Method, device and computer program for repairing a mask defect

Abstract

Method (100, 200) for repairing a defect in a mask (10, 20) for lithography, in particular an EUV mask, comprising: a. performing a first repair step (120) on the defect with a first repair dose (125), wherein the defect is thereby transformed from an initial topology (110) to a first defect topology (130); b. determining (140) an influence of the first repair step on the topology of the defect; c. determining a second defect topology (160) for the defect, which is to be achieved by a second repair step (150) on the defect; and d. determining a second repair dose for the second repair step, at least partially based on the determined influence of the first repair step on the topology of the defect and the second defect topology, wherein the topology of the defect comprises a height of the defect depending on a position on the mask; wherein step b., Determining the influence of the first repair step on the topology of the defect, includes determining a change in the topology of the defect caused by carrying out the first repair step, in particular a comparison of the first defect topology with the initial topology; wherein: step d., determining the second repair dose for the second repair step, furthermore taking into account one or more calibration curves for different defect types, which allow an estimation of the repair behavior of the defect; and / or the first process speed is determined as a positional factor; and / or wherein the defect corresponds to a known defect type for which one or more calibration curves are known, and wherein the method further comprises drawing conclusions about the stability of the method and / or a device on which the method is carried out from the repair behavior of the defect; and / or wherein reference process speeds are known for different materials and wherein the method further comprises drawing conclusions about the material composition of the defect from its repair behavior.

Description

1. Technical field

[0001] The present invention relates to a method, a device and a computer program for repairing a defect in a lithography mask, in particular a defect in a mask for EUV lithography (hereinafter referred to as "EUV mask"). 2. State of the art

[0002] As a consequence of the constantly increasing integration density in microelectronics, lithographic masks (hereafter often referred to simply as "masks") must image increasingly smaller structural elements onto the photoresist layer of a wafer. To meet these requirements, the exposure wavelength is shifted to ever shorter wavelengths. Currently, argon fluoride (ArF) excimer lasers, which emit light at a wavelength of 193 nm, are primarily used for exposure purposes. Intensive research is underway on light sources that emit in the extreme ultraviolet (EUV) wavelength range (10 nm to 15 nm) and corresponding EUV masks. To increase the resolution of wafer exposure processes, several variants of conventional binary lithographic masks have been developed concurrently. Examples include phase-shifting masks and masks for multiple exposures.

[0003] Due to the ever-shrinking dimensions of the structural elements, lithographic masks cannot always be produced without defects visible or printable on a wafer. Because of the costly manufacturing process, defective masks are repaired whenever possible.

[0004] Two important groups of defects in lithographic masks are, on the one hand, dark defects, and on the other hand, clear defects.

[0005] Dark defects are areas where absorber or phase-shifting material is present but should be free of this material. These defects are repaired by removing the excess material, preferably using a local etching process.

[0006] Clear defects, on the other hand, are defects on the mask that, when optically exposed in a wafer stepper or wafer scanner, exhibit greater light transmittance than an identical, defect-free reference position. In mask repair processes, such clear defects can be remedied by depositing a material with suitable optical properties. Ideally, the optical properties of the repair material should match those of the absorber or phase-shifting material.

[0007] The applicant develops and manufactures measuring systems for analyzing lithographic masks. Furthermore, the applicant develops and distributes repair devices for lithographic masks.

[0008] To remove dark defects, it is known to use an electron beam directed at the defect to be repaired. The use of an electron beam allows for particularly precise guidance and positioning of the beam on the defect. In conjunction with a precursor gas, also called a process gas, which can be introduced into the atmosphere of the mask being repaired, the incident electron beam can induce a reaction similar to a local etching process. This induced local etching process removes excess material (from the defect) from the mask, thus creating or restoring the desired absorber and / or phase-shifting properties for the lithographic mask.

[0009] Alternatively, the precursor gas used can be selected in such a way that a deposition process can be induced upon irradiation with the beam. As a result, additional material can be deposited at clearly defined defects to locally reduce the light transmission of the mask and / or increase its phase-shifting properties.

[0010] One possible method for repairing phase-shifting defects in a photomask is described, for example, in document US 6,593,040 B2. The method involves scanning the mask for defects and determining the location of at least one defect. The defect is then analyzed three-dimensionally, and the result of this analysis is used to direct a focused ion beam (FIB) at the defect to remove it. The FIB is controlled by an etching card, which was generated based on the results of the three-dimensional analysis. Furthermore, test patterns can be created with the FIB and analyzed three-dimensionally; these patterns are then used to further generate the etching card.

[0011] One disadvantage of the known methods, however, is that they either do not consider the dynamics of the repair process at all or only to an insufficient degree. For example, in the method mentioned above, the three-dimensional analysis is only performed once before the etching process. However, the way a defect reacts to ongoing repair measures can change during the repair process. This can lead to the pre-calculated repair steps not producing the expected results, and therefore the repair is not successful as desired (or the mask is even damaged).

[0012] The present invention therefore aims to provide a method that at least partially mitigates or eliminates the disadvantages of known methods and / or allows for a more precise and safer repair of mask defects. Furthermore, a corresponding device and a computer program with instructions for carrying out such a method are to be provided. 3. Summary of the invention

[0013] The above-mentioned problems are at least partially solved by the various aspects of the present invention, as described below.

[0014] In one embodiment, a method for repairing a defect in a lithography mask, in particular an EUV mask, comprises the following steps: (a.) performing a first repair step on the defect with a first repair dose, whereby the defect is thereby transformed from an initial topology to a first defect topology, (b.) determining an influence of the first repair step on the topology of the defect, (c.) determining a second defect topology for the defect to be achieved by a second repair step on the defect, and (d.) determining a second repair dose for the second repair step at least partially based on the determined influence of the first repair step on the topology of the defect and the second (intended) defect topology.

