Measuring device, method for operating mask measuring device, and computer program product

Abstract

To provide a method for operating a mask-metrology measuring apparatus.SOLUTION: An image of a section of a photomask is recorded with a first image sensor (20) and an aerial image is generated by subjecting an image raw data obtained by the image sensor (20) to a clear normalization (45) and is subjected to a non-linearity adaptation comprising the following steps. In step a., the aerial image is mathematically combined with a clear image (47). In step b., a linearity correction is applied to the image data generated in step a. so as to correct a linearity error of the first image sensor (20). In step c., a non-linearity adaptation (P-1lin2) is applied to the linearity-corrected image data obtained in step b. so as to imprint a linearity signature of a second image sensor not arranged in the beam path of the measuring apparatus on the image data. In step d., a clear normalization is applied to the linearity-adapted image data generated in step c..SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present invention relates to a measuring device, a method for operating a mask measuring device, and a computer program product. [Background technology]

[0002] Photomasks are used in microlithography projection exposure systems, particularly for producing integrated circuits with small structures. The photomask (=reticle), irradiated with short-wave deep ultraviolet or extreme ultraviolet radiation (DUV or EUV radiation), is imaged onto the lithography object to transfer the mask structure onto the object.

[0003] For high-quality image representations to be produced on lithographic objects, it is crucial that photomasks that perfectly match the specifications are used. Mask measuring devices are used to inspect photomasks so that a description of their dimensional accuracy can be created.

[0004] In a mask measurement device, sections of a photomask are irradiated with radiation emitted from a radiation source, and the irradiated sections of the photomask are imaged onto an image sensor. The unprocessed image data acquired by the image sensor is set in relation to a clear image, i.e., an image without any mask structure. A clear image can be generated, for example, by approaching a clear position in the photomask where radiation can pass through the glass substrate of the photomask without being affected by the mask structure. The image recording that has undergone clear normalization is hereafter referred to as a spatial image.

[0005] After several years of operation, it may be necessary to replace the image sensor of a measuring device. Since the nonlinear behavior of two image sensors is generally not identical, the spatial image received by the user after the sensor replacement will differ from the spatial image before the replacement. This is undesirable, as many users have pre-configured routines for further processing of the spatial image. These routines can be negatively affected if deviations exist between the spatial images.

[0006] The present invention aims to provide a method, a measuring apparatus, and a computer program product for operating a mask measurement device such that the measurement results after the replacement of the image sensor correspond to the measurement results before the replacement of the image sensor. This objective is achieved by the features of the independent claim. Advantageous embodiments are specified in the dependent claims. [Overview of the project]

[0007] In the method according to the present invention for operating a mask measurement device, an image of a section of a photomask is recorded by a first image sensor. A spatial image is generated by clear normalization of the unprocessed image data acquired by the first image sensor. The spatial image undergoes nonlinearity adaptation, which includes the following steps: Step a., the spatial image is mathematically coupled with the clear image. Step b., linearity correction is applied to the image data generated in step a. to correct the linearity error of the first image sensor. Step c., nonlinearity adaptation is applied to the linearity-corrected image data acquired in step b. to imprint the linearity signature of a second image sensor, which is not located in the beam path of the measuring device, into the image data. Step d., clear normalization is applied to the linearity-adapted image data generated in step c.

[0008] The present invention recognizes that when linearity adaptations found for old and new image sensors are applied directly to unprocessed images, accuracy and repeatability losses occur. The present invention proposes correcting the clear normalized spatial image. In this way, a good agreement can be obtained between the measurement results obtained by the old image sensor before replacement and the measurement results recorded by the new image sensor. For the purposes of the present invention, the new image sensor corresponds to a first image sensor. The old image sensor corresponds to a second image sensor that is not located within the beam path of the measuring device.

[0009] Clear normalization sets the measured image data in relation to the clear image data. The clear image data is acquired by a beam path guided through a structure-free glass substrate, particularly through the clear position of the photomask. Thus, the clear image data provides information about what the image sensor sees when radiation emitted by a radiation source is guided directly to the image sensor, which has no structure being inspected. Clear normalization can be used to prevent the measured data from being tampered with by image sensor drift and to calibrate the field non-uniformity. Since clear normalization is performed pixel by pixel, the ratio of the image data of the image recording to the image data of the clear image is formed for each pixel of the image sensor.

