DE102020208883B4 - Method and computer program for repairing a mask for lithography - Google Patents

Abstract

Method for repairing a mask for lithography, in particular a mask for EUV lithography, comprising the following steps:1a. determining a first value of an edge position (x) of a structure to be repaired on the mask,1b. Carrying out a repair process comprising at least a first repair step on the mask,1c. determining a second value of the edge position (x) after performing the first repair step,1d. comparing the second value of the edge position (x) with the first value of the edge position (x) and/or a target value (xo) for the edge position,1e. Determining a focus-exposure matrix for the edge position (x) with respect to an exposure system used for exposure,1f. Determining a deviation (dx) of the edge position (x) from the target value (xo) as a function of the focus depth (z) and the exposure dose (t) from the focus-exposure matrix, and1g. Providing the variable determined in step 1f as a quality metric relating to the repair process.

Description

1. Technical field

The present invention relates to a method, a device and a computer program for repairing a mask for lithography, in particular a mask for EUV lithography.

2. State of the art

As a result of the constantly increasing integration density in microelectronics, lithographic masks have to reproduce ever smaller structural elements in a photoresist layer (often also called "photoresist") on a wafer. In order to meet these requirements, the exposure wavelength is being shifted to smaller and smaller wavelengths. Argon fluoride excimer lasers, which emit light at a wavelength of 193 nm, i.e. in the so-called deep ultraviolet (DLTV) region of the spectrum, are currently mainly used for exposure purposes. Intensive work is being done on light sources that emit in the extreme ultraviolet (EUV) wavelength range (i.e. in the range 10 nm to 15 nm) and corresponding EUV masks.

Lithographic masks (often simply called "masks" or also "reticles") cannot always be produced without visible or printable defects on a wafer due to the ever smaller dimensions of the structural elements. Because masks are expensive to manufacture, defective masks are repaired whenever possible.

Two important groups of defects in lithographic masks are dark defects and clear defects.

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

Clear defects, on the other hand, are defects on the mask that, during optical exposure, e.g. in a wafer stepper or wafer scanner, have a higher light transmission than an identical defect-free reference position. In mask repair processes, such clear defects can be repaired by depositing a material with appropriate optical properties. Ideally, the optical properties of the material used for the repair should correspond to those of the absorber or phase-shifting material of the mask.

In addition to a defect-free mask, the exposure parameters during the exposure of the wafer with the mask also have an influence on the result of the exposure process and thus on the component that is ultimately produced, and there can be an interaction between the influence of the exposure parameters on the exposure process on the one hand and the repair measures on the other hand, which are necessary in order to achieve acceptable results in the later exposure.

For example, describes the U.S. 9,261,775 B2 the applicant proposes a method in which an aerial image stack of a mask or a mask feature is recorded, from which a so-called Bossung plot is then determined. The Bossung plot can be used to determine at least one parameter that characterizes the mask. A repair parameter can also be determined from the Bossung plot, which is forwarded to a repair system, which determines the process required for mask repair as a function of the repair parameter.

The disadvantage here, however, is that the only parameter used is the critical dimension (CD) of the mask features to be examined/repaired. With increasing integration density and modern exposure systems, in particular with EUV masks and corresponding exposure systems, this can be insufficient or problematic, since only relative changes in distance can be detected, which may not take sufficient account of the complexity of modern masks and exposure methods.

The DE 102 61 035 A1 describes a repairing device for repairing a defect of a photomask, comprising: a chamber accommodating the photomask; a gas inlet mechanism that introduces an etching gas into the chamber; an electron detection section which is supported by a surface detecting secondary electrons or reflection electrons emitted by the photomask by scanning an electron beam across the surface of the photomask; a pattern image forming section that forms a pattern image of the photomask based on information obtained by the detected electrons; and an irradiation mechanism that directs a shaped beam to a defective portion obtained from the pattern image thus formed, monitoring by electron beam scanning and etching by irradiation with the shaped beam being alternately repeated, and a non-defective edge position and a defective edge position from the Pattern image can be identified during monitoring, and etching is terminated when a repaired edge position reaches an appropriate position.

U.S. 2018/0 293 720 A1 , DE 10 2016 205 941 A1 and DE 10 2018 207 882 A1 describe devices for repairing a defect in a photomask.

The present invention is therefore based on the object of specifying a method which takes into account the ongoing development with regard to integration density and novel exposure methods and in particular provides improved possibilities for mask repair. Furthermore, a computer program with instructions for carrying out such a method is to be provided.

3. Summary of the Invention

The above objects are at least partially achieved by the various aspects of the claimed invention as defined in the independent claims. Further embodiments of the claimed invention are described in the dependent claims. Further "aspects", "embodiments", "examples", etc. described below that do not fall within the claimed invention as so defined are therefore to be considered as background information intended to promote an understanding of the claimed invention.

According to the claimed invention, a method for repairing a mask for lithography, in particular a mask for EUV lithography, comprises the following steps: (1a.) determining a first value of an edge position of a structure to be repaired on the mask, (1b.) Carrying out a repair process comprising at least a first repair step on the mask, (1c.) determining a second value of the edge position after carrying out the first repair step, and (1d.) comparing the second value of the edge position with the first value of the edge position and/or a target value for the edge position.

In the following, the value of the edge position is always denoted by the variable x and its target value is given as x _o . If necessary, the first and second determined value of the edge position are denoted by x ₁ and x ₂ , etc.

As mentioned at the outset, modern masks, in particular masks for ELTV lithography, are characterized by an ever-increasing structure density and, as a result, ever smaller structures. Furthermore, the use of light with ever-decreasing wavelengths results in further complications that, with conventional masks for longer wavelengths, do not have to be taken into account in this way, or at least not to the same extent.

On the one hand, masks for wavelengths at the lower end of the currently technically accessible range, in particular masks for EUV lithography, are not operated in transmission, but at the current state of affairs are reflective optical elements. This means that the chief ray angle (CRA) of the light beam exposing the mask will generally be non-zero (a typical value is, for example, 6°), so that the incoming light cone is separated from the outgoing light cone. A principal ray angle that is not equal to zero can cause distortion effects or shadowing effects at structure edges of the mask, to name just two possible effects that influence the imaging behavior of the mask.

Furthermore, the mask thickness is typically no longer small compared to the exposure wavelength used. On the contrary, an EUV mask, for example, typically consists of many layers stacked on top of each other with layer thicknesses that are greater than the wavelength of the actinic ELTV radiation. This means that the depth at which the radiation interacts with the mask can vary widely in units of exposure wavelength, resulting in so-called 3D effects. For example, the shadowing effect of a structure edge described above can result in the shadow area behind such an edge, the maximum interaction depth varies with the distance from the edge, which can affect the light reflected from the mask and thus its imaging properties.

The increasingly smaller dimensions of the structures also allow or require that the orientation of the individual mask structures will generally vary across the mask. For example, some structures may be horizontal and other structures may be vertical or diagonal, and so on.

In addition, the "environment" of a structure at one point on the mask can differ significantly from that of another structure at another point on the mask, and since diffraction effects and interference effects can influence the imaging properties of the mask, this also generally leads to different properties of the mask in different parts of the mask.

One can therefore say that in the case of masks for very short exposure wavelengths, in particular in the case of EUV masks, both the rotational symmetry and the translational symmetry of the mask will generally be broken by some or all of the effects described above.