[0015] In mask processing, particularly mask repair, one or more gases are typically directed to a specific reaction site and irradiated with a beam of energetic particles (e.g., photons, electrons, or ions) to initiate the desired processing steps (e.g., etching or deposition, as mentioned earlier and described in more detail below). The applied repair dose influences the outcome of each repair step.

[0016] According to the present procedure, an initial repair step with a first repair dose is performed on the defect to be repaired (for example, a clear or a dark defect as mentioned at the beginning). During this step, the defect changes; in particular, it transitions from an initial topology to a new topology, which is referred to here as the first defect topology.

[0017] In the simplest case, the topology is described by only one or a few parameters, such as the height of the defect. However, as described in more detail below, the topology of the defect can also be described and characterized more precisely within the framework of the present method, for example, by the position-dependent height of the defect, its lateral extent, its three-dimensional structure, etc., which can increase the accuracy of the mask repair.

[0018] After performing this first repair step, the described procedure is used to determine how the first repair step affected the defect. In particular, it can be determined how the topology of the defect changed as a result of the first repair step, or what influence the first repair step had on the topology of the defect. Determining the topology of the defect, or its changes, can be done in a manner generally known to those skilled in the art, for example, by using a scanning probe microscope, in particular an atomic force microscope, a profilometer, or another suitable device (see below). Interferometric measurement methods are also conceivable in principle, although these typically do not have sufficient lateral resolution.

[0019] This step is based on the understanding that when repairing defects on masks, a number of factors generally play a role that can influence the result of a repair step at a given repair dose, and the exact behavior of the defect is therefore not always precisely predictable. Or, put another way, the repair dose required for a specific repair success generally depends on a number of (defect) parameters, not all of which necessarily need to be known. Besides the topology of the defect itself, these can include, for example...The properties of the immediate defect environment on the mask can be relevant (for example, areas of the defect adjacent to the mask's absorber material may behave differently than areas adjacent to the mask's quartz material), the location of the defect on the photomask (for example, an area with a lot of quartz material behaves differently than an area with a lot of absorber material, and highly structured regions may behave differently than weakly structured regions), or the electrical charge and / or temperature of the mask surface. Especially for defects that need to be ablated, the material composition of the defect can also play a role. This composition may be inherent to the mask type itself (for example, different materials and / or layers of material that make up the mask) or may be influenced by contaminants or foreign substances on the mask or the defect.

[0020] All these factors can change during mask repair, and consequently, the required repair dose can also change during the process (in a way that is initially unpredictable). The repair dose must therefore be selected and adjusted during the mask repair process to ensure that the defect is repaired as desired and that damage to the photomask is minimized. According to the present method, the defect is therefore not repaired in a single step for which a required repair dose has been determined in advance. Instead, after the first repair step has been performed, its influence on the defect's topology is determined, and information is obtained from this to control the subsequent mask repair process.

[0021] More precisely, after determining the influence of the first repair step on the defect topology, a second defect topology is determined, which is to be achieved with a second repair step on the mask. As explained in detail below, the second defect topology can already be the final topology to be achieved (i.e., the desired repair success), or the second defect topology can itself represent an intermediate goal of the process, and the process comprises further iterations of the approach described here until the desired end result is achieved. Determining the second defect topology to be achieved can be done automatically or manually. Hybrid forms are also possible. A second defect topology or a sequence of defect topologies can also be determined before the process begins (if the process is to be iterative, see below).(below) have been specified, which, for example, result from empirical data from repairs of similar or identical defects, and which are then to be achieved sequentially using the procedure.

[0022] Once the second defect topology to be achieved in the subsequent second repair step has been determined, the necessary repair dose is calculated, taking into account the influence of the first repair step on the defect topology and the desired second defect topology. Other factors influencing the second repair step can, of course, also be considered at this point, such as those mentioned above.

[0023] Determining the repair dose required for the second repair step is therefore carried out by taking into account the information obtained from the first repair step, so that, according to the present method, a dynamic adaptation of the mask repair process to the possibly changing conditions, to (changing) properties of the defect currently being processed, or to other (uncontrolled and / or controllable) process parameters is made possible.

[0024] The procedure can then further include step (e.) of carrying out the second repair step with the second repair dose.

[0025] Performing this second repair step will change the defect topology once again. Ideally, the resulting topology will correspond to the target topology for the second repair step, which was used to determine the second repair dose. However, due to various factors, the defect topology after the second repair step may not exactly match the intended topology. If the topology generated by the second repair step was not intended as an intermediate goal of the process, further repair measures may be necessary to achieve the desired repair result. In any case, after the second repair step, the defect will have a topology that will henceforth be referred to as the second (actual) defect topology.

[0026] From the above explanations, it follows that the first repair step can be carried out, in particular, as a calibration step. For example, to avoid damage to the mask, e.g., due to an unpredictably "aggressive" course of the first repair step, the first repair step can be carried out, in particular, as a calibration step in which less than 50%, preferably less than 30%, and especially preferably less than 15% of the desired final repair success is achieved or intended to be achieved. The desired final repair success will, in most cases, represent the complete elimination of the defect in the mask; however, other cases are also conceivable and are included.

[0027] In particular, the repair dose for the first repair step (possibly dependent on the position) can be chosen so that the above percentage values ​​are not exceeded. In case of doubt, a more "conservative" value for the repair dose in the first repair step should be chosen, as this is the best way to avoid damage to the mask (for example, from over-etching).

[0028] Alternatively, an initial repair dose can be determined for the first repair step, aiming to achieve more than 50% of the desired final repair success. This minimizes the number of subsequent iterations and thus the overall process duration. At the same time, any deviations can be addressed in one or more second steps. The choice of the initial repair dose can, for example, be determined empirically based on similar repairs.

