Method, device and computer program for repairing mask defects

Abstract

A method, apparatus and computer program for repairing defects of lithographic masks, in particular EUV masks, are disclosed. A method for repairing defects of a lithographic mask, in particular an EUV mask, comprises the following steps: (a.) performing a first repairing step on the defect with a first repairing dose, such that the defect transitions from an initial topology to a first defect topology; (b.) determining the influence of the first repair step on the defect topology; (c.) determining a second defect topology at which a second repair step performed on the defect should reach; and (d.) determining a second repair dose of a second repair step, wherein at least one part of the determination is based on the measured influence of the first repair step on the defect topology and the second defect topology. In addition, the invention may also include a step (e.) of performing a second repair step with a second repair dose.

Description

[0001] This application is a divisional application of the invention application filed on March 28, 2022, with application number 202210311424.4 and entitled "Method, apparatus and computer program for repairing mask defects". Technical Field

[0002] This invention relates to a method, apparatus, and computer program for repairing defects in lithographic printing masks, particularly for repairing defects in EUV lithographic printing masks (hereinafter referred to as (EUV masks)). Background Technology

[0003] As the integrated density of microelectronics technology increases, lithographic printing masks (often simply referred to as "masks" in the following description) must image increasingly smaller structural components onto the photoresist layer of the wafer. To meet this requirement, the exposure wavelengths used are becoming increasingly shorter. Currently, argon fluoride (ArF) excimer lasers emitting wavelengths of 193 nm are primarily used for exposure. This process involves dense processing on a light source emitting extreme ultraviolet (EUV) wavelengths (10 nm to 15 nm) and a corresponding EUV mask. To improve the resolution of wafer exposure processes, the industry has simultaneously developed various traditional binary lithographic printing masks, such as phase masks or phase-shifting masks, as well as masks for multiple exposures.

[0004] However, as the size of structural components becomes smaller and smaller, it is not always possible to manufacture lithographic printing masks without visible or printable defects on the wafer. Since mask manufacturing is very costly, defective masks must be repaired as much as possible.

[0005] Defects in offset printing masks are divided into two main categories: dark defects and clear defects.

[0006] Dark defects refer to locations where absorbing or phase-shifting materials should not be present. Repairing dark defects is best done by removing excess material through localized etching.

[0007] Conversely, light leakage defects refer to locations on the mask where the transmittance is higher than that of a defect-free reference mask during exposure in a stepper or wafer scanner. These light leakage defects can be removed during mask repair by depositing or precipitating materials with appropriate optical properties. Ideally, the optical properties of the repair material should be the same as those of the absorbing or phase-shifting material.

[0008] The applicant has developed and proposed a measuring device for analyzing lithographic printing masks. In addition, the applicant has also proposed a repair device for lithographic printing masks.

[0009] One known method for removing dark defects is to directly target the defect to be repaired with a beam. The use of a beam allows for particularly precise guidance and positioning of the beam at the defect location. By combining this with a precursor gas (also known as a process gas) in the atmosphere guiding the beam to the mask to be repaired, the incident beam induces a reaction equivalent to localized etching. This induced localized etching process removes excess material from the mask (at the defect location), thus generating or restoring the desired absorption and / or phase shift characteristics of the lithographic mask.

[0010] Alternatively, a precursor gas could be selected that produces a precipitation effect upon exposure to the beam. This would allow additional material to be deposited at light leakage defects, thereby locally reducing the mask's transmittance and / or improving its phase-shifting properties.

[0011] For example, US 6,593,040 B2 describes a possible method for repairing phase-shift defects in a mask. This method involves scanning the mask for defects and identifying the location of at least one defect. The defect is then analyzed in three dimensions, and the results are used to align the defect with a focused ion beam (FIB) to remove it. The FIB is controlled by an etching map generated based on the results of the three-dimensional analysis. A test model can be generated using the FIB, subjected to three-dimensional analysis, and then the etching map can be generated using the test model.

[0012] However, a drawback of the aforementioned methods is that they largely fail to consider, or only minimally consider, the dynamics of the repair process. Thus, these methods perform only one 3D analysis before etching. However, during the repair process, the way defects respond to the ongoing repair measures may change. This could cause the repair steps calculated beforehand to fail to produce the desired results during method execution, thus failing to achieve the expected repair success (or even resulting in mask damage). Summary of the Invention

[0013] Therefore, the technical problem to be solved by the present invention is to provide a method for repairing mask defects, which can at least partially alleviate or eliminate the shortcomings of known methods, and / or repair mask defects more accurately and reliably. Furthermore, the repair equipment and computer program required to perform this method should also be provided.

[0014] The above-mentioned technical problems can be solved at least in part by the different aspects of the present invention described below.

[0015] In one embodiment, a method for repairing defects in a lithographic printing mask (particularly an EUV mask) includes the following steps: (a.) performing a first repair step on the defect using a first repair dose to transform the defect from an initial topology to a first defect topology; (b.) measuring the effect of the first repair step on the defect topology; (c.) determining a second defect topology that a second repair step should achieve on the defect; and (d.) measuring a second repair dose for the second repair step, at least in part, based on the measured effect of the first repair step on the defect topology and the second (to be achieved) defect topology.

[0016] In mask processing, particularly mask repair, one or more gases are typically directed to the desired reaction site and reacted with a beam of energy particles (such as photons, electrons, or ions) irradiating that site, thereby performing a desired processing procedure (such as the etching or deposition processes mentioned earlier and described in detail later). The repair dosage used in this process affects the repair steps.