For a successful repair of the mask, this means that it is no longer sufficient or in any case can be disadvantageous to only determine the relative distance between two mask features/structures and to take this into account as a characteristic value for the mask repair. This is where the claimed method comes in, in that it uses the (absolute) edge position x of a structure edge of the mask itself as a parameter. By using the edge position x itself, and not the relative distance between two edges (i.e. a critical dimension, CD), the sources of error and distortion, shadowing and 3D effects etc. described above can be better monitored and tracked, thereby making the Quality of mask repair can be improved.

Various options are available to the person skilled in the art for determining the edge position before or after the first repair step, all of which can in principle be used in the claimed method. One possibility is to take an aerial photograph or an aerial photograph stack, for example with the aid of a mask inspection system (eg a system from the applicant that is suitable for this purpose). Such a mask inspection system can be designed essentially as an optical exposure system of a lithography system, for example, which has a CCD camera instead of a wafer in order to measure the aerial image generated by the mask. Other options include scanning the mask with a scanning probe microscope/atomic force microscope (AFM) or a scanning particle microscope (SEM). The actual exposure of a “test wafer” can also be considered. Further details on these possibilities can be found, for example, in the publication DE 10 2011 079 382 A1 of the applicant, to which reference is accordingly made in connection with the analysis of a mask defect.

The (absolute) edge position x can be determined in relation to a suitably defined coordinate system. For example, when using a mask inspection system, as mentioned above, a coordinate system defined by the sensor of the CCD camera is appropriate. If another analysis method and/or another analysis tool is used, coordinate systems defined in an analogous manner will be apparent to the person skilled in the art. A coordinate system inherent in the mask itself would also be conceivable.

In order to monitor the progress of the repair process, the edge position x is determined according to the method described at least once before the first repair step is carried out and at least once afterwards, in order to compare the two values x ₁ and x ₂ with one another and/or a comparison with a ( absolute) target value x _o to allow the edge position. Of course, further measurements of the value of the edge position x are also possible and/or the repair process can include more than one repair step. For example, the repair process can run iteratively, after performing a certain number of repair steps (e.g. after each step, after every second step, etc., or after a varying number of steps), the value x of the edge position is determined anew and with one or more previous values and/or the target value x _o is adjusted.

The target value x _o can be specified, for example, by a previously determined mask design according to which the mask was produced and with which the mask is now compared. For example, the mask design can also define an acceptable deviation in edge position, within which an imperfect edge can still be considered acceptable.

The repair step(s) themselves can include, for example, an etching process and/or a deposition process (also called a deposition process), as is known in principle to the person skilled in the art for repairing masks, in particular for repairing ELTV masks. In this context, reference is made, for example, to the publications DE 103 38 019 A1 and DE 10 2013 203 995 A1 referenced, whose teachings regarding etching and deposition processes for mask processing and repair are hereby incorporated into the present disclosure.

By monitoring the edge position x (instead of e.g. a critical dimension CD), the repair process of the mask can thus be followed particularly precisely, also and especially with EUV masks that are subject to the shadowing, distortion and 3D effects etc. explained above and their Rotational and translational symmetry is broken in general.

The repair of the defect under consideration can be regarded as complete, for example, when the current value of the (absolute) edge position x is close enough to the (absolute) target value x _o , for example within a specified acceptance value Δ, that is to say when, for example, |x ₂ - x _o | < Δ (as already mentioned, this acceptance value Δ may have been specified together with the mask design, for example).

As an alternative or in addition to this, it can be assessed whether the value of the edge position x has changed by no more than a specified minimum change amount δ before and after a repair step has been carried out, ie whether, for example, | _x2 - _x1 | < δ in order to decide whether further repair steps are necessary or expedient. If one finds, for example, that the edge position has not changed by more than δ despite the implementation of a further repair step, this can be taken as a sign that a further repair step - at least one further repair step that affects the same repair mechanism (e.g. etching, material deposition). - will no longer be effective.

According to the claimed invention, the method further comprises the following steps: (1e.) determining a focus exposure matrix for the edge position with respect to an exposure system used for exposure, (1f.) determining a deviation of the edge position from the target value as a function of the depth of focus and the exposure dose from the focus-exposure matrix, and (1g.) providing the quantity determined in step 1f as a quality metric regarding the repair process.

An important measure of the quality of the repair of an edge of a mask structure can be the stability of the edge position with regard to (intended or unintended) variations in the exposure parameters with which an exposure system that is used to expose a wafer with the mask is operated. Important exposure parameters of such an exposure system used to expose a wafer are the focal depth (e.g. measured as the position of a focal plane relative to the wafer surface or to the surface of the photoresist on the wafer) and the exposure dose, which is essentially equivalent to the intensity of the radiation used. The exposure dose is also directly related to the so-called "threshold" or can be translated into it. The "threshold" refers to the light flux density on the wafer side at which the photoresist/wafer lacquer just reacts or just doesn't react.

In the following, the depth of focus is always denoted by the variable z, the exposure dose by the variable t. The deviation of the edge position x from the target value x _o is denoted by dx.

The focus-exposure matrix for the edge position x with regard to an exposure system used for exposure can thus be understood as a set of data or measured values that specifies the (absolute) edge position x for a set of measured points in the variables (z, t), where the measurement points will typically be discrete points, i.e. the depth of focus z and the exposure dose t are changed stepwise. For example, a mask inspection system can be used to determine such a data set, for example a system marketed by the applicant, with which aerial images are recorded with varying focal depth z and varying exposure dose t. If desired, a continuous and even differentiable data set can then be obtained from the data obtained in this way in a manner known to those skilled in the art, e.g. by suitable interpolation between the discrete data points. Accordingly, the determined deviation dx can be a data point, a set of discrete data points or a continuous and/or even differentiable data set.

In general, the aim will be to bring the edge position as close as possible to the target value x _o after the repair has been completed, ie to minimize the deviation dx.

In general, the deviation dx can have both a constant component dx̃ and a component that varies with the depth of focus z and the exposure dose t.

The constant component dx̃ can for example be defined or considered as the deviation of the edge position from the target value x _o at nominal focal depth z _o and nominal exposure dose t _o , ie dx̃ = x | (z _o , t _o ) - x _o ≡ x̃ - x _o .

As described immediately, the non-constant part of the deviation dx can be of particular interest, since it can contain further valuable information that can be used as additional quality metrics or “adjusting screws” for monitoring and controlling the repair process. However, it is also possible to monitor and control the repair using only the deviation dx as the only quality metric or "adjusting screw", in which context reference is again made to the advantages described above, which the use of the (absolute) Edge position x compared to using a critical dimension CD entails.

The method may further comprise the following steps: (2a.) Determining a dependence of the edge position on the depth of focus as a function of the depth of focus and the exposure dose from the focus-exposure matrix, (2b.) Determining a dependence of the edge position on the exposure dose as a function the depth of focus and the exposure dose from the focus-exposure matrix, and (2c.) providing the quantities determined in steps 2a and 2b as further quality metrics with regard to the repair process.

This embodiment is based on the finding that not only the offset dx of the edge position from a desired position x _o itself may be used to monitor and control the repair process, but that further relevant information about the quality and shape of the repaired or edge to be repaired may be included in the dependencies, and in particular the rates of change, with which the edge position x will respond to variations in exposure magnitudes, in particular variations in the values of z and t respectively.