[0010] The mathematical combination of the spatial image and the clear image according to the present invention can also be performed pixel by pixel by multiplying the image data measured for each pixel of the image sensor by the clear measurement value for the same pixel. The multiplication can be performed in normalized units, for example, on a scale ranging from 0 to 1. A value of 1 can be said to correspond to the saturation of the image sensor. The measuring device can be adjusted so that the clear image has an average intensity between 0.5 and 0.9, preferably between 0.6 and 0.8, on a relative scale.

[0011] Clear image C T2T The clear image used for calculations, referred to as a clear image, can be a clear image recorded by a new image sensor. For example, clear image C T2Tcan be the first clear image recorded within a measurement series. The linearity error of the new image sensor is the same as other image data recorded by the new image sensor, the clear image C T2T is affected. In one embodiment, before the clear image C T2T is used for clear normalization of the linearly adapted image data, non-linearity adaptation is also performed on the clear image C T2T .

[0012] The clear image C T2T can undergo linearity correction to correct the linearity error of the first image sensor. The first linearity correction can be the same first linearity correction that the non-linearly corrected measured image data also undergoes.

[0013] The clear image C T2T can undergo non-linearity adaptation to imprint the linearity signature of the second image sensor on the clear image C T2T . The second non-linearity adaptation can be the same non-linearity adaptation that the linearly corrected image data of image recording undergoes.

[0014] In a preceding method step that can also be a component of the method according to the invention, the measurement may have been performed to find the linearity error of the first image sensor. For this, a relationship can be established between the amount of radiation incident on the first image sensor and the image data that the first image sensor generates therefrom. The amount of radiation induced on the first image sensor can vary between zero and saturation of the image sensor in a relatively large number of measurements, from which a relationship can be derived as to what amount of radiation corresponds to which output signal of the image sensor.

[0015] To vary the amount of radiation induced in the image sensor during image sensor recording, the radiant power or exposure time may be adapted. If the radiant power is kept constant and the exposure time is varied, this has the advantage that the linearity error of the energy monitor can be eliminated. A beam splitter may be placed in the beam path of the measuring device so that a portion of the incident radiation is induced in the energy monitor. If the radiant power is kept constant and the exposure time is varied, this has the advantage that the linearity error of the energy monitor is not a problem because the energy monitor's reading remains constant.

[0016] If the first image sensor has a linearity error, a linear increase in the amount of radiation received during the exposure operation results in a nonlinear increase in the output signal. For this linearity correction, a calculation rule is found that can thereby convert the nonlinear profile of the output signal into a linear profile. The calculation rule can be in the form of a polynomial. The polynomial can be between the second and eighth order, and in particular, the polynomial can be a fifth-order polynomial.

[0017] The calculation rules for the linearity error of the second image sensor can be found in a corresponding manner. In many cases, it is useful to find the linearity correction for the second image sensor before the linearity correction for the first image sensor, because the second (older) image sensor is installed from the beginning, while the first (newer) image sensor is not yet installed.

[0018] In the method according to the present invention, the calculation rule for the linearity error of the first image sensor is applied directly so that non-linear image data is converted to linear image data. The calculation rule for the linearity error of the second image sensor is applied in reverse, on the other hand, so that the linearity signature of the second image sensor is imprinted on the image data which is linear after the application of the first linearity correction.

[0019] The radiation induced in the image sensor can be generated using a radiation source. The radiation source can be a laser radiation source. If the laser radiation source emits radiation pulses of constant energy, the amount of radiation induced in the image sensor during the exposure operation can be found by summing the energies of the radiation pulses. An energy monitor will show similar measurements for each of the radiation pulses. Despite the same radiation pulses, the measurements will generally not be completely identical due to noise. Any nonlinearity of the energy monitor does not affect the measurement at all. Alternatively, a lamp capable of maintaining a constant emitted radiation power can be used as a radiation source. Again, the amount of radiation received by the image sensor can thus be set by varying the exposure time.

[0020] The spatial image can be an energy-normalized spatial image. Therefore, both the image data of the image recording acquired by the first image sensor and the image data of the clear image used in clear normalization can be energy-normalized. Energy normalization means that the raw data recorded by the image sensor is set in relation to the amount of radiation used for recording. The amount of radiation used for recording can be found using an energy monitor as described above.