[0017] According to the method, a first repair step is first performed on the defect to be repaired (e.g., the aforementioned light leakage point defect or dark defect) using a first repair dose. During this process, the defect will change, specifically, the defect will transform from the initial topology to a new topology, which is referred to here as the first defect topology.

[0018] In the simplest case, the topology can be described using only one or a few feature values, such as the height of the defect. However, as will be detailed later, within the scope of this invention, the defect topology can be described and characterized more precisely, such as the location-related defect height, the lateral dimensions of the defect, and the three-dimensional structure of the defect, to improve the accuracy of mask repair.

[0019] According to the method described, after performing the first repair step, the effect of the first repair step on the defect can be determined. Specifically, it can be determined what changes occur in the defect topology due to the first repair step, or what effect the first repair step has on the defect topology. Here, the determination of changes in defect topology or defects can be achieved in ways and methods well known to those skilled in the art, for example using a scanning probe microscope, particularly a scanning microscope, profilometer, or other suitable equipment (see description below). Interferometric methods can also be considered in principle, but interferometric methods typically do not have sufficient lateral resolution.

[0020] This step is based on the understanding that in mask defect repair, multiple factors typically influence the repair outcome given a specific repair dose, making it impossible to predict the exact characteristics of the defect. In other words, the repair dose required for a particular repair result is usually determined by multiple (defect) parameters, and not all of them necessarily need to be known. Besides defect topology, other parameters may include features near the defect on the mask (e.g., the reaction of a defect region near the mask's absorbing material may differ from that of a defect region near the mask's quartz material), the defect's location on the mask (e.g., the reaction of a region containing a lot of quartz material may differ from that containing a lot of absorbing material, and the reaction of a highly structured region may differ from that of a less structured region), the charging state and / or temperature of the mask surface. Particularly for defects that must be removed, the material composition of the defect is also a potential influencing parameter. This composition may be determined by the mask type (e.g., different materials and / or layers constituting the mask), or it may be affected by contamination or impurities in the mask or defect.

[0021] Furthermore, all these parameters may change during mask repair, and therefore the required repair dose may also change (unforeseen at the beginning) during the repair process. Therefore, during mask repair, the repair dose must be appropriately selected and adjusted to repair defects as intended and minimize mask damage. Thus, according to the method, instead of repairing defects in a single step with a pre-determined repair dose, the impact of the first repair step on the defect topology is measured after the first repair step is performed, and information is obtained from this measurement to control subsequent mask repair processes.

[0022] To be more precise, after determining the impact of the first repair process on the defect topology, the second repair step is then used to determine the second defect topology that should be achieved on the mask. As will be detailed later, the second defect topology may already be the desired final topology (i.e., the desired repair result), or it may only be an intermediate goal achieved by performing the repair method. Therefore, the repair method includes other steps that iterate with the steps described here until the desired final repair result is achieved. The second defect topology to be achieved can be determined automatically or manually. A hybrid of automatic and manual methods can also be used. Alternatively, the second defect topology or a series of defect topologies (see later explanation if it should be performed iteratively) can be given before the method begins. These defect topologies are derived, for example, from empirical values ​​obtained from repairing the same or similar defects and should be achieved after the method is executed sequentially.

[0023] After determining the second defect topology to be achieved in the second repair step, the required repair dose must then be calculated, taking into account the impact of the first repair step on the defect topology and the desired second defect topology. In this case, other factors that may affect the operation of the second repair step can, of course, be taken into account, such as the various factors mentioned above.

[0024] By referencing the information obtained from the first repair step, the required repair dose for the second repair step is determined. Therefore, according to the method, the mask repair process can be dynamically adjusted based on conditions that may change during the process, the characteristics of the defects being processed (in change), or other (uncontrolled and / or controllable) process parameters.

[0025] The method may further include the step of performing a second repair step using a second repair dose (e.).

[0026] The defect topology is changed again by performing the second repair step. Ideally, the topology achieved in this way is equivalent to the topology given with the second repair step as the target, where the second repair dose has already been determined with reference to the given topology. However, due to various factors, the defect topology after performing the second repair step may not be exactly equivalent to the desired topology. If the topology generated by the second repair step is not merely an intermediate target of the method, then other repair measures may be necessary to achieve the desired repair result. In any case, after performing the second repair step, the defect will have a topology, and therefore this topology will be referred to from this point onward as the second (true) defect topology.

[0027] As can be seen from the above embodiments, the first repair step can be performed as a correction step. For example, to avoid mask damage, such as when the first repair step is subjected to unforeseen (aggressive) operation that causes mask damage, the first repair step can be performed as a correction step. For example, the first repair step should only achieve less than 50% of the desired final repair result, preferably less than 30%, and particularly preferably less than 15%. In most cases, the desired final repair result is the complete removal of mask defects, but other cases can also be considered and included.

[0028] In particular, the repair dose for the first repair step can be selected (which may be location-dependent) to avoid exceeding the percentages mentioned above. In cases of uncertainty, a "conservative" value can be selected for the first repair dose of the first repair step to minimize mask damage (e.g., damage due to over-etching).

[0029] Alternatively, a first repair dose required to achieve a final repair result of over 50% can also be calculated for the first repair step. This reduces the number of subsequent iterations, minimizing the overall process time. It also allows for responses to potential deviations in one or more subsequent second steps. Here, the selection of the first repair dose can be determined, for example, based on experience gained from similar repair work.

[0030] As described above, step (b.) of determining the effect of the first repair step on the defective topology may include determining the changes to the defective topology caused by performing the first repair step. In particular, this step may include a comparison of the first defective topology with the initial topology.