It has been found that the dependency of the edge position on the depth of focus z or the exposure dose t can quantify and describe different and essentially independent properties of the edge. This in turn turns out to be advantageous because, as described in more detail below, it can allow one or the other aspect of the edge shape to be influenced in a targeted manner by varying (largely) independent process parameters for the repair process. In addition, this can make it possible to identify those errors or deviations in the shape of the edges that play a particularly critical role in the subsequent exposure process, or vice versa, to determine which aspects of the requirement profiles/acceptance limits can be relaxed somewhat and still be acceptable Exposure result can be achieved.

Figuratively speaking, the inclusion of the dependencies or the rates of change of the edge position with regard to the depth of focus z and the exposure dose t increases the number of "adjusting screws" that are available for monitoring and controlling the progress of the repair process, so that the quality of the result of the repair process increases can be.

Although the rates of change (e.g. in terms of the first partial derivatives with respect to z or t, as described below) of the deviation of the edge position x from the target value x _o with variation of the depth of focus z or the exposure dose t are an important set of information or quality metrics regarding of the monitoring and control of the repair process, it should already be mentioned at this point that the dependencies of the edge position on the depth of focus z or the exposure dose t can also be more complicated, and can also include higher partial derivatives, for example, as also further described below. By considering such additional variables, the number of quality metrics and consequently the number of "adjusting screws" can be further increased with which the repair process can be monitored and controlled.

The method may further comprise the steps of: (3a.) adjusting a set of process parameters for performing the first and/or a second repair step of the repair process to improve at least one of the quality metrics, and (3b.) performing the first and/or second repair step using the adjusted process parameters.

In general, one will try to improve all quality metrics, with a trade-off between the time and effort of the repair process on the one hand and that on the other hand further improvement achieved may be appropriate on the other hand. However, it is also possible to focus on a single quality metric or a subset of the existing quality metrics when adjusting the process parameters, for example if experimental and/or theoretical findings show that these have a particularly noticeable or will have a critical impact.

Since the repair process, as already described above, can either only include the first repair step or several repair steps, in particular an iterative sequence of repair steps, this means that the quality metrics can be viewed at every point in the repair process and the process parameters can be adjusted accordingly. However, it would also be conceivable not to look at the quality metrics and adjust the process parameters after each step, but only after every second step, or after every third step, or whenever this may seem necessary. If the quality metric or quality metrics under consideration have (noticeably) improved in a process step, the subsequent process step can also be run with the same process parameters, and only when "nothing is happening anymore" in the quality metrics (e.g. measured against a suitably selected threshold value), the process parameters can be adjusted further.

The dependency of the edge position x on the focal depth z can include the first partial derivation of the edge position x according to the focal depth z. The dependence of the edge position x on the exposure dose t can include the first partial derivative of the edge position x with respect to the exposure dose t.

$$\frac{\partial }{\partial a}\left(\frac{\partial x}{\partial a}\right)=\frac{{\partial }^{2}x}{\partial a^2}\equiv x_{aa},\frac{\partial }{\partial a}\left(\frac{\partial x}{\partial a}\right)=\frac{{\partial }^{2}x}{\partial b\partial a}\equiv x_{ab},usw\text{.}$$

In the following, the partial derivatives of the edge position x according to one or more variables a, b, ... are optionally represented as follows: $$\frac{\partial x}{\partial a}\equiv x_a,\frac{\partial x}{\partial b}\equiv x_b,\text{etc}.$$

$$\frac{\partial }{\partial a}\left(\frac{\partial x}{\partial a}\right)=\frac{{\partial }^{2}x}{\partial a^2}\equiv x_{aa},\frac{\partial }{\partial a}\left(\frac{\partial x}{\partial a}\right)=\frac{{\partial }^{2}x}{\partial b\partial a}\equiv x_{ab},\mathrm{and}sw\text{.}$$

wobei auch nur eine der beiden ersten partiellen Ableitungen x_z bzw. x_t in dem Verfahren verwendet werden kann, falls gewünscht bzw. zur Erzielung der gewünschten Resultate erforderlich.The two options mentioned here, possibly together with the constant component dx̃ already described above, which can also be zero, represent a development of the edge position up to the first order in the exposure parameters: $$\text{i.e}x=\text{i.e}\tilde{x}+\frac{\partial x}{\partial \mathrm{e.g}}\text{i.e}\mathrm{e.g}+\frac{\partial x}{\partial t}\text{i.e}t\equiv \text{i.e}\tilde{x}+x_{\mathrm{e.g}}\text{i.e}\mathrm{e.g}+x_t\text{i.e}t$$

where only one of the two first partial derivatives x _z or x _t can be used in the method if desired or necessary to achieve the desired results.

The deviations dz and dt are measured from a specific reference point (z _o , t _o ), ie dz = z - z _o and dt = t - t _o . This reference point can, for example, be the intended/nominal working point of the exposure system, with which a wafer is later to be exposed using the mask.

It should be emphasized at this point that the position of the reference point (z _o , t _o ) can itself be regarded as a variable within the framework of the method described here, i.e. the operations and analysis steps described here can be started at a first working point (z _o ( ^1st ), t _o ⁽¹⁾ ) and then at another working point (z _o ⁽²⁾ , t _o ⁽²⁾ ) to determine, for example (e.g. as part of a simulation carried out before the actual repair of the same), with regard to which working point the repair would be easier or better to carry out or at which operating point the repaired mask would be more stable with regard to variations in the parameters z and t.

The first partial derivatives represent, at least in the vicinity of a certain nominal operating point (z _o , t _o ) of the exposure system, a measure of the rate of change of the edge position x with variation of the depth of focus z or exposure dose t, and thus give the "sensitivity “ of the edge position x compared to (intended or undesired) fluctuations or deviations of these parameters from the respective nominal/target value z _o or t _o .

These rates of change can therefore be of important importance for the mask repair because, as described immediately, they correspond directly to the relevant physical properties of the edge or edge structure, which can thus be specifically monitored and "controlled" in the repair process.

The dependency of the edge position x on the focus depth z, in particular its rate of change in terms of the first partial derivative x _z , can represent a quality metric in the form of a measure of a phase error and/or a telecentricity error.

In a repair process that includes an etching and/or deposition process, the quality metric with regard to the phase error/telecentricity error can then be influenced in a targeted manner by adjusting an end point of the etching and/or deposition process (e.g. the depth at which the process is terminated). For example, the quality metric can be optimized (as far as technically possible) by controlling the etching/deposition process in such a way that the dark/clear defect is eliminated as much as possible, but without unintentionally changing or affecting the surrounding mask substrate (e.g. due to unwanted etching away of the substrate near the edge to be repaired or unwanted material deposition on the substrate near the edge).

The dependency of the edge position x on the exposure dose t, in particular its rate of change in terms of the first partial derivative x _t , can represent a quality metric in the form of a measure of an edge steepness.

By adjusting a drift and/or jitter movement during the repair process, the quality metric regarding the edge steepness can be influenced in a targeted manner. In general, minimizing the drift and/or jitter motion increases the edge steepness, i.e. the corresponding quality metric is increased. However, it is by no means impossible for the method to be operated in such a way that, for example, an edge is made flatter by deliberately induced jitter movements, if this is desired or necessary to achieve a specific repair success.

