MEASUREMENT APPARATUS AND METHOD FOR MEASURING ROUGHNESS AND / OR DEFECTS ON A SURFACE - Patent application

Abstract

The apparatus for measuring roughness and defects on a plurality of surface sections of an investigated sample surface comprises an illumination device with at least two light sources for illuminating a measurement area of ​​the surface with measurement light, a detector device with a detector array with a plurality of detector pixels for capturing scattered light scattered by the surface, and an evaluation device for determining at least one roughness feature of the surface from the captured scattered light. The at least two light sources illuminate the measurement area along at least two illumination beam paths at different angles of incidence with respect to a surface normal of the surface and may be fixed with respect to the detector device. The detector device has imaging optics for imaging the measurement area of ​​the surface onto the detector array and detects at least two scattered light images of the surface section in the illuminated measurement area at a predefined viewing angle with respect to the surface normal of the surface. The portion of the scattered light received by the detector pixels and formed for each case by the illumination in one of the illumination beam paths has a common spatial frequency for each case. The evaluation device determines at least one roughness feature of the surface section from the at least two scattered light images.

Description

[Technical field]

[0001] The present invention relates to a measuring device and a method for roughness and / or defect measurements on multiple surface sections of a surface of an investigated sample, in particular for roughness and / or defect measurements of structured surfaces such as burls of a holding device for a wafer. The present invention can be applied to the investigation of surfaces, in particular the surfaces of a wafer holding device, for example to detect the use or processing state of the surface. [Background technology]

[0002] The following prior art is referred to in this specification to provide the technical background of the present invention: [1]Germany 10 2012 005 417 B4, [2] M. Zerrad et al., “Development of a goniometric light scatter instrument with sample imaging ability,” in Proc. of SPIE, Optical Fabrication, Testing, and Metrology III, Vol. 7102, pp. 710207 (2008), doi: 10.1117 / 12.797621. [3] R. Bousquet et al., “Scattering from multilayer thin films: theory and experiment”, Journal of the Optical Society of America, 71(9), 1115 (1981). [4] A. Duparre et al., “Surface characterization techniques for determining the root-mean-square roughness and power spectral densities of optical components” in Applied Optics 41(1), 154-171 (2002), [5] E. L. Church et al., “Residual surface roughness of diamond-turned optics,” Applied Optics, 14(8), 1788-1795 (1975). [6] J.C. Stover, Optical Scattering: Measurement and analysis, 3rd ed. (SPIE, Bellingham, Wash., 2012). [7] P. Bobbert et al., “Light scattering by a sphere on a substrate,” in Physica A: Statistical Mechanics and its Applications, 137, 209-242 (1986), and [8] Sven Schroder, Alexander von Finck, and Angela Duparre, “Standardization of light scattering measurements,” in Adv. Opt. Techn 2015; 4(5-6): 361-375.

[0003] Holding devices are generally known for holding components, for example silicon wafers in lithographic semiconductor processing, in particular chip manufacturing. If the holding is achieved using electrostatic forces, the holding device is also called electrostatic holding device, electrostatic clamping device, electrostatic clamp (ESC) or electrostatic chuck. The holding device typically comprises the shape of a plate with at least one planar carrier surface, by which the component to be held is carried. The carrier surface is formed by a plurality of protruding burls, in particular by the flat end faces of the burls. For example, several thousand burls are provided, each with an end face diameter in the sub-mm range.

[0004] It is generally known that the roughness of the end face of the burr and the possible appearance of defects such as contamination and / or material damage decisively influence the interaction properties such as friction, adhesion and slipperiness between the wafer and the ESC. The interaction properties between the burr and the wafer in turn decisively influence the lateral instabilities when clamping the wafer and therefore the achievable quality and the overlay of the structures coming from the different manufacturing steps in the case of lithographic structuring. For a highly accurate and reliable operation of the wafer, the roughness should be set as reproducibly as possible and possible defects should be reliably identified. In this case, for example, 0.3 μm -1 Roughness structures having high spatial frequencies exceeding 1000 nm are of particular interest. It is also intended to prevent the deterioration of the uniformity of the held wafer and the quality of the lithography process due to defects, especially coarse roughness structures or impurities, such as particles or material damage at the end face. Therefore, it is of interest to detect the roughness of the burl end face during production, during operation, and / or in case of rework of the ESC.

