point and the length of the straight section at the selected tolerance value of the straightness of the straight section in the horizontal profile for the set values of the working voltage and current of the particle beam, wherein the smallest distance of the first point from the edge of the sample is also the distance corresponding to the minimum thickness of the protective layer.
4. A method of delayering the area of interest in the sample according to any one of claims 1 to 3, wherein the device for performing the delayering method further comprises a second column having a second optical axis and generating a second particle beam distributed along the second optical axis, positioned such that the first optical axis and the second optical axis intersect in proximity of the area of interest.
5. The method of delayering the area of interest in the sample according to claim 4, which further comprises, after the step d) of forming the protective layer adjacent to the area of interest and before the step e) of placing the surface to be delayered of the area of interest into a position substantially parallel to the optical axis of the column, wherein the distance between the surface to be delayered of the area of interest andthe optical axis in the focal plane is less than half the diameter of said beam of charged particles in the focal plane, a step of forming a mark into the protective layer into a plane substantially parallel to the surface to be delayered, which is in a plane substantially parallel to the surface to be delayered in the field of view of the second column when the sample is scanned by the charged particle beam.
6. The method of delayering the area of interest in the sample according to any one of claims 1 or 5, wherein the focal plane passes between the sample and the column.
7. The method of delayering the area of interest in the sample according to claim 6, wherein the focal plane passes through the protective layer.
8. The method of delayering the area of interest in the sample according to any one of the preceding claims, wherein the charged particle beam is an ion beam.