[0029] As already indicated, step (b.) of determining the influence of the first repair step on the defect topology can include determining a change in the defect topology caused by performing the first repair step. In particular, the step can involve comparing the first defect topology with the initial topology.

[0030] If necessary, this may include determining the initial topology before carrying out the first repair step.

[0031] Within the scope of this procedure, changes in the defect topology can be specified and considered at various levels of detail. As mentioned earlier, the topology can, for example, be described simply by a value for the defect's height / thickness (e.g., a height / thickness averaged over the defect's base area, a maximum height / thickness, or a minimum height / thickness), and in this simple case, the change in topology can be characterized solely by a change in the defect's height / thickness. However, by comparing the initial defect topology with the initial topology, far more detailed information about the topology change can be obtained. The more precise the information thus obtained, the more accurately the second repair dose can generally be determined (possibly even position-dependent, cf.(further below), which is necessary to achieve the desired second defect topology. However, this can increase the analysis effort of the procedure, so a trade-off must be made here.

[0032] Step (b.) of determining the influence of the first repair step on the topology of the defect may in particular include determining a first process rate of the first repair step at least partially based on the determined change in the topology of the defect and the first repair dose.

[0033] The determined change in the defect's topology, combined with the first repair dose, can be used to determine how quickly the initial repair step alters the defect's topology and, for example, removes or deposits material. For an etching process, it's possible to determine the extent to which the first repair step, applied with the initial repair dose, etches away material from the defect. For a deposition process, it's possible to determine how quickly, or how much, the first repair step, applied with the initial repair dose, deposits material onto the mask. This information can then be used to determine "how much more needs to be done" and thus inform the calculation of the second repair dose.Since the initial process speed of the first repair step was determined on the defect currently being processed, it can automatically take into account the aforementioned additional factors that may influence the process speed, allowing for particularly precise and also "dynamic" control of the further process sequence.

[0034] The topology of the defect can, in particular, include the height of the defect depending on its position on the mask.

[0035] Such a “height map” of the defect allows, for example, the determination of the repair dose for the subsequent repair step(s) with spatial resolution, as it provides spatially resolved information about the influence of the first repair step (or further iteratively performed repair steps) on the defect and thus allows a location-dependent calibration of the process.

[0036] The topology of the defect may also include, for example, one or more of the following information: a lateral extent of the defect, a three-dimensional structure of the defect.

[0037] As already mentioned, the inclusion of further properties or information about the topology of the defect will generally allow for even more precise control of the process, but will also increase its complexity and resource consumption, so that a balance must or can be struck between the complexity and accuracy of the procedure.

[0038] Step (d.) of determining the second repair dose for the second repair step may include one or more calibration curves for different defect types, which allow an estimation of the repair behavior of the defect.

[0039] Beyond the "dynamic" calibration of the process based on the topology change of the defect currently being repaired (hereinafter often referred to as the "current defect"), the present method can also utilize reference data / calibration curves to further increase the accuracy and / or speed of the process. For example, these calibration curves can be used to adjust the repair dose for the subsequent repair step(s) to achieve the desired process speed. Thus, the topology change experienced by the current defect during the first repair step can be compared with one or more calibration curves for a defect type that corresponds to or is similar to the current defect, in order to verify whether the current repair process is behaving as expected or deviating from the reference values.In the latter case, the calibration data from the curves can be incorporated into the determination of the second repair dose for the second repair step. For example, the determined actual influence of the first repair step on the defect topology can be averaged with values ​​from the calibration curves, or a deviation from the expected behavior can be determined, and the magnitude of this deviation is then taken into account when determining the second repair dose (e.g., as a correction factor to the second repair dose initially determined without considering the calibration curves). A weighted sum of the determined actual influence and one or more values ​​from the calibration curve(s) can also be used. Furthermore, such stored reference data can also be used to eliminate intermediate steps, i.e.,to be able to achieve the desired end topology faster than would be possible if one had to "blindly feel one's way" towards the desired repair result.

[0040] In this process, one or more calibration curves can be created manually or automatically during the current run and / or one or more previous runs of the procedure.

[0041] The calibration curves created in one or more previous iterations of the process can, for example, be stored in a database that is accessed during the current iteration. Other data storage methods are, of course, also possible. However, particularly if the process is iterative, as described in more detail below, it is also possible that the calibration curves from previous steps of this iteration process originated within the current repair process itself, and therefore can particularly well reflect the properties and characteristics of the current repair process.

[0042] For a defect spanning multiple material layers, the procedure can proceed as follows: The first repair step with the first repair dose can be performed on a first material layer containing a first material, and the second repair step with the second repair dose can be performed on a second material layer containing a second material that differs from the first material. Furthermore, the first repair step's process rate can be determined with respect to the first material, and the second repair dose can then be determined by considering the thickness of the second material layer and the process selectivity of the second repair step on the second material relative to the first repair step on the first material.

[0043] Taking the thickness of the second material layer into account when determining the second repair dose is optional, i.e., the option explained here can also be applied without including the thickness of the second material layer as a relevant factor.

[0044] In any case, the first repair step on the first material layer, which contains the first material, can be used to extrapolate the behavior and effects of the repair process on the second material layer, and thus to determine and select a suitable second repair dose for treating this second material layer in the second repair step. The process selectivity of the ongoing reaction (e.g., etching or deposition process) with respect to the first and second materials plays a particularly important role here, e.g., the process rate or rate at which the repair process proceeds under otherwise identical conditions with respect to the first and second materials.