[0031] If necessary, this step can be included before performing the first patching step to determine the initial topology.

[0032] In this invention, changes in defect topology can, in principle, be described and considered at different levels of detail. For example, as mentioned at the beginning of this document, the topology can be described using only one value (e.g., the height / thickness of the defect) (e.g., the height / thickness measured on the bottom surface of the defect, or the maximum height / thickness of the defect, or the minimum height / thickness of the defect). Correspondingly, in this simplest case, the change in topology can be described using only the change in the defect height / thickness. Comparing the first defect topology with the initial topology provides more detailed information about the topology changes. Generally, the more accurate the data obtained, the more accurately the second repair dose required to achieve the desired second repair defect topology can be calculated (sometimes the second repair dose is location-dependent, see later description). However, this may increase the analytical work required by the method, and therefore a trade-off must or can be made.

[0033] The step (b.) of determining the effect of the first repair step on the defect topology may include calculating the first process rate of the first repair step, which is at least partly based on the measured change in defect topology and the first repair dose.

[0034] Knowing the first repair dose, the rate at which the first repair step changes the defect topology can be deduced from the measured change in defect topology, such as the rate of material removal or deposition. For example, in an etching process, the extent of etching of the defect material by the first repair step using the first repair dose can be determined. In a deposition process, the rate at which material is deposited onto the mask, or the amount of material deposited onto the mask, by the first repair step using the first repair dose can be determined. This information can then be used to determine "how much more needs to be done," serving as the basis for calculating the second repair dose. Since the first process rate of the first repair step is measured from the defect being processed, other factors mentioned earlier that may affect the process rate can be automatically taken into account, thus enabling very precise and "dynamic" control of subsequent processes.

[0035] The defect topology can include the defect height, which varies with the position of the defect on the mask.

[0036] For example, such a "height map" of a defect allows for the location-analytical determination of the repair dose for one or more subsequent repair steps, because location-analytical information about the effect of the first repair step (or other iteratively performed repair steps) on the defect is available, and thus allows for location-related corrections to the process.

[0037] In addition, the defect topology may include one or more of the following information: the lateral dimensions of the defect, the three-dimensional structure of the defect.

[0038] As mentioned earlier, in general, incorporating more characteristics or data of the defect topology allows for more precise control over the process, but it also makes the process more complex and resource-intensive. Therefore, a trade-off must be made between the complexity and accuracy of the method.

[0039] The step (d.) of determining the second repair dose in the second repair step can take into account one or more calibration curves for different defect types, which can enable the prediction of the repair characteristics of the defect.

[0040] The "dynamic" correction of the method is based on the topological changes of the defect in the process (hereinafter often simply referred to as "the current defect"). In addition, the method can utilize reference data / calibration curves to further improve its accuracy and / or speed. For example, calibration curves can be used to adjust the repair dose for one or more subsequent repair steps to achieve a desired process speed. For instance, the topological changes experienced by the current defect during the first repair step can be compared with one or more calibration curves for the same or similar defect type to see if the current repair process is proceeding as expected or deviating from the reference value. If a deviation from the reference value is found, the correction data taken from the calibration curve can be used to determine the second repair dose for the second repair step. For example, the measured true effect of the first repair step on the defect topology can be averaged with the value taken from the calibration curve. If the response of the current repair process deviates from the expected response, this deviation should be taken into account when calculating the second repair dose (e.g., using this deviation as a correction parameter for the second repair dose calculated without considering the calibration curve). Alternatively, a weighted sum calculated from the measured true effect and one or more values ​​taken from one or more calibration curves can be taken into consideration. Furthermore, these stored reference data can be used to save intermediate steps, meaning that the desired final topology can be achieved more quickly, instead of having to "blindly" explore and find the desired repair results.

[0041] For this purpose, one or more calibration curves can be added manually or automatically to the current job of the method and / or one or more previous runs of the method.

[0042] For example, one or more calibration curves from one or more jobs added before the method can be stored in a database for access during the current operation of the invention. Other data storage methods can also be used. In particular, if the method is performed in an iterative manner as described later herein, calibration curves from steps prior to this iteration can be generated automatically during the current repair process, thus effectively reflecting the characteristics and features of the current repair process itself.

[0043] For defects comprising multiple material layers, the method can be performed as follows: a first repair step can be performed on a first material layer containing a first material with a first repair dosage, and a second repair step can be performed on a second material layer containing a second material with a second repair dosage, wherein the second material is different from the first material. Furthermore, a first process rate of the first repair step on the first material can be calculated, while the calculation of the second repair dosage can selectively take into account the thickness of the second material layer and the process of the second repair step on the second material relative to the first repair step on the first material.

[0044] Taking the thickness of the second material layer into account when calculating the second repair dose is optional; that is, in principle, the thickness of the second material layer may not be taken into account.

[0045] In any case, a first repair step on a first material layer containing a first material can be used to infer the reaction and effect of the repair process on the second material layer, and to calculate and select an appropriate second repair dose required to repair the second material layer in a second repair step. The ongoing reaction (e.g., an etching process or a precipitation process) has a significant impact on the process selectivity of the first and second materials, such as the process rate of the repair process on the first or second material, all other things being equal.

[0046] If the process selectivity for different materials is unknown, it can be determined from the repair process itself during the operation of the method and / or in one or more runs prior to the method. For example, before proceeding to the second repair step, a step of “testing” the second material layer (e.g., a processing step with the same repair material but a lower repair dosage) can be added before the second repair step to calculate the process selectivity for the second material.