[0021] When recording spatial images, it is possible to proceed such that multiple image recordings are acquired after the recording of a clear image. Conventionally, it has been customary to perform clear normalization of measurement records based on the last clear image recorded up to that point. As a standalone solution of the present invention that can be independently applied to the nonlinearity adaptation described, it is proposed to use a first clear image and a second clear image for clear normalization, where the first clear image is temporarily recorded before the measurement record and the second clear image is temporarily recorded after the measurement record.

[0022] Between the first clear image and the second clear image, interpolation can be performed over time between the average intensity of the first clear image and the average intensity of the second clear image. The interpolation corresponds to the time profile of the clear state of the image sensor between the time of the first clear image and the time of the second clear image. The clear state derived from the interpolation can be used for clear normalization of the measurement record. This is advantageous when the clear state to be interpolated corresponds to the same time point when the image record was recorded.

[0023] The interpolation between the first clear image and the second clear image can be linear interpolation. Higher-order interpolation is also possible. According to current knowledge, the ratio between effort and effect in linear interpolation seems to be advantageous.

[0024] A further deviation between the first image sensor and the second image sensor can be attributed to the fact that the point imaging characteristics of the image sensors are different. This means that a point object does not result in an accurate point image representation on the image sensor and that the spread of the point object is different for the first image sensor compared to the second image sensor. The deviation can be mathematically described by the modulation transfer function MTF (modulation transfer function). In the method of the present invention, a first modulation transfer function MTF1 can be found for the first image sensor and a second modulation transfer function MTF2 can be found for the second image sensor. The modulation transfer function can be found, for example, by imaging different structures on the image sensor and then inferring the correction of the image sensor.

[0025] The correction can be performed by applying the modulation transfer function MTF1 to the image data recorded by the first image sensor. Different from linearity correction, the correction of the point imaging characteristics is performed not pixel by pixel, but rather as a correction in the frequency or Fourier domain over the entire area of the image sensor. In a further course of the method according to the invention, the inverse modulation transfer function MTF2 -1However, this can be applied to image data. The first modulation transfer function MTF1 can be applied before or after linearity correction. The second modulation transfer function MTF2 -1 This can be applied before or after nonlinearity adaptation.

[0026] Correction of the linearity error of the energy monitor can also be performed in the method according to the present invention. Therefore, it is convenient in the preceding method step to find the linearity error of the energy monitor by passing a signal to the energy monitor across its bandwidth and determining the ratio of the radiation incident on the energy monitor to the energy monitor measurement found therein. Image data recorded by the image sensor can be used as a reference for comparison. To eliminate any linearity error of the image sensor, the exposure time of the image sensor can be adapted such that the amount of radiation incident on the exposure sensor remains constant during the exposure operation. This means that the exposure time of the image sensor is kept short in measurement operations where a large amount of radiation is induced on the energy monitor during that time. The smaller the amount of radiation on the energy monitor, the longer the exposure time of the image sensor may be. The linearity-corrected measurement of the energy monitor can be used for energy normalization of the spatial image.

[0027] There are measuring devices in which image data recorded by older image sensors has already been corrected. These corrections, which may be due to adaptations to yet another image sensor performed at an earlier stage, are referred to below as default corrections. The method according to the present invention provides spatial image and clear image C T2T After mathematically combining the two and before applying the linearity correction, the default correction may be performed in reverse. After the nonlinearity adaptation is complete, the default correction may be performed again.

[0028] The present invention also relates to a mask measurement device comprising an image sensor, a computing module, and a correction module. The image sensor is designed to record an image of a section of a photomask. The computing module is designed to generate a spatial image by subjecting the raw image data acquired using the image sensor to clear normalization. The correction module is designed to subject the spatial image to nonlinearity adaptation. Nonlinearity adaptation consists of the following steps: Step a. The spatial image is mathematically coupled with the clear image. Step b. Linearity correction is applied to the image data generated in step a. to correct the linearity error of the first image sensor. Step c. Nonlinearity adaptation is applied to the linearity-corrected image data acquired in step b. to imprint the linearity signature of a second image sensor, which is not located in the beam path of the measuring device, into the image data. Step d. Clear normalization is applied to the linearity-adapted image data generated in step c.