The method can also include the following steps: (10a.) Determining at least one partial derivative of the second or higher order of the edge position according to the depth of focus and/or the exposure dose from the focus-exposure matrix, and (10b.) Providing the at least one partial Second or higher order derivation as another quality metric related to the repair process.

As already mentioned, the dependencies of the edge position x on the depth of focus z and the exposure dose t can also be considered beyond the first partial derivations, and from this further relevant quality metrics or "levers" for the repair process can be derived.

$$\frac{\partial }{\partial z}\left(\frac{\partial x}{\partial t}\right)=\frac{{\partial }^{2}x}{\partial z\partial t}\equiv {\partial }^{2}x/\left(\partial z\partial t\right)\equiv x_{tz}$$

$$\frac{\partial }{\partial z}\left(\frac{\partial }{\partial z}\left(\frac{\partial x}{\partial t}\right)\right)=\frac{{\partial }^{3}x}{\partial z^2\partial t}\equiv {\partial }^{3}x/\left(\partial z\partial z\partial t\right)\equiv x_{tzz}$$

In particular, the following partial derivatives of the second or higher order can be relevant for monitoring or controlling the mask repair process: $$\frac{\partial }{\partial \mathrm{e.g}}\left(\frac{\partial x}{\partial \mathrm{e.g}}\right)=\frac{{\partial }^{2}x}{\partial {\mathrm{e.g}}^2}\equiv {\partial }^{2}x/\left(\partial \mathrm{e.g}\partial \mathrm{e.g}\right)\equiv x_{\mathrm{e.g}\mathrm{e.g}}$$

$$\frac{\partial }{\partial \mathrm{e.g}}\left(\frac{\partial x}{\partial t}\right)=\frac{{\partial }^{2}x}{\partial \mathrm{e.g}\partial t}\equiv {\partial }^{2}x/\left(\partial \mathrm{e.g}\partial t\right)\equiv x_{t\mathrm{e.g}}$$

$$\frac{\partial }{\partial \mathrm{e.g}}\left(\frac{\partial }{\partial \mathrm{e.g}}\left(\frac{\partial x}{\partial t}\right)\right)=\frac{{\partial }^{3}x}{\partial {\mathrm{e.g}}^2\partial t}\equiv {\partial }^{3}x/\left(\partial \mathrm{e.g}\partial \mathrm{e.g}\partial t\right)\equiv x_{t\mathrm{e.g}\mathrm{e.g}}$$

It should be noted at this point that, at least after suitable processing of the data of the focus-exposure matrix, it can generally be assumed that all of these derivatives are continuous and even differentiable, so that, based on Schwarz’s theorem, the sequences of the partial derivatives then does not matter. Other partial derivatives of the second or third order, which would be possible in principle, have turned out to be less relevant in the context of the claimed method and are therefore not discussed further below. In principle, it would of course be possible to include them.

In particular, it is possible for the quality metrics to include a quality metric in the form of a depth of field measure, where this quality metric is proportional to x _tzz .

wobei im letzten Schritt die Umformung $$\frac{1}{1-\text{y}}\approx 1+\text{y}+\dots \text{f}\ddot{\text{u}}\text{r}\left|\text{y}\right|<1$$

benutzt wurde und die drei Parameter „tce“, „sm“ und „zw“ eingeführt wurden (solche Parameter werden zu Deutlichmachung und Unterscheidung von Variablen im Folgenden fett und in aufrechter Schriftart angegeben). Die im Folgenden als Parameter „sme“ bezeichnete Kombination $$\frac{1}{\mathbf{sm}\ast {\mathbf{zw}}^2}\equiv \mathbf{sme}$$

tritt dabei als Faktor des (dt dz dz) - Terms auf und stellt eine Qualitätsmetrik in Form eines Maßes für die Kantenqualität dar. Diese wiederum hat Einfluss darauf, wie groß der Bereich der Tiefenschärfe ist, für den die Kante ein akzeptables Abbild auf dem Wafer liefert, und sie enthält Informationen, die über diejenigen hinausgehen, die in dem x_t - Term enthalten sind. Ihre Einbeziehung in das beanspruchte Verfahren erhöht somit abermals die Anzahl der Parameter oder „Stellschrauben“, die zur Überwachung und Steuerung des Reparaturprozesses zur Verfügung stehen.Based on Eq. (3) Above, an advantageous model for describing and controlling the mask repair process in the context of the claimed method can be given as follows: $$\text{i.e}x=\text{i.e}\tilde{x}+\mathbf{tce}\ast \text{i.e}\mathrm{e.g}{\left(\mathbf{sm}\ast {(1-\left(\frac{\text{i.e}\mathrm{e.g}}{\text{between}}\right)}^2\right))}^{-1}\ast \text{i.e}t\approx \text{i.e}\tilde{x}+\mathbf{tce}\ast \text{i.e}\mathrm{e.g}+\frac{\text{i.e}t}{\text{sm}}\ast \left(1+{\left(\frac{\text{i.e}\mathrm{e.g}}{\text{between}}\right)}^2\right)$$

where in the last step the transformation $$\frac{1}{1-\text{y}}\approx 1+\text{y}+...\text{f}\ddot{\text{and}}\text{right}\left|\text{y}\right|<1$$

was used and the three parameters "tce", "sm" and "zw" were introduced (such parameters are given in bold upright font for clarity and differentiation of variables in the following). The combination referred to below as the "sme" parameter $$\frac{1}{\mathbf{sm}\ast {\mathbf{between}}^2}\equiv \mathbf{sme}$$

occurs as a factor of the (dt dz dz) term and represents a quality metric in the form of a measure of edge quality. This in turn affects how large the range of depth of field is for which the edge is an acceptable image on the wafer provides, and it contains information beyond that contained in the x _t term. Their inclusion in the claimed method thus further increases the number of parameters or "adjustments" available to monitor and control the repair process.

The claimed method may further comprise evaluating at least one quality metric at a nominal depth of focus z _o and a nominal exposure dose t _o to obtain at least one quality number related to the repair process.

In other words, in order to obtain numerical variables that can be compared, for example, with one or more threshold values or acceptance intervals, one or more of the quality metrics (the ones that are currently to be considered) can be set at a desired operating point of the exposure device for later exposure Wafers are evaluated with the mask, each yielding a particular numerical value or set of numerical values, referred to herein as a quality number(s). As already indicated above, several working points, for example (z _o ⁽¹⁾ , t _o ⁽¹⁾ ) and (z _o ⁽²⁾ , t _o ⁽²⁾ ), can be considered (e.g. as part of a prior repair performed simulation or similar) to determine whether and if so which operating point would be more suitable.

In line with what has been said up to this point, the goal of the repair process will generally be to process the mask so that this quality number or numbers are as close as possible to, or at least within a certain acceptance interval, a respective target value. The acceptance interval can be quite different for different quality numbers. For example, a quality number that corresponds to a “critical” quality metric (e.g., in the sense that a deviation/degradation from that metric would mean a significant error in imaging a wafer) may have a more stringent measure than a more critical one "uncritical" quality number.

In this context, it is once again emphasized that it is precisely the inclusion of several quality metrics within the framework of the method described that may allow individual requirement profiles and acceptance intervals to be relaxed, since this allows different error contributions (as far as possible) to be monitored separately from one another and also influenced separately . If, for example, one considers only one critical dimension CD as the only parameter, one will always only be able to see and track the "interaction" of all error contributions.