[0047] As mentioned above, the second defect topology can be the desired final topology. That is, the present invention has the potential to achieve the desired repair result (or at least very close to the desired repair result) after the second repair step is completed. In particular, the second repair step can self-correct the defect at the present time, that is, it can be corrected by measuring the effect of the first repair step on the defect topology.

[0048] Another feasible approach is that the second defect topology is merely an intermediate goal of the method, and steps (b.) through (e.) are iterated until the desired final defect topology is reached.

[0049] In other words, after completing the second repair step with the second repair dose, the effect of the second repair step on the defect topology (and / or the average or weighted effect of the first and second repair steps) is measured. Then, a third defect topology to be achieved with the third repair dose is determined for the defect. At least part of this calculation involves determining the third repair dose based on the measured effect of the second repair step on the defect topology (and / or based on the aforementioned average or weighted effect), and measuring the third defect topology. The third repair step can then be performed using the third repair dose.

[0050] Generally, after completing the i-th repair step with the i-th repair dose, the impact of the i-th repair step on the defect topology can be measured. Then, the (i+1)-th defect topology that the (i+1)-th repair step should cause to the defect is determined, and at least part of the (i+1)-th defect topology is calculated based on the measured impact of the (i+1)-th repair step on the defect topology. Then, the (i+1)-th repair step is performed with the (i+1)-th repair dose. This process is repeated continuously.

[0051] When calculating the (i+1)th repair dose in the (i+1)th repair step, one can use the data and knowledge obtained from the previous (i)th executed iteration (i.e., the correction feed) as a starting point. Alternatively, one can use the correction feed obtained from one or more executed iterations (i.e., the (i-1)th iteration and / or the (i-2)th iteration and / or the (i-3)th iteration, etc.) as a starting point. In this way, the number of correction feeds taken into account can increase with the number of iterations. For example, one can use the average feed from different iterations or the feed calculated based on (reality).

[0052] Verifying the final topology achieved can also be part of the method.

[0053] For example, this verification could include checking whether the defect topology generated by the previous repair step is indeed consistent with the defect topology intended to be achieved in this repair step. If the verification passes, the repair process can be considered complete; if the verification fails, it indicates that iterations need to continue.

[0054] The first and / or second repair doses can vary depending on the location. The iterative operation of the method also applies to third and / or fourth repair doses, etc., that may be included by the method.

[0055] By using location-dependent repair dosages, the processing of defects can be controlled very precisely, with the aim of removing or depositing only the necessary amount of material at each location of the defect to achieve the desired defect topology after the repair process is completed.

[0056] The first and / or second repair steps may include an etching process and / or a deposition process. The iterative operation of the method also applies to third and / or fourth repair steps, etc., that may be included by the method.

[0057] As mentioned at the beginning of this article, mask defects can generally be divided into two types. One type is the presence of excess absorber material in locations where it shouldn't be, while the other type is the opposite, a lack of absorber material. Each type of defect has typical repair methods. One repair method is to remove the excess absorber material, particularly by etching it away. Another repair method is to deposit absorber material in locations where it is lacking.

[0058] In particular, the first process rate can be determined in relation to the position. The iterative operation of the method also applies to the second and / or third process rates that may be included in the method.

[0059] As mentioned earlier, this can be used to calculate the position-dependent repair dose for the next repair step, thus allowing for very precise control of the next repair step to achieve the predetermined defect topology as accurately as possible. Typically, the topological change from the recently completed repair step can be measured, at least in position-dependent terms, to calculate the position-dependent process rate; however, data on position changes from earlier iterations (if any) can also be used for calculation.

[0060] One advantageous approach is to keep the external conditions constant during the execution of the method, particularly to execute it in a vacuum.

[0061] If this is not the case, prior art methods, while allowing for one or more intermediate controls on the repair process, require removing the mask during processing to feed it into the analysis tool. However, during analysis, the mask surface (especially the defect surface) or defect composition may change. For example, when returning to the repair process, data obtained before or during analysis may become outdated due to oxidation, thus failing to accurately reflect the current situation. In contrast, the described method can be performed under constant conditions, particularly in a vacuum, thus avoiding these effects that could lead to errors and inaccuracies in the repair process.

[0062] In particular, the method can be performed on a combined repair and analysis device.

[0063] This is closely related to the previously mentioned possibility of performing the method under unchanged conditions, which is highly advantageous for the repair process because it eliminates the need to remove the mask from the repair equipment and feed it into a separate analysis device. Furthermore, this simplifies and speeds up the repair process.

[0064] For example, this combined repair and analysis equipment can be combined with the repair system equipped with a scanning microscope proposed in this case.

[0065] If the defect belongs to a known defect type, that is, one or more correction curves for that defect type are known, the method may further include inferring the stability of the method from the repair characteristics of the defect, and / or inferring the stability of the device performing the method.

[0066] The above inferences can be made by taking the correction curve into account.

[0067] As mentioned earlier, this paper explains how to use stored calibration data and calibration curves to verify whether the currently performed repair step is proceeding as expected (e.g., verifying based on the defect type of the defect in the process and the reference data stored for that defect type) to optimize and accelerate subsequent repair steps. This calibration using existing reference data / calibration curves can also be used to study the stability of the executed method against known or unknown influences. For example, if the current defect belongs to a well-studied and known defect type, and the repair characteristics of the current defect deviate significantly from one or more calibration curves, this indicates that the executed method and / or the equipment executing the method are highly susceptible to these influences. Sometimes, appropriate countermeasures may need to be taken (e.g., checking temperature, vacuum, mask charging, etc.) to avoid jeopardizing the current mask repair results.