[0029] The present invention also relates to a computer program product or set of computer program products comprising program components, the program components being designed to perform a method according to the present invention when loaded onto a computer or networked computer connected to a measuring device according to the present invention.

[0030] This disclosure includes the development of a mask measuring device having features described in the context of the method according to the present invention. This disclosure includes the development of a method described in the context of the mask measuring device according to the present invention.

[0031] The present invention will be described hereafter by example using advantageous embodiments with reference to the accompanying drawings. [Brief explanation of the drawing]

[0032] [Figure 1] This is a schematic diagram of the measuring device according to the present invention. [Figure 2] This is a view from above of the photomask being inspected by the measuring device shown in Figure 1. [Figure 3] This is a schematic cross-sectional view of the beam path of the measuring device from Figure 1. [Figure 4] Figure 1 shows a schematic diagram of the irradiation beam path of the measuring device. [Figure 5] This figure shows multiple consecutively recorded clear images and the average intensity of the clear interpolation derived from them. [Figure 6] This figure shows the various positions on the image sensor where the image representation of the measurement field can be generated. [Figure 7] This is a block diagram of the sequence of the method according to the present invention. [Figure 8] This is a diagram showing the linearity error of the image sensor. [Figure 9] This is another diagram of the linearity error from Figure 8. [Figure 10] Figure 7 shows an alternative embodiment of the present invention. [Figure 11] Figure 7 shows an alternative embodiment of the present invention. [Figure 12] Figure 7 shows an alternative embodiment of the present invention. [Modes for carrying out the invention]

[0033] The mask measurement device according to the present invention is used to inspect the structure of a photomask 17. The photomask 17 is intended for use in a microlithography projection exposure apparatus (not shown). In the microlithography projection exposure apparatus, the photomask 17 is irradiated with deep ultraviolet radiation (DUV radiation) having a wavelength of, for example, 193 nm, to image the structure formed on the photomask 17 onto the surface of a lithographic object in the form of a wafer. The wafer is coated with a photoresist that reacts to DUV radiation. The measurement device is used to inspect whether the structure on the photomask 17 corresponds to dimensional specifications.

[0034] In the measuring apparatus, the photomask 17 is positioned according to Figure 1 such that the beam path emanating from the laser emission source 14 passes through the photomask 17 and is guided to the image sensor 20. The emission has a wavelength of 193 nm, corresponding to the DUV emission used in microlithography projection exposure apparatuses. An irradiation system 16 is positioned between the laser emission source 14 and the photomask 17, and the irradiation system 16 expands the laser beam emitted by the laser emission source 14 so as to uniformly irradiate the measurement field 22 on the surface of the photomask 17 (see Figure 2). Using an imaging system 19, the structure of the photomask 17 is imaged onto the image sensor 20. The portion of the beam path between the laser emission source 14 and the photomask 17 is called the irradiation beam path 15. The portion of the beam path between the photomask 17 and the image sensor 20 is called the imaging beam path 21.

[0035] The photomask 17 is positioned within the measuring apparatus on an XY positioner 18, schematically shown in Figure 1. By moving the photomask 17 in the XY plane, the irradiation beam path 15 can be directed to different regions of the photomask 17. The measurement field 22 is limited by a field aperture 28 through which the irradiation beam path 15 is incident via a condenser optical unit 31.

[0036] Figure 4 schematically shows the irradiation beam path 15 between the laser source 14 and the condenser optical unit 31. The laser beam emitted by the laser source 14 is initially guided through a beam attenuator 32. The beam attenuator 32 is set so that the intensity of the laser beam matches the sensitivity of the image sensor 20. Using a prism arrangement configuration 33, the irradiation beam path 15 is deflected toward a first optical assembly 34, and the exit aperture of the laser beam source 14 is imaged onto a pupil-shaping mirror element 35 by the first optical assembly 34. The pupil-shaping mirror element 35 can be a mirror array comprising multiple mirror elements, which are movably suspended on a frame structure, and their orientation relative to the frame structure can be set individually.