[0045] If the process selectivity with respect to the different materials is unknown, it may have been determined during the current run of the process and / or one or more previous runs of the process from the repair process itself. For example, the second repair step may be preceded by a further step in which the second material layer is "tested" (e.g., a processing step with a small repair dose but of the same type as the second repair step to be performed) and the process selectivity with respect to the second material is thus determined before the actual repair is then continued with the second repair step.

[0046] As already mentioned, the second defect topology can already represent a desired final topology. This means that the present method may achieve the desired repair success (or at least come very close to it) upon completion of the second repair step. This is made possible, in particular, by calibrating the second repair step on the actual defect itself, namely by determining the influence that the first repair step had on the defect's topology.

[0047] However, it is also possible that the second defect topology is merely an intermediate goal of the procedure, and that steps (b.) to (e.) are iteratively repeated within the procedure until a desired final defect topology is reached.

[0048] That is, after performing the second repair step with the second repair dose, the influence of the second repair step on the defect topology is determined (and / or an averaged or weighted influence of the first and second repair steps). Then, a third defect topology is defined for the defect, which is to be achieved with a third repair step. A third repair dose for the third repair step is then determined, at least partially, based on the determined influence of the second repair step on the defect topology (and / or based on the aforementioned averaged or weighted influence) and the third defect topology. The third repair step can then be performed with the third repair dose.

[0049] Generally, after performing the i-th repair step with the i-th repair dose, the influence of the i-th repair step on the defect topology can be determined. Then, an i+1-th defect topology can be defined for the defect, which is to be achieved with an i+1-th repair step. An i+1-th repair dose for the i+1-th repair step can be determined, at least partially, based on the determined influence of the i-th repair step on the defect topology and the i+1-th defect topology. Finally, the i+1-th repair step can be performed with the i+i-th repair dose. And so on.

[0050] When determining the i+1th repair dose for the i+1th repair step, the information and insights—i.e., the "calibration input"—from the i-th iteration just completed can be included, as well as calibration input from one or more earlier iterations (i.e., from the i-1th iteration and / or the i-2th iteration and / or the i-3rd iteration, etc., if available), so that the calibration input considered can increase with each iteration. For example, the input from the different iterations can be averaged or weighted according to "recognition."

[0051] Verification of the achieved final topology can also be part of the process.

[0052] This verification can, for example, include checking whether the defect topology resulting from the most recent repair step actually matches the topology that was specified as the target for the repair step. If so, the procedure can be considered complete; if not, further iterations may be necessary.

[0053] The first and / or second repair dose can vary depending on the position. The same applies, in the case of an iterative process, to a third and / or fourth repair dose, etc., that may be included in the process.

[0054] By varying the repair dose(s) depending on the position, the processing of the defect can be controlled with pinpoint accuracy, ensuring that only as much material is removed or deposited at each point of the defect as is necessary to achieve the desired defect topology after completion of the repair process.

[0055] The first and / or second repair step may include an etching process and / or a deposition process. The same applies, in the case of an iterative process, to a possible third and / or fourth repair step, and so on.

[0056] As mentioned at the beginning, mask defects can typically be classified into one of two types: defects where excess absorber material is present in places where it should not be, or defects where, conversely, absorber material is missing. These two types of defects correspond to specific typical repair measures: removing the excess material, particularly by etching it away, or deposition of absorber material in the areas where it is lacking.

[0057] The first process speed can be determined, in particular, as a function of position. The same applies, in the case of an iterative process, to a second and / or third process speed, etc., that may be included in the process.

[0058] This can be used, in particular, to determine a position-dependent repair dose for the next repair step, as just described, so that it can be controlled with pinpoint accuracy to achieve the intended defect topology as precisely as possible. To determine the process speed in a position-dependent manner, the change in topology can generally be measured in the most recent repair step, but position-resolved information from earlier iterations (if available) can also be used.

[0059] It is advantageous if the process takes place under constant external conditions, especially in a vacuum.

[0060] If this is not the case, as in prior art methods where one or more intermediate checks of the repair progress may be possible, but the mask being processed must be removed from the processing environment to be fed into an analysis tool, then the mask surface and, in particular, the defect surface or composition of the defect can change during such an analysis, for example, due to oxidation processes. As a result, the findings obtained before or during the analysis may already be outdated when the repair process is resumed and no longer accurately reflect the prevailing conditions. In contrast, the present method can be carried out under constant conditions, especially in a vacuum, thus avoiding such effects that can lead to errors and inaccuracies in the repair process.

[0061] The procedure can be carried out in particular on a combined repair and analysis device.

[0062] This goes hand in hand with the aforementioned possibility of running the process under consistent conditions and favors such a procedure, as the mask does not need to be removed from the device and, for example, fed to a separate analysis device. Furthermore, this can simplify and accelerate the repair process. As an example, the combined repair and analysis device can be designed in a combination of a repair system distributed by the applicant with an atomic force microscope.

[0063] If the defect corresponds to a known defect type for which one or more calibration curves are known, the method may also include drawing conclusions about the stability of the method and / or the stability of a device on which the method is carried out, based on the repair behavior of the defect.

[0064] Conclusions can be drawn taking the calibration curves into account.

[0065] It was mentioned and explained earlier how stored calibration data and calibration curves can be used to verify whether the current repair process is behaving as expected (for example, based on the defect type of the defect currently being addressed and reference data recorded for that defect type), in order to optimize and accelerate the further repair process. Conversely, such a comparison with existing reference data / calibration curves can also serve the purpose of investigating the stability of the ongoing process against known or unknown influences.

[0066] If, for example, the current defect corresponds to a well-researched and known defect type, and the repair behavior of the current defect deviates significantly from the calibration curve(s), this may indicate that the ongoing process and / or the device on which it is running is highly susceptible to such influences. Appropriate countermeasures can then be taken (e.g., checking the temperature, vacuum, electrical charge of the mask, etc.) to avoid jeopardizing the success of the repair on the mask in question.