[0068] If the reference process speeds for different materials are already known, the method may further include inferring the material composition of the defect from the repair characteristics of the defect.

[0069] The above inferences can be made by taking the reference repair speed into account.

[0070] For example, a fixed repair dose can be used to perform several or more repair steps on a defect, and the process rate of each repair step can then be inferred from the various topological changes of the defect. This can then be corrected using a known reference repair rate to determine the material composition of the defect (e.g., the location of the defect at each repair step).

[0071] Another aspect of the invention is a computer program containing instructions that causes a repair device to perform the method described herein for repairing defects in lithographic printing masks (particularly EUV masks).

[0072] Different aspects and features of the methods described herein can be combined with each other in this computer program, or individual aspects that are irrelevant to achieving the desired result can be excluded from the computer program.

[0073] The computer program may include instructions to control the device to perform a repair defect, such that the device automatically performs a first repair step on the defect as a correction step with a first repair dose, automatically determines the effect of this correction step on the defect topology, and automatically calculates, at least in part, based on the effect of the correction step and the first repair dose, the second repair dose required to perform a second repair step to achieve a second defect topology.

[0074] Since users do not need to operate manually, the mask repair process can be accelerated and / or the chance of errors reduced. This also facilitates the automation of the repair method.

[0075] Another aspect of the invention is an apparatus for repairing defects in lithographic printing masks (particularly EUV masks), wherein the apparatus is designed to perform the methods described herein.

[0076] The device can also be designed to perform repair work on it by combining different aspects and features of the methods described herein with each other, or by excluding individual aspects that are irrelevant to achieving the desired result from the computer program.

[0077] This device may include: a means for performing a first step on a defect using a first repair dose, wherein the defect is transformed from an initial topology to a first defect topology; a means for measuring the effect of the first repair step on the defect topology; a means for measuring a second defect topology that a second repair step performed on the defect should result in; and a means for measuring a second repair dose for the second repair step, the measurement being at least in part based on the measured effect of the first repair step on the defect topology and the second defect topology.

[0078] As mentioned earlier, this device can be a combined repair and analysis apparatus. Therefore, the methods being performed can be carried out entirely or at least almost entirely under constant external conditions, particularly in a vacuum.

[0079] For example, the aforementioned operations can be accomplished by a combination of the repair system and scanning microscope proposed in this case.

[0080] In addition, the device may also include a storage device for storing the aforementioned computer program and an apparatus for executing the instructions of the computer program. Thus, when executing the instructions, the device will automatically perform a first repair step as a correction step on the defect with a first repair dose, automatically calculate the effect of this correction step on the defect topology, and at least in part automatically calculate the second repair dose required to perform a second repair step to achieve a second defect topology based on the effect of the correction step and the first repair dose.

[0081] The advantages of this automatic execution have been pointed out above, and this also relates to the corresponding execution.

[0082] Finally, it should be noted that, to date, this invention has primarily described and discussed the repair of mask defects, which is also an important application of the principles proposed in this invention. However, it should be pointed out that the application of the teachings disclosed in this invention is not limited to this, but can also be applied to the (surface) treatment of other components in the field of microelectronics, such as for modifying and / or repairing the surface of structured wafers or microchips. Even if these applications are not explicitly stated herein, these possible applications still fall within the scope of the teachings disclosed in this invention, unless a certain application is explicitly excluded or is physically or technically infeasible. Attached Figure Description

[0083] The following will describe in detail possible embodiments of the present invention with reference to the accompanying drawings, in which:

[0084] Figure 1 : This shows an exemplary embodiment of the method described herein, in which excess absorbent material is removed;

[0085] Figure 2 : Shows the results of using an embodiment of the method described above to address defects in a mask having linear and spatial structures;

[0086] Figure 3 This section illustrates possible methodological flows for performing the methods described herein using a combination of analysis and repair equipment. Detailed Implementation

[0087] The embodiments of the present invention described below primarily relate to the repair of defects in offset printing masks. However, for the sake of completeness, it should be emphasized again that the present invention is not limited thereto, but can in principle be applied to other mask processing methods, or more generally, to the (surface) processing of other components in the field of microelectronics, such as for modifying and / or repairing the surface of structured wafers or microchips. Although the application described below primarily relates to the repair of mask defects for the purpose of clarity and ease of understanding, other possible applications will remain within the scope of the teachings disclosed in this invention for those skilled in the art.

[0088] It should also be noted that the following description only details certain embodiments of the present invention. However, those skilled in the art should understand that the features and variations described in these embodiments can be further modified and / or combined with each other in other combinations or sub-combinations, but still fall within the scope of the present invention. Individual features or sub-features may be excluded, provided that such individual features or sub-features are irrelevant to achieving the desired result. To avoid unnecessary repetition, the following description will not repeat the content already detailed above.

[0089] Figure 1An embodiment 100 of a method for repairing defects in mask 10 by etching away excess absorbent material is shown schematically.

[0090] When the method is started, the defect has an initial topology 110. The initial topology 110 may be known, for example, from previous analysis of mask 10, or measured, for example, by scanning microscopy during the execution of the method.

[0091] Next, in the first repair step 120, the defect is treated with a first repair dose (sometimes simply referred to as dose) 125. In this example, the first repair step 120 involves an etching process, such as directing one or more precursor gases around the mask defect and then irradiating the site, for example, irradiating energetic particles to excite a reaction that etches away the absorbing material. In this example, the repair dose 125 is location-dependent, meaning it is irradiated according to a so-called "dose map." This dose map can be determined to some extent based on the initial topology 110 of the defect.