[0037] The pupil-shaping mirror elements or mirror array 35 are positioned in the pupil plane 38 of the illumination system 16 so that the reflected radiation is distributed with uniform intensity across the measurement field 22 on the photomask 17. By adjusting one or more pupil-shaping mirror elements, the illumination setting, i.e., the angular distribution of radiation incident on the measurement field 22 thereon, can be varied.

[0038] Using a second optical assembly 36, the irradiation radiation is transmitted to a third optical assembly 37. The third optical assembly 37 comprises a diffractive optical element in the form of a field of view DOE (diffractive optical element) 43, the field of view which is generated by it, and a field aperture 28, the measurement field 22 which is defined by it. In the beam splitter 41, the irradiation beam path 15 is split such that a first portion of the radiation is guided to a photomask 17 via a condenser optical unit 31, and a second portion of the radiation is guided to an energy monitor 40.

[0039] The irradiation radiation is divided at a fixed ratio between the energy monitor 40 and the condenser optical unit 31, resulting in the measurement of the energy monitor 40 forming a measure of the amount of radiation incident on the photomask 17 via the condenser optical unit 31. Based on the measurement of the energy monitor 40, the image data recorded by the image sensor 20 can undergo energy normalization.

[0040] In the case of inspecting the photomask 17, multiple images are stored by the image sensor 20 in a time sequence. The photomask 17 is moved between recordings using the XY positioner 18 so that different areas of the photomask 17 are inspected.

[0041] For example, at regular time intervals of 15 minutes, clear images are recorded as the beam path passes through clear positions on the unstructured photomask 17. The clear images provide a reference for the measurement by recording an image unaffected by ordered structures within the beam path. The clear images may be used to perform clear normalization of recorded image data so as to reduce the effect of drift on the image data. The image produced by clear normalization of unprocessed image data is called a spatial image. The measuring device includes a computing module 23, in which the unprocessed image data recorded by the image sensor 20 undergoes clear normalization to generate a spatial image.

[0042] Previously, it was customary to use the last clear image recorded before image recording for clear normalization. Clear normalization establishes a relationship between the raw data of the image recording and the clear image. Clear normalization can be performed on a scale ranging from 0 to 1 in normalized units. In this case, 0 corresponds to the value provided by the image sensor 20 when the laser light source 14 is inactive, and therefore radiation from the irradiation system 16 is not directed towards the image sensor 20. A value of 1 corresponds to the value provided by the image sensor 20 in a saturated state. For example, the average intensity of the clear image can be 0.7. The average intensity of the image recording is lower than the intensity of the clear image.

[0043] Drift in the image sensor 20 can cause the average clear intensity to change significantly between two clear images. Therefore, changes in the clear images used for clear normalization result in abrupt changes between two consecutive spatial images.

[0044] According to the present invention, it is proposed to perform clear interpolation between two consecutive clear images. Therefore, the average intensity of the first clear image recorded at time T1 and the average intensity of the second clear image recorded at the second time point T2 are found, and linear interpolation is performed between time points T1 and T2. In Figure 5, the average intensity of the clear images recorded at four time points T1, T2, T3, and T4 is shown. CThis is shown as an example. The interval between clear images forms the period for linear clear interpolation, which is also shown in Figure 5.

[0045] The image representation 25 of the measurement field 22 generated on the image sensor 20 does not cover the entire area of ​​the image sensor 20 (see Figure 6). The position of the image representation 25 on the image sensor 20 shifts accordingly according to the position of the measurement field 22 on the photomask 17. Various possible positions of the image representation 25 on the image sensor 20 are shown in Figure 6. Average clear intensity I forms the basis for clear interpolation. C This is cropped to the target image size and found by section 26, which is positioned in the center on the image sensor 20.

[0046] Clear correction coefficient F CI This can be read from clear interpolation for each time point T at each time interval. Image correction is performed by applying a clear correction coefficient F to the spatial image acquired by the measuring device at time T, which is applicable at time T. CI This is done by dividing it pixel by pixel. The result is a corrected spatial image AI. CI That is the case.

[0047]

number

[0048] After prolonged use of the measuring device according to the present invention, it may be necessary to replace the image sensor installed in the measuring device with a new image sensor. If possible, this should be done in such a way that there is no change for the user of the measuring device, i.e., so that the measurement results for the new image sensor appear exactly the same as the measurement results for the old image sensor. This correspondence does not occur on its own because, generally, the linearity errors of the two image sensors are not the same.