Nevertheless, the present method can also include determining a critical dimension (CD) of the structure to be repaired and including the critical dimension in a quantification of the effect of the repair process.

Furthermore, it is emphasized that, generally speaking, the edge position x can also always be replaced by a critical dimension CD in everything that has been said so far, and that this is also covered by the present disclosure. However, this will generally be associated with the loss of some of the advantages mentioned at the outset, which result from using the (absolute) edge position x as a parameter.

At this point it should also be mentioned that the possible features, options and possible modifications of the claimed method have been described in a specific order up to this point, but this is not necessarily intended to express a specific interdependence of the features - unless this is explicitly stated or it results from technical, physical and/or mathematical necessities. Rather, the various features and options can also be combined with one another in other orders and permutations - as far as physically and technically possible - and such combinations of features or even sub-features are also covered by the claimed method. Individual features or sub-features can also be omitted if they are not required to achieve the desired technical result.

Another aspect of the claimed invention is a computer program that includes instructions that when executed cause a mask repair apparatus to perform the steps of an embodiment of the claimed method.

character list

• 1 einen 3D-Effekt/Abschattungseffekt zeigt, welcher bei der Verwendung von EUV-Strahlung an einer Maske auftreten und deren Abbildungseigenschaften beeinflussen kann;
• 2a-c Probleme zeigen, die bei der Verwendung von ELTV-Strahlung hinsichtlich der Auswahl eines globalen Isofokalpunktes auftreten können;
• 3 verschiedene mögliche Kantenfehler sowie deren Überlagerung zeigt;
• 4 die Grundlage eines Modells 3. Ordnung zur Definition geeigneter Qualitätsmetriken zur Überwachung eines Reparaturprozesses einer (EUV)Maske zeigt;
• 5a-d einen Vergleich der Vorhersagen des Modells 3. Ordnung mit einer numerischen Simulation der Abbildungseigenschaften einer Maske zeigen; und
• 6 weitere Anwendungsmöglichkeiten der besprochenen Qualitätsmetriken bei der Reparatur und Analyse von Maskendefekten illustriert.

In the following detailed description, possible embodiments of the invention are described with reference to the figures, wherein the

• 1 shows a 3D effect/shading effect which can occur on a mask when using EUV radiation and can influence its imaging properties;
• 2a-c show problems that can arise when using ELTV radiation regarding the selection of a global isofocal point;
• 3 shows various possible edge defects and their superimposition;
• 4 shows the basis of a 3rd order model for defining suitable quality metrics for monitoring a repair process of an (EUV) mask;
• 5a-d show a comparison of the predictions of the 3rd order model with a numerical simulation of the imaging properties of a mask; and
• 6 further possible applications of the quality metrics discussed in the repair and analysis of mask defects are illustrated.

5. Detailed description of possible embodiments

In the following, embodiments of the present invention are described primarily with reference to the repair of a mask for ELTV lithography. However, the invention is not limited to this and can also be used for other types of masks and mask processing.

It is also pointed out that only individual embodiments of the invention can be described in more detail below. However, a person skilled in the art will understand that the features and possible modifications described in connection with these embodiments can also be further modified and/or combined with one another in other combinations or sub-combinations, without departing from the scope of the present invention. Individual features or sub-features can also be omitted if they are not required to achieve the intended result. In order to avoid unnecessary repetition, reference is therefore made to the statements and explanations in the previous sections, which are also valid for the detailed description that now follows.

The 1 10 shows a section 100 of an EUV mask 110 exposed with ELTV radiation 120, which includes an absorber edge 112 and a substrate 115. FIG. The substrate 115 has a multilayer structure, as is known in principle to those skilled in the art of constructing ELTV masks, and the ELTV radiation 120 incident on the EUV mask 110 interacts with the substrate 115 in principle at the points where it is not Absorber material 112 is covered.

Since EUV masks like the mask 110 shown here are not operated in transmission, but are designed as a reflecting optical component, the chief ray angle θ (“chief ray angle”, CRA), cf. reference sign 122, is generally not equal to 0° , as well as in 1 shown. This can lead to the following 3D/shading effect: If the absorber edge 112 as in 1 shown is sharply formed, there is a shadow area directly behind the edge 112 in the "direction of inclination" of the incident light 120, the width of which depends on the edge plane 140 of the interaction depth of the radiation 120 with the layers of the substrate 115. Thus, an interaction point 135 with the top substrate layer viewed from the direction of incidence can be closer to the edge plane 140 than an interaction point 136 in the middle of the substrate 115. This in turn can be closer to the edge plane 140 than an interaction point 137 with the deepest layer of the layer viewed from the direction of incidence Substrate 115. This influenced the respective reflected beams 125, 126 and 127 of the EUV radiation from these interaction points 135, 136 and 137 or their superimposition with reflected radiation from other areas and layers of the mask 110, and thus has an influence on the imaging properties of the Mask 110, in particular since the thickness of the substrate 115 iA will no longer be small compared to the wavelength of the EUV light 120, so that interference effects between the various reflected beams, for example the beams 125, 126, 127, gain weight.

The 2a-c show further effects regarding the imaging behavior of modern masks, which can result from the ever increasing integration density and the ever decreasing exposure wavelength.

The 2a shows a section 200 of an intensity profile (coded by different colors) as a function of the mask position, which results from the exposure of a largely defect-free ELTV mask with a very simple structure, namely a mask with a sequence of vertically running structures. Masks with which modern chips or the like are to be produced are often of a considerably more complex structure (they can contain structures with different orientations, for example, as mentioned at the beginning). The in the 2 B and 2c Details shown relate to the area 210 of the intensity profile of 2a .

The 2 B FIG. 12 shows a section 201 along the horizontal axis through the center of area 210 of FIG 2a , where in 2 B the intensity (in relative units) is plotted on the ordinate. The zero point of the abscissa, which indicates the horizontal position on the mask, is in 2 B chosen arbitrarily in the middle of the figure. The intensity curve has also been truncated at high and low values, and only the area around the critical dose 225 is shown (relative to a wafer intended for exposure), since this is exactly the intensity or threshold area that is used in the Structure formation in the photoresist of the wafer will be relevant or can occur in practice, for example due to fluctuations in the light intensity or (slight) variations in the triggering threshold of the photoresist/varnish for the polymerization. In addition, a whole family of intensity curves is shown for different values of the depth of focus z (the different curves are shown in 2c better recognizable), and three intersection areas 220 of this set of curves with the critical dose 225 are highlighted, which are numbered “1”, “2” and “3”.

In 2c enlarged sections 202 of the cut areas just mentioned are shown. The section "1" is shown at the top at reference number 230, the section "2" in the middle at reference number 231, and the section "3" at the bottom at reference number 232. The ordinate again shows the intensity (in relative units) and on the abscissa represents the horizontal position on the mask, the zero point of the abscissa again being chosen arbitrarily at the (average) point of intersection of the family of curves with the critical dose. The curves denoted by reference number 235 represent those curves which have the greatest defocusing (ie deviation from the nominal value z _o ) of all the curves in the family of curves.

As can be seen in particular from the enlarged sections 230 and 232 (i.e. the intersection areas "1" and "3"), even for the largely defect-free and simply structured mask considered here, there is generally no common intersection point in the ELTV wavelength range of all curves of the family of curves with the critical dose more. In other words, even for the "good" and "simple" mask considered here, it is generally no longer possible to find a global isofocal point.