[0092] After performing the first repair step 120, the topology described and represented by the height map resolved by the location of the defect will change, and the defect will have a first defect topology 130. As will be explained later, the first repair step 120 not only performs a substantial repair on the mask 10, but also acts as a correction step to control and optimize subsequent repair steps.

[0093] After performing the first repair step 120, the impact of the first repair step 120 on the defect topology is investigated in analysis step 140. In this case, the initial topology 110 is compared with the first defect topology 130 to determine changes in the defect topology, particularly changes in the location-related defect height, which result in a location-related change profile in the topology with respect to the first repair step 120. Considering the first repair dose 125 used, the process rate of the etching process can be calculated, which is used to control subsequent steps.

[0094] After performing analysis step 140, the second defect topology to be achieved in the next second repair step (indicated by arrow 150) is first calculated. For example, the second defect topology could be the final repair result, in which case excess absorbent material is removed all the way down to the quartz layer of mask 10. However, the second defect topology could also be an intermediate goal achieved by performing the repair method, for example, if the method is designed from the beginning to perform multiple iterations.

[0095] Based on the information obtained in analysis step 140, that is, based on the measured effect of the first repair step 120 on the defect topology, especially the measured process speed of the first repair step 120, and the desired second defect topology, the second repair dose of the second repair step 150 is determined, and the second repair step 150 is performed accordingly.

[0096] In the best-case scenario, the execution of the second repair step 150 precisely (or within an acceptable range) achieves the desired repair result, meaning that the resulting defect topology is exactly the desired second defect topology. However, as shown in Figure 1, the actual second defect topology 160 achieved by executing the second repair step 150 still deviates from the desired result, thus requiring post-processing and / or subsequent repair steps.

[0097] For example, it can be verified whether the second defect topology 160 reached after the second repair step 150 is as described, for example by comparing the second defect topology with an acceptable or similar interval.

[0098] If further processing is required, the steps described above can be repeated iteratively until the desired repair result is achieved.

[0099] Figure 2 Based on the teachings disclosed herein, a schematic diagram illustrating the results and process of mask repair method 200 is shown using a mask 20 having a linear and spatial structure. Excess absorbent material exists in the space that has been removed by the operation of method 200.

[0100] Method 200 is performed in a combined repair and analysis apparatus under constant external conditions, i.e., in a vacuum. This apparatus includes a scanning electron microscope (SEM) to perform the repair steps, wherein, with the introduction of a suitable etching gas (or gas mixture), the electron beam of the SEM is used to initiate an etching step at the defect location. Simultaneously, the SEM images the mask defect location. The SEM images are shown in [the diagram / image description]. Figure 2 The top of subgraphs 210, 220, and 240.

[0101] In addition, to more accurately analyze the defect topology, an analog scanning microscope (AFM) was used, and the corresponding images captured by the AFM were displayed on [the screen]. Figure 2 The middle and bottom images of sub-images 210, 220, and 240. The middle image shows a plan view of mask 20 (similar to an image taken by SEM) with a height profile drawn with contour lines (similar to a topographic map). Figure 2 The bottom of subfigures 210, 220, and 240 respectively show cross sections of the height profile taken along the horizontal line 28 via the corresponding AFM.

[0102] The initial defects are shown in sub-figure 210, as in a SEM image (above sub-figure 210) and an AFM image (in the middle of sub-figure 210, with contour lines). As previously described, a height profile (bottom of sub-figure 210) is cut out from the AFM image along horizontal line 28. From Figure 2 It can be seen that in a region 21 of the mask 20, there is a defect containing excess absorbing material located in the space between the two lines 25.

[0103] The first repair step is performed on this defective topology using a given dose map (repair dose) as the speed of the correction process.

[0104] In this correction step, the repair dose can be adjusted not only by changing the current intensity of the electron beam, but also by changing the number of times the electron beam irradiates a specific processing point in a repair step. Furthermore, the teachings disclosed in this invention explicitly include other possibilities for controlling and changing the repair dose.

[0105] After the correction / first repair step is completed, the following situation occurs: Figure 2 As shown in the middle sub-figure 220: Region 22 of mask 20 still has a (reduced) defect containing excess absorbing material in the space involved.

[0106] By re-taking one or more AFM images (see the middle and bottom of subfigure 220), the change in defect topology caused by the first repair dose can be determined positionally. Then, based on this change, the dose map used in the correction step, and the target topology to be achieved in the subsequent second repair step, a new dose map is calculated to achieve this target topology. This dose map is then used for defect repair in the next repair step 230, such as... Figure 2 As shown by the arrow in the image.

[0107] It should be noted that the processing of repair step 230 may also include multiple analysis and repair steps, in which the dose map of each step can be calculated based on the dose map of one or more previous steps, one or more previous topological changes, and the target topological dose map given by the individual or common repair target.

[0108] Method 200 can be designed to proceed in multiple iterations. This means that the defect topology to be achieved in a specific repair step does not necessarily correspond to the desired final repair result, such as removing excess absorbent material. Therefore, even after the first repair step has ended, a given repair step can still be used for further micro-correction and fine-tuning of repair method 200. Thus, the data obtained from the first repair step and this repair step (and possibly information from other corrective repair steps) can ultimately be applied to one or more subsequent repair steps to precisely repair defects in mask 20. For example, over-etching of defects can be avoided.