[0049] To find the linearity error of the image sensor 20, a sequence of exposure operations can be performed that covers the bandwidth of the image sensor 20 from zero incident radiation to saturation. The actual amount of radiation incident on the image sensor 20 is measured by the measurement E of the energy monitor 40. mon The measurements of the image sensor 20 are obtained from the energy monitor 40. To prevent the measurements of the image sensor 20 from being tampered with by the linearity error of the energy monitor 40, the exposure operation may be performed such that the energy monitor 40 measures a constant value so that the possible linearity error of the energy monitor 40 does not have an effect.

[0050] Therefore, the laser beam source 14 of the measuring device can be set to emit laser pulses having a constant energy. The energy monitor 40 always sees the same amount of radiation when such laser pulses are incident. The amount of radiation incident on the image sensor 20 can be varied by changing the number of laser pulses during the exposure operation. The energy monitor 40 then measures the total energy of all the radiation pulses. mon To supply.

[0051] The measurement sequence is determined by the image sensor 20, which measures any value E without incident radiation. cam To find out what to provide, it can start with zero laser pulses. The number of laser pulses can be gradually increased until the image sensor 20 reaches saturation. An ideal image sensor 20 is E mon Along with, that is, with the number of laser pulses, the measured value E increases linearly. cam This will provide the necessary information, see Figure 8. The real image sensor 20 deviates from the ideal profile. In Figure 8, this is shown by the curves in the lower region below the ideal linear profile and the curves in the upper region above the ideal linear profile. In Figure 9, the same linearity error is given by the amount of incident radiation E mon It is shown again as a relative quantity to [the given quantity].

[0052] For example, the correction function is the measured value E camThe nonlinear profile can be calculated in the form of a fifth-degree polynomial, which thereby converts the nonlinear profile to a linear profile. When replacing an image sensor, such a correction function can be found for both the old image sensor and the new image sensor 20. The result is a first polynomial P for the new image sensor 20. lin1 and a second polynomial P for the old image sensor lin2 These correction functions can be used to convert image recordings made by the new image sensor 20 to appear as if they were made by the old image sensor.

[0053] Figure 7 shows the image sensor 20 storing raw data in the buffer 44 of the computing module 23. In step 45, the raw data of the image recording undergoes clear normalization to generate a spatial image stored in memory 46. The spatial image is affected by the linearity error of the image sensor 20 and therefore deviates from the spatial image that would have been generated if a different image sensor had been used instead of the image sensor 20.

[0054] The correction module 57 is used for image correction, and the correction module 57 reads the image data of the spatial image from memory 46. In the first step 47 of image correction, the spatial image is cleared. T2T It is mathematically coupled with the clear image C. T2T This is a clear image recorded at the beginning of the measurement sequence, and its average intensity is calibrated against a constant value stored in the database. This value corresponds to the average value of the clear image. The linearity error of the first image sensor 20, on which the unprocessed data is recorded, is given by the first polynomial P, which is used to determine the image data acquired from the spatial image. lin1 This is calculated in step 48 after applying a linearity correction using [the specified method].

[0055] In the subsequent step 49, nonlinearity adaptation of the second form of inverse linearity correction may be applied to imprint the linearity signature of the second image sensor onto the linearity-corrected image data. For this purpose, the second inverse polynomial P -1 lin2However, this is applied to the linearity-corrected image data to generate linearity-adjusted image data, and the linearity error of the first image sensor 20 is calculated from the linearity-adjusted image data. The linearity signature of the second image sensor is then imprinted on the linearity-adjusted image data.

[0056] In step 50, a final clear normalization is performed to generate an adapted spatial image from the adapted image data. To image the transition from the old image sensor to the new image sensor 20, and possibly during clear normalization, a clear image C T2T It undergoes a corresponding nonlinearity adaptation. In the first step, the linearity error of the first image sensor 20 is given by a polynomial P lin1 By applying this, a clear image C T2T This is how it is calculated. In the second step, the inverse polynomial P -1 lin2 However, linearity corrected clear image C T2T Applied to, clear image C T2T Adapting it to the linearity signature of older cameras. Linearly adapted spatial image AI T2T To generate the linearly adapted image recording, the image data is linearly adapted clear image C T2T It is divided by. When formulated as an equation, linearity-adapted spatial image AI T2T This arises from the spatial image AI1 recorded by the first image sensor 20 as follows:

[0057]

number

[0058] Linearly-adapted spatial image AI T2T This is output by the correction module 57 as a result of the measurement and can be made available for further use. Linearity-adapted spatial image AI for the user. T2T Because it appears as if it were recorded by an older image sensor, users can continue using it without modifying their normal routines.