For example, due to the in connection with the 1 and the 2a-c In the circumstances described, it is therefore of great relevance for the repair of modern masks, in particular EUV masks, not only to know the relative distance between the various mask structures and, if necessary, to to correct, but it is desirable to have additional "adjusting screws" and quality metrics available in order to take into account and incorporate the complicated imaging properties of the mask depending on and in the event of deviation from a selected nominal point (t _o , z _o ) in the repair process (where the nominal values can be chosen globally for a mask, as just described).

This is where the method according to the invention comes into play, of which in the following two 3 and 4 certain aspects should be emphasized.

The 3 shows first schematically a section of a defect-free mask 300, with a (possibly multi-layer) substrate 305 and an absorber layer 302, which forms a clearly defined edge in the intended position 308 (for example the position specified in the mask design).

Three sections of masks 310, 320 and 330 are shown underneath, each of which has a specific error, as can be monitored and eliminated using the options described herein during a corresponding repair process.

For the mask 310, the edge position 318 is shifted compared to the desired position 308. FIG. Since the claimed method allows the absolute edge position x to be monitored, such an error can be corrected, for example, by etching away the excess material. The progress of the repair can be controlled and monitored continuously and/or at specific time intervals by (re)determining the edge position x. Furthermore, this is also possible under "difficult conditions", e.g. if the mask no longer has rotational and/or translational symmetry, since instead of the critical dimension CD otherwise used, i.e. a relative position measure, the absolute edge position x can be used as a characteristic value.

Although the mask 320 has the correct edge position, the mask 320 has a void 328 in the substrate 305 near the edge. This defect 328 could, for example, be due to the fact that the excess material of the overhanging edge 318 was etched away in the mask 310, but the etching process was terminated too late, so that it continued into the substrate 305 (analogously, a certain part of the overhanging edge 318 may have remained on the substrate 305 if the etch process was terminated too early, leading to a situation similar to that described here). In order to avoid such an error, it is advantageous to take into account information that goes beyond the edge position x when controlling the repair process. The error 328 corresponds, for example, to the first partial derivation x _z of the edge position x according to the focus depth z, as obtained in the context of the claimed method from the focus-exposure matrix with regard to the edge position x and used to monitor the repair process as a quality metric or "adjusting screw “ can be used.

In the case of the mask 330, the edge steepness 338 is not as intended or desired (it should be mentioned that the figures in 3 are schematic representations, ie the desired edge steepness is shown as a vertical edge for simplicity; in reality, a beveled edge may well be desirable, but this case can also be covered with the claimed method). The edge steepness 338 (which is often described in the literature in the form of the "normalized image logslope", NILS) corresponds to the first partial derivative x _t of the edge position x according to the exposure dose t, as calculated from the focus Exposure matrix regarding the edge position x can be obtained and used to monitor the repair process as a quality metric or "adjusting screw".

In a real mask 350, these errors often occur in combination, i.e. due to the superimposition of the individual error contributions - as in 3 indicated by the arrow 340 - a complex mask defect 358 occurs, the repair of which can be difficult without suitable control options and "adjusting levers". The quality metrics provided by the method described here provide a remedy here, since they enable the composite error 358 to be resolved according to the individual error contributions (e.g. according to the error contributions indicated by the reference symbols 318, 328 and 338) and, moreover, a controlled influencing of the respective Allow contribution through suitable repair measures. For example, the error contribution as indicated at 328 can be avoided or corrected by precisely ending the etching/deposition process, which can be monitored and controlled largely independently of the other error contributions in the combined error 358 by monitoring the x _z quality metric. The edge steepness as indicated at 338 can be adjusted by adjusting the drift or Adjusted for jitter movement during the repair process and monitored and controlled largely independently of the other error contributions by monitoring the x _z quality metric.

4 shows the basis for a third-order model in the partial derivatives according to the depth of focus z and the exposure dose t, which is based on the in 3 goes beyond the possibilities described and can also be used to monitor and control a mask repair process.

A section 400 of the data of a focus-exposure matrix with regard to the edge position x is shown (possibly after this data has been "smoothed" by suitable interpolation or the like), with the deviation dx of the edge position x from a target value being shown on the horizontal axis x _o is plotted, and on the vertical axis the deviation dt of the exposure dose t from a nominal value t _o . Also shown are a number of curves 410, 420, 425, 430, 435, which correspond to different values of the depth of focus z. The curve 410 corresponds to the nominal value z _o , which represents the “best” focus for the data set shown here (at least in a local area surrounding the parameter point ( _zo , t _o )).

It should be noted at this point that in the following, for the sake of simplicity, it is assumed that the value of the edge position x at the nominal value (z _o , t _o ) corresponds to the target value x _o of the edge position (the nominal value (z _o , t _o ) becomes hereinafter also referred to as "isofocal point"). Using the nomenclature introduced at the beginning, dx̃ = x | applies (z _o , t _o ) - x _o = 0, and therefore dx = x - x _o . If the value of the edge position at the nominal value (z _o , t _o ) does not correspond to the target value x _o of the edge position, this deviation must be taken into account by adding a constant dx̃ to the deviation dx.

For the curve 410, which is the curve for the “best” focus z _o , it applies in the vicinity of the isofocal point ( _zo , t _o ) that it can be approximated by a tangent. Their gradient is referred to below as “sm” (from “slope max”). That means off 4 : $${\frac{\partial t}{\partial x}|}_{\mathrm{e.g}=\mathrm{e.g}0}=\mathbf{sm}$$

mit dem Korrekturparameter „zw“ (von engl. „z-width“). Dieser Paramater ist ein Maß für die Tiefenschärfe (engl. „depth off focus“, DoF) der Belichtungsanordnung, d.h. wie stark sich eine Abweichung vom „besten“ Fokus z_o in der Belichtung der Maske bemerkbar macht, oder anders gesprochen, wie groß der tolerierbare Fehlerbereich hinsichtlich einer (ungewollten) Defokussierung ist.If the depth of focus z deviates from the “best value” z _o , this can be described according to the invention by a correction term of the following form (in the lowest order): $${\frac{\partial t}{\partial x}|}_{\mathrm{e.g}\ne \mathrm{e.g}0}=\mathbf{sm}\ast \left(1-{\left(\mathrm{i.e}\mathrm{e.g}/\mathbf{between}\right)}^2\right)$$

with the correction parameter “zw” (from “z-width”). This parameter is a measure of the depth of focus (DoF) of the exposure arrangement, ie how much a deviation from the “best” focus z _o becomes noticeable in the exposure of the mask, or to put it another way, how large it is tolerable error range in terms of (unwanted) defocusing.