[0109] In other words, the desired defect topology for each step (e.g., the second and / or third and / or fourth defect topologies, etc., in the case of using an iterative method) does not necessarily have to be equivalent to the final result desired by the repair method 200 (though it can be equivalent to the desired final result), but can only be an intermediate target. In this case—and this applies universally to the teachings disclosed herein, not just to the embodiment 200 described herein—the intermediate defect topology to be achieved by the repair step does not need to be given in great detail, for example, not as detailed as the final result of the repair process. For example, it can be given that the achieved intermediate defect topology should not be greater than a certain percentage of the remaining defect size (e.g., defect height and / or defect width, or, in the case of a deposition process, only a given percentage) to avoid mask damage and / or (over-repair). For example, it can be given that the desired defect topology should not be greater than 50%, 30%, or 15% of the repair result not yet achieved. In other words, a repair step (or the next step after the first repair step) can be performed as (another) correction step, for example, to better understand and correct the kinetics of the repair process while avoiding mask damage and / or (over-repair).

[0110] In this scenario, the process rate of the etching process can be calculated from the measurement of the defect height change performed by AFM at each repair step, along with the dose map used. Then, in an interpolation step, using the remaining defect height also determined by AFM location resolution, a new dose map or repair dose (e.g., the number of cycles) can be determined to advance to the next step and / or the desired final defect topology.

[0111] Figure 2 Subfigure 240 on the right shows the defects (or remaining defects) after the second repair step (and possibly subsequent repair steps) 230. The combination of SEM and AFM images is used to verify the repair results.

[0112] In this example, from two AFM images ( Figure 2 As can be seen from the middle and bottom of sub-figure 240, residual defects still exist, which may be removed in the next repair step. However, if the topology of the residual defect is within the acceptable range for successful repair, the repair work can still be considered successful even with the presence of residual defects. In particular, as long as the residual defect is less than a specific threshold (for example, in wafer exposure, as long as mask 20 does not cause wafer damage), the residual defect can remain on mask 20.

[0113] Figure 3 This illustrates possible method flows 300 for performing the methods described herein, such as those that can be performed using components of a combined analysis and repair device, for example, a combined device comprising a repair system (e.g., a repair system comprising SEM) and an analysis system (e.g., an analysis system comprising AFM).

[0114] When repair 310 begins, the repair dose 320 required for a mask to be processed can be determined. The dose can be determined using information such as the target topology 315, the current topology 330, and the process rate 325. Specifically, in the first iteration step, the process rate 325 can be set by user input or by using a system-given standard value (e.g., a standard value derived from a reference value of a similar repair and correction curve). However, in the second iteration step or later iteration steps, the process rate can also be calculated or determined by measuring one (or more) previous topology changes 375 and one (or more) previous doses 365.

[0115] The dose can be given or defined by the number of times the electron beam emitted by the SEM during the repair step irradiates a processing point. However, other dose definitions can also be used.

[0116] The measured dose 320 can then be used in the first (partial) repair step 335 to repair the defect. One possibility is to use a SEM as a component of the repair equipment. Mask repair can be performed in one step or in multiple (e.g., iterative) steps. The first (partial) repair step 335 can also serve as a correction step, meaning that the desired first defect topology and corresponding first dose can be relatively (conservatively) selected and calculated to avoid mask damage or (over-repair). After the repair steps are completed, post-processing steps may sometimes be required.

[0117] In the next step 340, it can be decided whether the patching process should end 345 or whether the next iteration of the patching process should proceed 350.

[0118] If the next iteration 350 of the repair process should be performed, then the new topology 355 or the new defect height of the mask defect can be determined. Alternatively, although a next iteration 350 is not necessary, it is still necessary to determine the new topology, for example, to verify the final repair result. If the verification result unexpectedly does not match the desired result, another iteration 350 can be performed.

[0119] The new topology 355 can be stored as the (new) current topology 330. Based on the new topology 355 and data obtained from one or more previous iterations 360 (e.g., a previous topology 370), the topology change 375 caused by the first iteration step or the first repair step can be calculated. The data on the topology change 375 obtained in this way can be viewed alone, or combined with a previous dose 365 (the dose used in one or more previous iteration steps or repair steps) to calculate a (new) process rate 325.

[0120] Then, in step 320, the dose for the next (partial) repair step can be determined again. This can be determined based on the (new) process speed 325, the (new) current topology 330, and the target topology 315. Reference data stored in the system and / or dynamically generated calibration curves can be used at this location.

[0121] Then the next repair step 335 can be performed using the dose calculated in this way, and subsequent (partial) repair steps 335 can also be performed if necessary.

[0122] The method can be continued in this manner until the desired repair result or target topology 315 is achieved. Finally, the repair result can be verified, for example, by AFM and / or SEM.

[0123] At least some of the components of the repair equipment can be implemented using hardware and software, with a combination of hardware and software being a clearly possible implementation method.

[0124] Furthermore, an advantageous embodiment of the invention involves, after determining the initial topology of the defect, providing a target topology (manually or automatically) as the desired final result of the repair process, and then automating the repair process (to a great extent) on a system or device. The system or device can be designed to eliminate the need for manual intervention, such as finding a defect level similar to or the same as the current defect based on a reference curve, selecting an appropriate process execution method, and determining how many iterations are currently most beneficial for repairing the current defect. The intermediate topology to be achieved can be automatically calculated based on this target. As described above, the system can then automatically iterate during the repair process, where it is highly likely that the initially given number of iterations is insufficient (or excessive) to achieve the desired repair result. In such cases, the system can automatically perform more repair steps (or omit unnecessary steps), or the system can interrupt the repair process to await new user input. In either case, the system can self-correct during the iteration process, which is a significant advantage of the methods and systems described herein.