[0059] In the alternative embodiment shown in Figure 10, energy normalization is performed in the calculation module 23 in a further step 51. Therefore, the raw data recorded by the image sensor 20 is set in relation to the measurement of the energy monitor 40 at the time of image recording. Thus, the clear image used for clear normalization can also undergo energy normalization. Therefore, the clear image is set in relation to the measurement of the energy monitor 40 at the time the clear image is recorded. The result is an energy-normalized and clear-normalized spatial image stored in memory 46. Image recording I, clear image C, energy normalized image recording E I , and the energy normalization E of the clear image C Therefore, this can be expressed as the following equation:

[0060]

number

[0061] During energy normalization, linearity correction of the measurements of the energy monitor 40 may also be considered. Similar to the image sensor, a correction function in the form of a polynomial may be found in place, and by this correction function, the non-linear measurements of the energy monitor 40 are converted into linear data. The required measurement data regarding the characteristics of the energy monitor 40 can be obtained by varying the energy measured by the energy monitor 40 while the image sensor 20 is observing an invariant radiation. This can be achieved by decreasing the exposure time of the image sensor 20 as the energy of the laser pulse increases. At the level of the correction module 57, linearity correction of energy normalization is not necessary because the error in question is averaged anyway when the image recording and clear recording are set relative to each other.

[0062] In further modifications, it should be considered that the new image sensor 20 and the old image sensor may have different point imaging characteristics. To compensate for this, a first modulation transfer function MTF1 may be found for the new image sensor 20, and a second modulation transfer function MTF2 may be found for the old image sensor.

[0063] As shown in Figure 11, two further steps will occur in the correction module 57. Following the correction of the linearity error of the new image sensor 20 and the Fourier transform into the frequency domain in step 48, the linearity-corrected image data in step 52 is divided by the modulation transfer function MTF1 to calculate the point imaging error of the new image sensor 20. Immediately thereafter, in step 53, multiplication by the modulation transfer function MTF2 may be performed to imprint the point imaging error of the old image sensor into the image data. An inverse Fourier transform back to the spatial domain is then performed. Alternatively, it is also possible to correct the point imaging error of the new image sensor 20 immediately after the image recording has been considered for the Fourier transform by directly dividing the raw data of the image recording by the modulation transfer function MTF1.

[0064] In a further embodiment shown in Figure 12, the default correction is performed in the calculation module 23 in a further step 54. The default correction may be introduced early to calculate the inaccuracies that occurred when using older image sensors. The default correction is a polynomial P def It is mathematically described by [the formula / method].

[0065] In the nonlinearity adaptation of the present invention, the results may be manipulated by such default corrections. Therefore, the method is modified by the polynomial P in step 55 prior to the linearity correction. def The default correction may be reversed by inverse application and by repeating the default correction after nonlinearity adaptation in step 56. This method involves the image recording itself and the clear image C T2TApplicable to both. In mathematical representation, linearity-adapted spatial image AI T2T It is then obtained as follows:

[0066]

number

[0067] Thus, the method according to the present invention can be used similarly when the image data of an older image sensor has already undergone default correction. [Explanation of Symbols]

[0068] 14. Laser emission source 15. Irradiation beam path 16. Irradiation System 17 Photomasks 18 XY Positioner 19. Imaging System 20 Image Sensors 21 Imaging beam path 22 Measurement field 23 Computation Modules 25 Image expression 26 sections 28 Field of View Aperture 31. Capacitor Optical Unit 32 Beam attenuator 33 Prism Arrangement Configuration 34 First Optical Assembly 35. Pupil reshaping mirror elements 36. Second Optical Assembly 37 Third Optical Assembly 38 Pupil plane 40 Energy Monitor 41 Beam Splitter 43. Field of View (DOE) (Diffraction Optical Element) 44 buffers 46 memory 57 Correction Module