In addition to these contributions, there is also a telecentric error/phase error, which is described by the parameter “tce” (from “telecentric error”) and is included in the model according to the invention by the following contribution $$\frac{\partial x}{\partial \mathrm{e.g}}=\mathbf{tce}$$

$$\approx \mathbf{tce}\ast \text{d}z+\left(1/\mathbf{sm}\right)\ast \left(1+{\left(\text{d}z/\mathbf{zw}\right)}^2\right)\ast \text{d}t$$

Solving Eq. 10 after dx and addition of the contributions from Eq. 10 and Eq. 11 delivers $$\text{i.e}x=\mathbf{tce}\ast \text{i.e}\mathrm{e.g}+{\left[\mathbf{sm}\ast \left(1-{\left(\text{i.e}\mathrm{e.g}/\mathbf{between}\right)}^2\right)\right]}^{-1}\ast \text{i.e}t$$

$$\approx \mathbf{tce}\ast \text{i.e}\mathrm{e.g}+\left(1/\mathbf{sm}\right)\ast \left(1+{\left(\text{i.e}\mathrm{e.g}/\mathbf{between}\right)}^2\right)\ast \text{i.e}t$$

lässt sich Gl. 13 wie folgt schrieben: $$\text{d}x\approx \mathbf{tce}\ast \text{d}z+\left(1/\mathbf{sm}\right)\ast \text{d}t+\mathbf{sme}\ast \text{d}t\text{ d}z\text{ d}z$$

In the last step, the approximation 1 / (1 - y) ≈, 1 + y + ... was used. By introducing the further parameter $$\mathbf{sme}=\left(1/\mathbf{sm}\right)\ast {\left(1/\mathbf{between}\right)}^2$$

can Eq. 13 wrote as follows: $$\text{i.e}x\approx \mathbf{tce}\ast \text{i.e}\mathrm{e.g}+\left(1/\mathbf{sm}\right)\ast \text{i.e}t+\mathbf{sme}\ast \text{i.e}t\text{i.e}\mathrm{e.g}\text{i.e}\mathrm{e.g}$$

This equation is generally valid in the vicinity of a certain isofocal point (z _o , t _o ) (and assuming dx̃ = o, see above).

If one regards the position of the isofocal point itself as variable, one obtains if one uses Eq. 15 uses the definitions dx = x - x _o , dz = z - z _o and dt = t - t _o , a model for the edge position x as a function of the focal depth z and exposure dose t with the three parameters x _o , z _o and t _o and the three independent metrics tce, sme and zw (alternatively eg tce, sme and sm, or tce, sm and zw; cf. Eq. 14, which allows a conversion), which is used to describe and control a repair process of a mask as herein described can be used: $$\begin{array}{l}x-x_0\approx -\mathbf{tce}\ast {\mathrm{e.g}}_0-\mathbf{<figure-callout\; id="0"\; label="sme\; \ast \; t"\; filenames="DE102020208883B4\_0020.png,DE102020208883B4\_0021.png"\; state="\{\{state\}\}">sme</figure-callout>}\ast t_{}{}_0\ast \left({\mathrm{e.g}}_0^2+{\mathbf{between}}^2\right)\\ +t\ast \mathbf{sme}\ast \left({\mathrm{e.g}}_0^2+{\mathbf{between}}^2\right)\\ +\mathrm{e.g}\ast \left(2\ast \mathbf{sme}\ast {\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="DE102020208883B4\_0020.png,DE102020208883B4\_0021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0\ast {\mathrm{e.g}}_0+\mathbf{tce}\right)\\ -{\mathrm{e.g}}^2\ast \mathbf{sme}\ast {\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="DE102020208883B4\_0020.png,DE102020208883B4\_0021.png"\; state="\{\{state\}\}">t</figure-callout>}}_0\\ -t\ast \mathrm{e.g}\ast 2\ast \mathbf{sme}\ast {\mathrm{e.g}}_0\\ +t\ast {\mathrm{e.g}}^2\ast \mathbf{sme}\end{array}$$

The 5a-d , show a comparison of the related to the 4 and the Eq. 9 to 16 described model with simulated data of a mask for the DUV lithography under different exposure conditions. This data was used here because the simulation of a DLTV mask is well understood and the simulated data therefore provide a good "benchmark" for how the described model describes reality. The applicability of the model to EUV masks is fundamentally unaffected by this.

The data correspond to a mask with structural lines and intervening “lines and spaces” in a periodic sequence, which are to be imaged on the wafer with a period (“pitch”) of 76 nm in the x-direction (i.e. with a recurring line and free space width of 38 nm each on the wafer). The simulated data show the imaging properties of the mask at different values of depth of focus z and different values of exposure dose t.

The focal depth was varied stepwise between several focal planes, with "focal plane no. 7" corresponding to the nominal value z _o and two adjacent focal planes corresponding to a variation of the variable z of 20 nm each. The simulated data covers the “Focal Plane #1 - #13”, ie a focal depth range from -120 nm to +120 nm in steps of 20 nm in relation to the wafer surface.

The exposure dose t was also varied stepwise between seven different, equidistant doses, with the mean value, ie "dose no. 4", corresponding to the nominal value t _o . The range over which the exposure dose was varied corresponds to ±5% of the nominal dose t _o .

5a shows a section 500 of the (simulated) intensity distribution of the DLTV radiation on the wafer surface as a function of the x-position on the wafer. The zero point of the x-axis was chosen arbitrarily. A family of curves is shown which corresponds to the various values of the depth of focus z explained above. The curve identified by the reference sign 520 is the one at the nominal focal depth z _o and has the greatest amplitude of all the curves in the family of curves. The curves denoted by reference numeral 525 are the maximum defocused curves (ie at z= _zo ±120 nm) and these have the smallest intensity amplitude. In addition, a nominal value for the critical dose is drawn in at reference sign 530, and around this, at reference signs 532 and 535, a range of -5% and +5%, respectively.

In the area 510, the intensity of the curves is below the nominal value designated by reference number 530, so that here (at least assuming a perfect contrast of the photoresist) there is no removal of the photoresist during exposure, while outside this area the intensity is higher than the nominal value 530 the critical dose, i.e. the photoresist is removed during exposure (again, under the simplifying assumption of perfect contrast). Consequently, when the intensity curves pass through the critical dose 530 , structure edges form on the wafer, for example in the area designated by the reference symbol 550 . When the critical dose or the exposure dose t is varied, the position of this transition point, ie the position of the structure formation on the wafer, and thus the edge position x, generally shifts.

5b shows this relationship in the form of a section 501 of a focus-exposure matrix for the edge position x of the edge forming in the area 550 compared to its target value x _o . The vertical dashed line 520 in the middle of the figure corresponds to the nominal depth of focus z _o , ie the curve 520 in 5a with the greatest intensity amplitude. To the left of this there is negative defocusing (up to -120 nm), to the right of this there is positive defocusing (up to +120 nm).

The differently colored, essentially horizontal lines correspond to the different exposure doses t, with the middle line 540 corresponding to the nominal exposure dose t _o . The bottom line 542 corresponds to a deviation in the exposure dose from t _o by -5%, and the top line 545 to a deviation in the exposure dose from t _o by +5%.

The same facts are also presented by the 5c again, but this time in a different type of representation 502, in which the depth of focus z is plotted on the abscissa, the exposure dose t is plotted on the ordinate, and the deviation dx of the edge position x from its target value x _o for each pair of values (e.g , t) according to the at the right edge of the 5c color legend shown.

5d finally shows a comparison 503 in the 5a-c plotted (simulated) data with the predictions of the third-order model as given by Eq. 16 is described. The type of representation is the same as in 5c , ie the depth of focus z is plotted on the abscissa and the exposure dose t on the ordinate. However, the deviation of the data is now the 5a-c from the predictions from Eq. 16 (after fitting the parameters involved to the present dataset in a manner known to those skilled in the art) in color. As can be seen, at least in the area indicated by the broken line 560, the agreement of the model with the data (after appropriate parameter fitting) is extremely good and the difference between the model predictions with regard to the edge position x and those in the 5a-c The data shown is consistently below 0.02 nm in the range 560, mostly even below 0.01 nm.