[0067] If reference process speeds are known for different materials, the method can also include drawing conclusions about the material composition of the defect from its repair behavior.

[0068] The conclusions can be drawn taking into account the reference process speeds.

[0069] For example, using a constant repair dose, several or more repair steps can be performed on the defect, and the respective changes in the defect topology can then be used to draw conclusions about the process speed of each repair step. From this, by comparing with known reference process speeds, the material composition of the defect (e.g., at the location of the respective repair step) can be derived.

[0070] Another aspect is a computer program with instructions which, when executed, cause a device for repairing a defect in a lithography mask, in particular an EUV mask, to carry out a procedure as described herein.

[0071] In this process, various aspects and features of the procedure described herein may have been combined in the program, and individual aspects may also be omitted if they are unnecessary for achieving a desired result.

[0072] The computer program may, in particular, include instructions which, when executed, cause the device for repairing the defect to automatically perform the first repair step as a calibration step on the defect with the first repair dose, to automatically determine the influence of this calibration step on the topology of the defect, and, at least partially based on the determined influence of the calibration step and the first repair dose, to automatically determine the second repair dose so that the second repair step can be performed with it in order to achieve the second defect topology.

[0073] Manual user intervention may therefore be eliminated, which can speed up mask repair and / or reduce its susceptibility to errors. This is also advantageous for automating the process.

[0074] Another aspect is a device for repairing a defect in a lithography mask, in particular an EUV mask, wherein the device is configured to carry out a method as described herein.

[0075] Here too, the device can be designed in such a way that, in the embodiment that follows, various aspects and features of the method described herein are combined, and individual aspects of the described method can also be omitted if they are unnecessary for achieving a desired result.

[0076] The device may comprise the following: means for performing a first repair step on the defect with a first repair dose, whereby the defect is thereby transformed from an initial topology to a first defect topology; means for determining an influence of the first repair step on the topology of the defect; means for determining a second defect topology for the defect, which is to be achieved with a second repair step on the defect; and means for determining a second repair dose for the second repair step, at least partially based on the determined influence of the first repair step on the topology of the defect and the second defect topology.

[0077] As already mentioned, the device can be a combined repair and analysis device. This can, in particular, allow the process to take place completely or at least largely under constant external conditions, especially in a vacuum.

[0078] As an example, the above-mentioned means can be configured in a combination of a repair system distributed by the applicant with an atomic force microscope.

[0079] The device can also include a memory in which the aforementioned computer program is stored, as well as means for executing the instructions contained therein, so that when the instructions are executed, the device automatically performs the first repair step as a calibration step on the defect with the first repair dose, automatically determines the influence of this calibration step on the topology of the defect, and, at least partially based on the determined influence of the calibration step and the first repair dose, automatically determines the second repair dose so that the second repair step can be performed with it in order to achieve the second defect topology.

[0080] The advantages of such an automated process have already been pointed out, and reference is therefore made here to the corresponding implementation.

[0081] Finally, the following should be noted: Up to this point, the invention has been primarily described and discussed in the context of repairing a mask defect, which represents an important application of the disclosed teaching. However, it should be pointed out that the disclosed teaching is not fundamentally limited to this, but can also be applied to the (surface) processing of other objects used in the field of microelectronics, e.g., for modifying and / or repairing structured wafer surfaces or microchip surfaces, etc. These further application possibilities are therefore always also included in the disclosed teaching, even if not explicitly mentioned, unless they are explicitly excluded or physically or technically impossible. List of characters

[0082] The following detailed description describes possible embodiments of the invention with reference to the following figures: Fig. Figure 1 shows an exemplary embodiment of the method described herein, in which excess absorber material is removed; Fig. Figure 2 shows the results of applying an embodiment of the method described herein to a defect in a mask with a lines-and-spaces structure; Fig. Figure 3 shows conceivable process flows for a possible implementation of the method described herein, which can be carried out using components of a combined analysis and repair device. 5. Detailed description of possible embodiments

[0083] In the following, embodiments of the present invention are described primarily with reference to the repair of defects in lithographic masks. For the sake of completeness, however, it should be mentioned again that the invention is not limited to this application but can also be used in principle for other types of mask processing, or more generally for the surface processing of other objects used in the field of microelectronics, e.g., for modifying and / or repairing structured wafer surfaces or microchip surfaces, etc. Although the following description focuses primarily on the application of repairing mask defects in order to keep it clear and easier to understand, the other possible applications of the disclosed teaching remain familiar to those skilled in the art.

[0084] It should also be noted that only individual embodiments of the invention can be described in more detail below. However, those skilled in the art will understand that the features and modifications described in connection with these embodiments can be further modified and / or combined in other combinations or sub-combinations without this taking the invention outside the scope of the present invention. Individual features or sub-features can also be omitted if they are unnecessary for achieving an intended result. To avoid unnecessary repetition, reference is therefore made to the explanations and statements in the preceding sections, which also apply to the detailed description that follows.

[0085] The Fig. Figure 1 schematically shows an embodiment 100 of the described method for repairing a defect in a mask 10 by etching away excess absorber material.

[0086] At the beginning of procedure 100, the defect has an initial topology 110. The initial topology 110 may already be known, for example from previous analyses of the mask 10, or it can initially be determined within the framework of procedure 100, for example using an atomic force microscope.

[0087] The defect is then treated in a first repair step 120 with a first repair dose (sometimes also simply called dose) 125. In this case, the first repair step 120 is an etching process; that is, for example, one or more precursor organases are introduced into the mask area containing the defect, where they are irradiated, for example, with a beam of energetic particles, causing them to react and etch away the absorber material. In the case shown here, the repair dose 125 varies depending on the position; that is, the irradiation is carried out according to a so-called "dose map." This dose map can already be adapted to a certain extent to the initial topology 110 of the defect.