Claims

1. A method (100, 200) for repairing defects in lithographic printing masks (10, 20), particularly in EUV masks, the method comprising the following steps: a. performing a first repair step (120) on the defect with a first repair dose (125), wherein, The defect thus transforms from the initial topology (110) to the first defect topology (130); b. Determine (140) the effect of this first repair step on the defect topology; c. Determine the second defect topology (160) that should be achieved at the defect using the second repair step (150); d. Determine the second repair dose for the second repair step, at least in part, based on the measured effect of the first repair step on the defect topology and the second defect topology; The defect topology contains the feature values ​​of the defect related to its position on the mask.

2. The method according to claim 1, further comprising the following steps: e. Perform the second repair step using the second repair dose.

3. The method according to claim 1 or 2, wherein, This first repair step is performed as a correction step, specifically in which the desired final repair result is achieved only at 50% or less, preferably 30% or less, particularly preferably 15% or less, or somewhere in between. The first repair dose required to achieve a final repair result of more than 50% is determined for this first repair step.

4. The method according to any one of claims 1 to 3, wherein, The step (b.) of determining the effect of the first repair step on the defective topology includes determining the change caused to the defective topology by performing the first repair step, and in particular includes a comparison of the first defective topology with the initial topology.

5. The method according to claim 4, wherein, The step (b.) of determining the effect of the first repair step on the defect topology includes determining the first process rate of the first repair step based at least in part on the measured change in the defect topology and the first repair dose.

6. The method according to any one of claims 1 to 5, wherein, The defect topology includes the defect height relative to the defect's position on the mask.

7. The method according to claim 6, wherein, The defect topology also includes one or more of the following: the lateral dimensions of the defect, and the three-dimensional structure of the defect.

8. The method according to any one of claims 1 to 7, wherein, The step (d.) of determining the second repair dose in the second repair step also considers one or more calibration curves for different defect types, which can achieve a prediction of the repair characteristics of the defect.

9. The method according to claim 8, wherein, The one or more correction curves are manually or automatically added during the current run of the method and / or during one or more previous runs.

10. The method in combination with claim 2 according to any one of claims 5 to 9, wherein, The defect comprises multiple material layers, and among them: The first repair step is performed using the first repair dose in a first material layer containing the first material; The second repair step is performed using the second repair dose on a second material layer containing a second material, the second material being different from the first material; The first process speed of the first repair step with respect to the first material was determined, and The determination of the second repair dose also includes consideration of 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.

11. The method according to any one of claims 1 to 10, wherein, This second defective topology is the desired final topology.

12. The method according to any one of claims 2 to 10, wherein, The second defect topology is merely an intermediate goal of the method, and the method iteratively runs steps b through e until the desired final defect topology is reached.

13. The method according to claim 11 or 12, further comprising verifying the final topology achieved.

14. The method according to any one of claims 1 to 13, wherein, The first repair dose and / or the second repair dose are location-dependent.

15. The method according to any one of claims 1 to 14, wherein, The first repair step and / or the second repair step includes an etching process and / or a deposition process.

16. The method according to any one of claims 1 to 15, wherein, The speed of this first process is determined in relation to the position.

17. The method according to any one of claims 1 to 16, wherein, The method is performed under constant external conditions, particularly in a vacuum environment.

18. The method according to any one of claims 1 to 17, wherein, The method is operated on a combined repair and analysis device.

19. The method according to any one of claims 1 to 18, wherein, The defect is a known defect type, one or more correction curves for the defect type are known, and the method further includes inferring the stability of the method and / or the device performing the method from the repair characteristics of the defect.

20. The method according to any one of claims 1 to 19, wherein, The method includes inferring the material composition of a defect from its repair characteristics, based on known reference process rates for different materials.

21. A computer program comprising instructions that, upon execution, cause an apparatus for repairing defects in a lithographic printing mask to perform the method according to any one of claims 1 to 20.

22. The computer program of claim 21, further comprising instructions that, upon execution, cause the device for repairing a defect to automatically perform a first repair step as a correction step on the defect using the first repair dose, automatically determine the effect of the correction step on the defect topology, and determine a second repair dose at least in part based on the measured correction step and the effect caused by the first repair dose, thereby enabling the device to perform a second repair step using the second repair dose to achieve the second defect topology.

23. An apparatus (300) for repairing defects in lithographic printing masks (10, 20), particularly for repairing defects in EUV masks, wherein, The device is designed to perform the method (100, 200) according to any one of claims 1 to 20.

24. The device according to claim 23, wherein, The device includes: a. A device (330) for performing the first repair step on the defect using the first repair dose, wherein the defect is transformed from the initial topology to the first defect topology; b. Devices (310, 350) for determining the effect of the first repair step on the defective topology; c. A device for determining a second defect topology that should be achieved at the defect using the second repair step; d. A device (320) for determining the second repair dose of the second repair step, the determination being at least in part based on the measured effect of the first repair step on the defect topology and the second defect topology; The defect topology includes the feature values ​​of the defect related to its position on the mask.

25. The device of claim 23 or 24, further comprising a memory storing the computer program of claim 22, the device further comprising means for executing instructions of the computer program, wherein, when executing the instructions, the device automatically performs the first repair step as a correction step on the defect using the first repair dose, automatically determines the effect of the correction step on the defect topology, and automatically determines a second repair dose at least in part based on the measured correction step and the effect caused by the first repair dose, thereby enabling the second repair step to be performed using the second repair dose to achieve the second defect topology.

Images