This means that in a significant range around the nominal operating point (z _o , t _o ) the model described (cf. Eq. 16) can make very good statements about the influence of possible variations of the exposure parameters on the "effective" edge position x, and therefore can provide essential information for controlling and influencing the process as part of a repair process that attempts to adapt or correct this position.

The 6 finally shows further possible applications of the claimed method and the models described herein, eg in relation to 3 and 4 models described, in the mask repair.

The cutouts 600, 601 and 602 relate to a repair process that includes etching away excess material. Section 600 shows an SEM image of a mask section before the etching process, and section 601 shows an SEM image of the same mask section after the process has been completed. For better identification, the three "trenches" in section 601, for which excess material was etched away, are labeled "1", "2" and "3" in sequence.

The excerpts 610, 611 and 612, on the other hand, relate to a repair process that includes a material deposition of missing absorber material. Section 610 shows an SEM image of a mask section before the deposition process and section 611 shows an SEM image of the same mask section after this process. For better identification, the two "absorber lines" in section 611, in which additional absorber material was deposited, are marked "1" and "2" in sequence.

The two excerpts 602 and 612 each show a graphically prepared analysis of the results of these two repair processes, the numbering of the repaired trenches with “1”, “2” and “3” and the repaired absorber lines with “1” and “2” being retained became. The result of the respective repair process with regard to one or more of the qualities disclosed herein is shown in a color-coded manner ity metrics (e.g. deviation dx from the target value x _o , phase error/telecentricity error - x _z and/or edge steepness - x _t ).

For example, a certain color (e.g. the color green, cf. the areas 605 and 615) expresses that after the repair or at the moment during the repair process under consideration, the quality metric or quality metrics under consideration have already assumed a very good value, i.e. the repair was successful with regard to this metric(s) (e.g. no more than 5% deviation from a predefined target value). A different color (e.g. the color orange, see areas 608 and 618) can indicate that the corresponding quality metric(s) is currently in a medium range (e.g. between 5% and 10% deviation from the predefined target value). Another color (e.g. the color red) or no color at all could mean that the quality metric(s) under consideration has not yet reached an acceptable result (e.g. more than 10% deviation) and that the repair process, for example, should therefore continue.

Which quality metric or quality metrics are displayed visually in this way can be preset or can be selected by a user, possibly with the option of switching between different displays. If several quality metrics are selected for visualization, their "overlay" (e.g. their sum or average) can be displayed in a single image, or all metrics are displayed in a separate image, or mixed forms thereof.

In any case, such a visual (e.g. color-coded) representation allows an accurate and targeted monitoring of the repair process not only with regard to a single parameter, but several parameters in the form of the quality metrics presented herein can be included.

Finally, it should be mentioned that such a representation of the repair success can also be translated into a numerical or automated process (or vice versa), e.g. by defining corresponding intervals for the respective quality metric with the different colors (see above for a specific example for this purpose with the deviation ranges 0-5%, 5-10%, and >10%), so that in this way an automatic control of a repair process can take place (largely) without human intervention.

Claims (15)

Method for repairing a mask for lithography, in particular a mask for EUV lithography, comprising the following steps: 1a. determining a first value of an edge position (x) of a structure to be repaired on the mask, 1b. Carrying out a repair process comprising at least a first repair step on the mask, 1c. determining a second value of the edge position (x) after performing the first repair step, 1d. comparing the second value of the edge position (x) with the first value of the edge position (x) and/or a target value (x _o ) for the edge position, 1e. determining a focus exposure matrix for the edge position (x) with respect to an exposure system used for exposure, 1f. determining a deviation (dx) of the edge position (x) from the target value (x _o ) as a function of the focus depth (z) and the exposure dose (t) from the focus-exposure matrix, and 1g. Providing the variable determined in step 1f as a quality metric relating to the repair process. procedure after claim 1 , the method further comprising the following steps: 2a. Determining a dependency of the edge position (x) on the depth of focus (z) as a function of the depth of focus (z) and the exposure dose (t) from the focus-exposure matrix, 2b. determining a dependency of the edge position (x) on the exposure dose (t) as a function of the focus depth (z) and the exposure dose (t) from the focus-exposure matrix, and 2c. Provision of the variables determined in steps 2a and 2b as further quality metrics with regard to the repair process. procedure after claim 1 or 2 , further comprising: 3a. Adjusting a set of process parameters to perform the first and/or a second repair step of the repair process to improve at least one of the quality metrics, and 3b. Performing the first and/or second repair step using the adjusted process parameters. Procedure according to one of claims 2 - 3 , wherein the dependency of the edge position (x) on the focal depth (z) comprises the first partial derivative (x _z ) of the edge position (x) with respect to the focal depth (z). Procedure according to one of claims 2 - 4 , where the dependence of the edge position (x) on the exposure dose (t) comprises the first partial derivative (x _t ) of the edge position (x) with respect to the exposure dose (t). Procedure according to one of claims 2 - 5 , wherein the dependency of the edge position (x) on the depth of focus (z) represents a quality metric in the form of a measure for a phase error and/or a telecentricity error. procedure after claim 6 , wherein the repair process comprises an etching and/or deposition process and wherein the quality metric with regard to the phase error and/or the telecentricity error is influenced in a targeted manner by adjusting an end point of the etching and/or deposition process. Procedure according to one of claims 2 - 7 , where the dependence of the edge position (x) on the exposure dose (t) represents a quality metric in the form of a measure of an edge steepness. procedure after claim 8 , whereby the quality metric with regard to the edge steepness is purposefully influenced by adjusting a drift and/or jitter movement during the repair process. Procedure according to one of Claims 1 - 9 , further comprising the following steps: 10a. Determining at least a partial derivative of the second or higher order of the edge position (x) according to the depth of focus (z) and/or the exposure dose (t) from the focus-exposure matrix, and 10b. Providing the at least one partial derivative of the second or higher order as a further quality metric regarding the repair process. procedure after claim 10 , wherein the partial derivative of the second or higher order determined in step 10a comprises at least one of: $$x_{\mathrm{e.g}\mathrm{e.g}}={\partial }^{2}x/\left(\partial \mathrm{e.g}\partial \mathrm{e.g}\right),x_{t\mathrm{e.g}}={\partial }^{2}x/\left(\partial \mathrm{e.g}\partial t\right),x_{t\mathrm{e.g}\mathrm{e.g}}={\partial }^{3}x/\left(\partial \mathrm{e.g}\partial \mathrm{e.g}\partial t\right).$$

procedure after claim 11 , wherein the further quality metric includes a quality metric in the form of a measure of a depth of field, and wherein this quality metric is proportional to x _tzz . Procedure according to one of Claims 1 - 12 , further comprising evaluating at least one quality metric at a nominal depth of focus (z _o ) and a nominal exposure dose (t _o ) to obtain at least one quality number related to the repair process. Procedure according to one of Claims 1 - 13 , further comprising determining a critical dimension (CD) of the structure to be repaired and including the critical dimension in a quantification of the effect of the repair process. Computer program including instructions which, when executed, cause a mask repair device to carry out the steps of a method according to one of Claims 1 - 14 to perform.

Images