[0088] After performing the first repair step 120, the topology, which is described and characterized here by a spatially resolved height map of the defect, has changed, and the defect now has a first defect topology 130. As described below, the first repair step 120 not only serves the actual repair of the mask 10, but also functions as a calibration step to control and optimize the further process flow.

[0089] Thus, after the first repair step 120 has been carried out, an analysis step 140 examines the influence of this first repair step 120 on the topology of the defect. In this case, this is done by determining the change in the topology of the defect, in particular the spatially dependent change in the height of the defect, by comparing the initial topology 110 with the first defect topology 130. This results in a spatially resolved change profile of the topology with respect to the first repair step 120. Taking into account the repair dose 125 used, a kind of resolved process speed of the etching process can be determined, which is then used for further process control.

[0090] Thus, after performing analysis step 140, a second defect topology is determined, which is to be achieved in a subsequent second repair step, indicated by arrow 150. For example, the second defect topology may already correspond to the final desired repair result, which in this case involves removing the excess absorber material down to the quartz level of mask 10. However, the second defect topology may also only represent an intermediate goal of the process, e.g., if the process is designed from the outset to involve several iterations.

[0091] Based on the information obtained in analysis step 140, i.e., based on the determined influence of the first repair step 120 on the topology of the defect, in particular the determined process rate of the first repair step 120, as well as the desired second defect topology, a second repair dose for the second repair step 150 is then determined, and the second repair step 150 is carried out accordingly.

[0092] Ideally, performing this second repair step 150 leads exactly (or within certain acceptance intervals) to the desired result, i.e., the resulting defect topology corresponds exactly to the desired second defect topology. As in Fig. As indicated in 1, the actual second defect topology 160 achieved when carrying out the second repair step 150 may still deviate from the desired result, so that further rework and / or additional repair steps may be necessary.

[0093] Whether this is the case can be determined, for example, by verifying the defect topology 160 achieved after the second repair step 150, for example by comparing the defect topology 160 with certain acceptance intervals or the like.

[0094] If further processing is necessary, the steps described above can be repeated iteratively until the desired repair result is achieved.

[0095] Fig. Figure 2 shows a schematic representation of the results and the process of a mask repair procedure 200 according to the teaching disclosed herein, using a mask 20 with a lines-and-spaces structure. Excess absorber material was located in one of the spaces, which was removed during the process 200.

[0096] Method 200 was carried out on a combined repair and analysis apparatus under constant external conditions, namely in a vacuum. The apparatus comprised a scanning electron microscope (SEM) that performed the repair steps by using its electron beam to trigger the etching steps at the defect site by introducing a suitable etching gas (or gas mixture). Simultaneously, the SEM acquired images of the mask in the vicinity of the defect site. These images acquired by the SEM are shown in the partial figures designated with reference numbers 210, 220, and 240 of the Fig. 2 each shown at the top.

[0097] For a more detailed analysis of the defect topology, an atomic force microscope (AFM) was also used, the corresponding images of which are shown in subfigures 210, 220 and 240 of the Fig. Figures 2 and 2 are shown, one in the middle and one below. The middle image shows mask 20 in plan view (analogous to the SEM images) and with an elevation profile in the form of equidistant contour lines (analogous to a topographic hiking map). Below, in sub-figures 210, 220, and 240 of the Fig. Figure 2 shows a section through the corresponding elevation profile measured by the AFM along section line 28.

[0098] In subfigure 210, the initial defect is shown as it appeared in a SEM image (top of subfigure 210) and an AFM image (middle of subfigure 210, with contour lines). As mentioned previously, a height profile was extracted from the AFM image along the horizontal line 28 (bottom of subfigure 210). In the area of ​​mask 20 marked with reference symbol 21, a defect in the form of excess absorber material is visible on one of the spaces located between two lines 25.

[0099] A first repair step was performed on this defect topology using a predefined dose map (repair dose), which served to calibrate the process speed.

[0100] The repair dose in this calibration step can be adjusted not only by changing the current intensity of the electron beam, but alternatively or additionally by varying the number of times the electron beam strikes a specific processing point in a repair step. Other methods for controlling and changing the repair dose are also expressly included in this teaching.

[0101] After performing this calibration step / first repair step, the situation resulted as shown in the middle subfigure designated with reference number 220. Fig. Figure 2 shows: In the area of ​​mask 20 marked with reference symbol 22, a (weak) defect in the form of excess absorber material is still visible in the space in question.

[0102] By reacquiring one or more AFM images (see center and bottom of subfigure 220), the change in defect topology caused by the first repair dose could be determined with spatial resolution. From this change, the dose map used in the calibration step, and a target topology for the subsequent second repair step, a new dose map was calculated to achieve this target topology. This dose map was then applied to the defect in a further repair step, resulting in Fig. 2 is indicated by the arrow 230.

[0103] It should be noted that the processing indicated by arrow 230 may also include several analysis and repair steps, whereby the dose map for each step can be calculated from the previous dose map(s), the previous topology change(s) and a target topology specified as a repair goal, either individually or jointly.

[0104] The process 200 can certainly be designed to run in several iterations; that is, a defect topology targeted in a specific repair step does not necessarily have to correspond to the final desired repair outcome, e.g., the complete removal of the excess absorber material. Rather, a given repair step can also serve, even after the first repair step has already been completed, specifically for the further fine-tuning and adjustment of the repair process 200, so that the information obtained in the first and this repair step (and possibly information from further repair steps serving calibration) can ultimately be used in the subsequent repair step(s) to achieve an extremely precise repair of the defect in the mask 20. For example, this should prevent over-etching of the defect.