Claims

1. A method for operating a mask measurement device, wherein an image of a section (26) of a photomask (17) is recorded by a first image sensor (20), a spatial image is generated by clear normalization (45) of the unprocessed image data acquired by the image sensor (20), the spatial image is subjected to nonlinearity adaptation, the nonlinearity adaptation is performed by the following steps, i.e. a. The spatial image is converted into a clear image (C T2T ) and mathematically combine them to generate image data of the mathematically combined spatial image (47), b. Apply linearity correction (P) to the image data of the mathematically combined spatial image generated in step a. lin1 Step (48) to apply the formula to correct the linearity error of the first image sensor (20) and obtain linearity-corrected image data. c. A step (49) to generate nonlinearity-adapted image data by applying nonlinearity adaptation (P -1 lin2) to the linearity-corrected image data obtained in step b, and imprinting the linearity signature of the second image sensor, which is replaced by the first image sensor (20) and is not located within the beam path of the mask measurement device, into the image data. d. Step (50) to apply clear normalization to the nonlinearity-adapted image data generated in step c. Methods that include...

2. The method according to claim 1, wherein the spatial image in step a is multiplied pixel by pixel by the clear image.

3. The clear image (C T2T The method according to claim 1, wherein ) is a clear image recorded by the first image sensor (20).

4. The clear image (C T2T ) undergoes nonlinearity adaptation, and the nonlinearly adapted clear image (C T2T The method according to claim 1 or 2, wherein the clear normalization in step d. is used.

5. The clear image (C T2T The method according to claim 4, wherein the nonlinearity adaptation of the first image sensor (20) includes linearity correction to correct the linearity error of the first image sensor (20).

6. The clear image (C T2T The nonlinearity adaptation of the second image sensor (20) is used to determine the linearity signature of the clear image (C T2T The method according to claim 4, comprising a nonlinear adaptation for imprinting on ).

7. The method according to claim 1 or 2, wherein in the preceding method step, the measurement is performed to find the linearity error of the first image sensor (20).

8. The method according to claim 7, wherein the measurement is performed while the energy monitor (40) of the measuring device provides a constant measurement value.

9. The method according to claim 1 or 2, wherein the spatial image is an energy-normalized spatial image.

10. The default correction (P def ) applicable to the second image sensor is calculated backward before step a. The method according to claim 1 or 2.

11. The method according to claim 1 or 2, wherein the clear normalization of the spatial image is based on a first clear image (T1) and a second clear image (T2), the first clear image (T1) being recorded before recording the image of section (26) of the photomask (17), and the second clear image (T2) being recorded after recording the image of section (26) of the photomask (17).

12. The method according to claim 11, wherein linear interpolation is performed over time between the average intensity of the first clear image (T1) and the average intensity of the second clear image (T2).

13. The method according to claim 1 or 2, wherein the point imaging error of the first image sensor (20) is corrected in the context of nonlinearity adaptation, and the point imaging error of the second image sensor is imprinted in the image data.

14. A mask measurement device comprising a first image sensor (20) for recording an image of a section (26) of a photomask (17), a calculation module (23) for generating a spatial image by clear normalization (45) of the unprocessed image data acquired by the first image sensor (20), and a correction module (57) which is designed to subject the spatial image to nonlinearity adaptation, the nonlinearity adaptation comprising the following steps, namely, a. The spatial image is converted into a clear image (C T2T ) and mathematically combine them to generate image data of the mathematically combined spatial image (47), b. Apply linearity correction (P) to the image data of the mathematically combined spatial image generated in step a. lin1 Step (48) to apply the formula to correct the linearity error of the first image sensor (20) and obtain linearity-corrected image data. c. A step (49) to generate nonlinearity-adapted image data by applying nonlinearity adaptation (P -1 lin2) to the linearity-corrected image data obtained in step b, and imprinting the linearity signature of the second image sensor, which is replaced by the first image sensor (20) and is not located within the beam path of the mask measurement device, into the image data. d. Step (50) to apply clear normalization to the nonlinearity-adapted image data generated in step c. A mask measuring device, including a mask measuring device.

15. A computer program for causing a computer to perform the method described in claim 1.

Images