[0105] To reiterate, a particular targeted defect topology (e.g., the second defect topology and / or the third defect topology and / or the fourth defect topology, etc., in the case of an iterative process flow) does not necessarily always have to correspond to the desired end result of the repair process 200 (although this can, of course, also be the case), but can also merely represent an intermediate goal of the process. In such a case—and this applies generally to the teaching disclosed herein and not only to the embodiment 200 currently described—the intermediate topology of the defect targeted in a repair step can also be specified in less detail than, for example, the targeted end result of the repair process might be.For example, the desired intermediate topology of the defect can be specified only to the extent that no more than a certain percentage of the remaining defect size (e.g., defect height and / or defect width) should be removed (or deposited in the case of a deposition process) in order to avoid damaging the mask and / or over-repairing. For instance, it can be specified that the desired defect topology should achieve no more than 50%, 30%, or 15% of the remaining repair success. In other words, a repair step (even after the first repair step) can be specifically performed as a (further) calibration step, for example, to better understand and calibrate the dynamics of the repair process, while simultaneously avoiding damaging the mask and / or over-repairing, as is done here.

[0106] Here, spatially resolved measurements of the defect height change at each repair step, performed with AFM, combined with the respective dose map, allow for the determination of a spatially resolved process speed of the etching process. Furthermore, in conjunction with the remaining defect height, which can also be determined spatially resolved by AFM, a new dose map or repair dose (e.g., number of loops) can then be determined in an extrapolation step to achieve the desired defect topology in the subsequent step and / or ultimately.

[0107] The right-hand part of figure 240 of the Fig. Figure 2 shows the defect (or what remains of it) after the second repair step (and any subsequent repair steps) 230. The combination of SEM and AFM images served to verify the success of the repair.

[0108] In the case shown here, the two AFM images (in the middle and below in subfigure 240) are used to determine the result. Fig. 2) to recognize that a residual defect remained, which could have been eliminated by a further repair step. However, if the topology of the residual is within the specification limits for a successful repair, the repair can already be considered successfully completed. In particular, if the remaining residue is below a certain threshold (e.g., does not lead to a defect in the wafer during the actual irradiation process using mask 20), it can remain on mask 20.

[0109] The Fig.Figure 3 shows conceivable process sequences 300 for a possible implementation of the method described herein, which can be carried out, for example, by means of components of a combined analysis and repair device, which can be realized in a combination of a repair system (e.g. with a SEM) and an analysis system (e.g. with an AFM).

[0110] At the start of the repair process (310), the dose or repair dose (320) can be determined for a mask being processed. The target topology (315), the current topology (330), and the process speed (325) can all be factored into the dose determination. In a first iteration, the process speed (325) can be set by user input or predefined as a system default (e.g., based on a reference value for a similar repair, a calibration curve, etc.). However, the process speed, particularly in a second or subsequent iteration, can also be determined from a measurement of one (or more) previous topology changes (375) and one (or more) previous doses (365), or may already be known.

[0111] The dose can be defined as the number of times the electron beam of a SEM strikes a processing point during the repair step. However, other definitions of dose are also conceivable.

[0112] The determined dose 320 can then be applied in a first (partial) repair step 335 to repair a defect, for which a SEM, which can be part of the repair device used, is particularly suitable. The mask repair can be performed in one step or in several (e.g., iterative) steps. The first (partial) repair step 335 can also be performed as a calibration step, meaning that the initial defect topology to be achieved and the corresponding initial dose can be chosen and determined relatively conservatively to avoid mask damage or over-repair. After the repair step has been carried out, post-processing steps can be performed if necessary.

[0113] In the next step 340, a decision can be made as to whether the repair should be completed 345 or whether a next iteration 350 of the repair process should take place.

[0114] If another iteration 350 of the repair process is to be performed, the new topology 355, or the new defect height of the mask defect, can then be determined. However, it is also possible to determine the new topology even if no further iteration 350 is planned, for example, to check the final result of the repair. If, unexpectedly, this result is not as desired, it would also be possible to perform another iteration 350 and proceed accordingly.

[0115] The new topology 355 can be stored as the (new) current topology 330. Based on the new topology 355 and the information from one or more previous iterations 360, e.g., a previous topology 370, a topology change 375 achieved by the first iteration or repair step can be determined. The information thus obtained about the topology change 375 can be considered in isolation or used in conjunction with a previous dose 365 (from one or more previous iteration or repair steps) to determine a (new) process rate 325.

[0116] Now, in step 320, the dose for the next (partial) repair process can be determined. This can be done based on the (new) process speed 325, the (new) current topology 330, and the target topology 315. Reference data from calibration curves stored by the system and / or dynamically generated ones, etc., can also be incorporated at this point.

[0117] Afterwards, the next repair step 335 can be carried out with the dose determined in this way and, if necessary, further (partial) repair steps 335 can be performed.

[0118] In this way, the procedure can be continued until the desired repair success or target topology 315 has been achieved, possibly including a final verification of the repair result, for example using an AFM and / or SEM.

[0119] The components of the repair device can be implemented at least partially in hardware as well as in software, with combinations of hardware and software implementations clearly being possible.

[0120] A further advantageous implementation of the method can be such that, for example, after determining the initial topology of the defect, a target topology is specified as the desired end result of the repair process (manually or automatically), and the method then continues (largely) automatically on a system or device. For this purpose, the system or device can be designed to select a suitable process flow without manual intervention, for example, based on reference curves concerning a class of defects that are similar to or identical to the present defect, and in particular to decide how many iterations initially appear most advantageous for repairing the present defect. Based on this number, the intermediate topologies to be achieved can be determined automatically.The system can then automatically iterate through the repair process as described above. However, it is quite possible that the number of iterations initially targeted by the system will be insufficient (or too large) to achieve the desired repair result. In such cases, the system can automatically perform further repair steps (or omit steps that are no longer needed), or it can temporarily pause the repair process to await further user input. In any case, the system can calibrate itself during the iteration of the repair process, which can be a particular advantage of the described method and system. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] US 6593040 B2

[0010]

Claims

[1] Method (100, 200) for repairing a defect in a mask (10, 20) for lithography, in particular an EUV mask, comprising: a. Performing a first repair step (120) on the defect with a first repair dose (125), whereby the defect is thereby transformed from an initial topology (110) to a first defect topology (130); b. Determine (140) the influence of the first repair step on the topology of the defect; c. Determining a second defect topology (160) for the defect, which is to be achieved with a second repair step (150) on the defect; and d. Determining a second repair dose for the second repair step, at least partially based on the determined influence of the first repair step on the topology of the defect and the second defect topology. [2] Method according to claim 1, further comprising: e. Performing the second repair step with the second repair dose. [3] Method according to one of claims 1-2, wherein the first repair step is carried out as a calibration step, in particular as a calibration step in which less than 50%, preferably less than 30%, particularly preferably less than 15% of a desired final repair success is achieved, or wherein a first repair dose is determined for the first repair step which aims to achieve more than 50% of a desired final repair success. [4] Method according to one of claims 1-3, wherein step b of determining the influence of the first repair step on the topology of the defect comprises determining a change in the topology of the defect caused by carrying out the first repair step, in particular a comparison of the first defect topology with the initial topology. [5] Method according to claim 4, wherein step b. of determining the influence of the first repair step on the topology of the defect comprises determining a first process rate of the first repair step at least partially based on the determined change in the topology of the defect and the first repair dose. [6] Method according to any one of claims 1-5, wherein the topology of the defect comprises a height of the defect depending on a position on the mask. [7] Method according to claim 6, wherein the topology of the defect further comprises one or more of the following information: a lateral extent of the defect, a three-dimensional structure of the defect. [8] Method according to any one of claims 1-7, wherein step d. of determining the second repair dose for the second repair step further takes into account one or more calibration curves for different defect types that allow an estimation of a repair behavior of the defect. [9] Method according to claim 8, wherein the one or more calibration curves were created manually or automatically during the current run and / or one or more previous runs of the method. [10] Method according to any one of claims 5-9 in combination with claim 2, wherein the defect comprises several layers of material and wherein: The first repair step is carried out with the first repair dose on a first layer of material that contains a first layer of material. The second repair step is carried out with the second repair dose on a second material layer, which contains a second material that differs from the first material. the determination of the initial process speed of the first repair step with respect to the first material is carried out, and Determining the second repair dose also involves taking into account the thickness of the second material layer and the process selectivity of the second repair step on the second material in relation to the first repair step on the first material. [11] Method according to any one of claims 1-10, wherein the second defect topology represents a desired end topology. [12] Method according to any one of claims 2-10, wherein the second defect topology is an intermediate goal of the method, and wherein the method iterates through steps b. to e. until a desired final defect topology is achieved. [13] Method according to one of claims 11 or 12, further comprising a verification of the achieved final topology. [14] Method according to any one of claims 1-13, wherein the first repair dose and / or the second repair dose varies depending on the position. [15] Method according to any one of claims 1-14, wherein the first repair step and / or the second repair step comprises an etching process and / or a deposition process. [16] Method according to any one of claims 1-15, wherein the first process speed is determined as a function of position. [17] Method according to any one of claims 1-16, wherein the method takes place under constant external conditions, in particular in a vacuum. [18] Method according to any one of claims 1-17, wherein the method is carried out on a combined repair and analysis device. [19] Method according to any one of claims 1-18, wherein the defect corresponds to a known type of defect for which one or more calibration curves are known, and wherein the method further comprises drawing conclusions about the stability of the method and / or a device on which the method is carried out from the repair behavior of the defect. [20] Method according to one of claims 1-19, wherein reference process speeds are known for different materials, and wherein the method further comprises drawing conclusions about the material composition of the defect from its repair behavior. [21] Computer program with instructions which, when executed, cause a device for repairing a defect in a lithography mask to carry out the method according to one of claims 1-20. [22] Computer program according to claim 21, further comprising instructions which, when executed, cause the device for repairing the defect to automatically perform the first repair step as a calibration step on the defect with the first repair dose, to automatically determine the influence of this calibration step on the topology of the defect, and, at least partially based on the determined influence of the calibration step and the first repair dose, to automatically determine the second repair dose in such a way that the second repair step can be performed with it in order to achieve the second defect topology. [23] Device (300) for repairing a defect in a mask (10, 20) for lithography, in particular an EUV mask, wherein the device is configured to carry out the method (100, 200) according to one of claims 1-20. [24] Device according to claim 23, wherein the device comprises: a. Means (330) for performing a first repair step on the defect with a first repair dose, whereby the defect is thereby transformed from an initial topology to a first defect topology; b. Means (310, 350) for determining an influence of the first repair step on the topology of the defect; c. Means for determining a second defect topology for the defect, which is to be achieved with a second repair step on the defect; and d. Means (320) for determining a second repair dose for the second repair step at least partially based on the determined influence of the first repair step on the topology of the defect and the second defect topology. [25] Device according to claim 23 or 24, further comprising a memory in which the computer program according to claim 22 is stored, and means for executing the instructions contained therein, such that when the device is executed, it automatically performs the first repair step as a calibration step on the defect with the first repair dose, automatically determines the influence of this calibration step on the topology of the defect, and at least partially based on the determined influence of the calibration step and the first repair dose, automatically determines the second repair dose in such a way that the second repair step can be performed with it in order to achieve the second defect topology.

Images