[0005] A conventional investigation method typically used in practice consists of measuring the roughness using an atomic force microscope (AFM). Individual burls are approached and investigated with the AFM. It is possible to directly measure roughness structures with high spatial frequencies by AFM, but the AFM only covers very small areas, typically 50x50 μm. 2 The disadvantages are that AFM has a smaller measurement field than 100 nm and AFM measurement is very time-consuming, so the use of AFM is limited to the observation of a few representative burls. For example, measuring all ESC burls on a surface with a diameter of 300 mm requires approximately 200 days, which is not practical. Therefore, in practice, local roughness changes or impurities on the carrier surface cannot be detected or can only be detected unreliably by AFM measurement. Also, AFM cannot perform comprehensive detection of defects.

[0006] Alternatively, the burls can be measured by white light interferometry (WLI). This allows for a higher measurement speed than when using AFM, but requires measurement times that are too long for practical applications. Also, due to the low-pass nature of the imaging objective, the burls can be measured with a resolution of approximately 0.3 μm. -1 Only smaller spatial frequencies can be reliably detected with a larger WLI measurement field of view.

[0007] From [1], [3], [4], [5] and [6] it is known to detect the power spectral density function (PSD function) of optical components, such as lenses, which are characterized by roughness, by scattered light measurements. Typically, for this purpose, the scattered light distribution at a local measurement point of the sample is acquired by an angle scanning detector. In the method described in [1], a measurement spot on the surface of the sample to be investigated is illuminated at various angles or using various wavelengths, and the light scattered at the measurement spot is acquired by several detectors at several viewing angles. The PSD function is determined from the detector signals. However, the disadvantage of this method is that it measures only in time and locally, and is therefore very time-consuming as a lateral scanning method. The detection of all the individual bars of the ESC lasts approximately two days, which is approximately the same as the WLI measurement time. Moreover, the conventional method provides only limited information about the cause of the increase in scattering. Furthermore, the method according to [1] has a limited field of application, since the spatial resolution is limited by the diameter of the illumination spot.

[0008] Another scanning scattered light measurement is the goniometric method known from [2], in which an illumination source is scanned over the sample and scattered light images of the light scattered at different illumination angles are acquired by a camera detector. This method is suitable for example for detecting particles on a wafer surface. From the scattered light image it is possible to determine whether and where in the imaged area a scattering particle is present. If a particle is detected, a precise observation of the surface section can be provided in a locally localized manner. The method according to [2] is also characterized by high time requirements and limited fields of application.

[0009] Conventional scanning scattered light methods further have the disadvantage that, for measurements on an ESC, several burls are illuminated simultaneously, since typically a measurement field diameter of 1 mm to 3 mm is achieved. Complex illumination optics may make it possible to achieve a measurement field diameter of less than 0.2 mm, but this requires a highly accurate absolute positioning system (sample coordinates), and the individual positions of the burls must be accurately detected and transmitted to the positioning system. Also, when detecting the bur end faces, there would be a risk that the scattered light of the edge structures, i.e. the diffraction effects, would be superimposed on the scattered light of the end face topography. Measures to prevent this risk would slow down the scanning process. Furthermore, small scattering angles, smaller than approximately 2° to 5°, cannot be detected, because for measurement field diameters smaller than 0.2 mm the illumination beam must be focused on the sample, and not in the detector plane, as is otherwise conventional. Thus, the specularly reflected illumination beam is reflected divergently and superimposed on the scattered light at small scattering angles.

[0010] The above mentioned limitations, especially with regard to the speed, spatial resolution and spatial frequency of the roughness measurement or defect identification, arise not only in the case of measurements on the burl end faces, but also in the case of measurements on the surfaces of other components, e.g. optical components, in particular lenses, wedges or mirrors.

[0011] The object of the present invention is to provide an improved measuring device and an improved method for roughness and / or defect measurements on multiple surface sections of the surface of a sample to be investigated, whereby the drawbacks and limitations of the prior art are avoided. The invention is intended in particular to make it possible to perform roughness and / or defect measurements with increased speed and / or accuracy, to detect structures with increased spatial frequency, to reduce the processing complexity of the data processing and / or to deliver roughness and / or defect measurement results with increased information content. The roughness and / or defect measurements should in particular be suitable for a reliable and rapid detection of the roughness of burls of a holding device for holding a wafer, as far as possible with a measurement time of approximately one hour or less when detecting all burls.

[0012] This object is in each case achieved by a measuring device and a method with the features of the independent claims. Preferred embodiments and applications of the invention follow from the dependent claims. Summary of the Invention

[0013] According to a first general aspect of the invention, the above mentioned object is achieved by a measurement device configured for roughness and / or defect measurements on multiple surface sections of a surface of a sample to be investigated, in particular on structured surfaces such as burl end faces of a holding device for holding a wafer.

[0014] The measuring apparatus comprises an illumination device having at least two light sources arranged to illuminate a measurement area of ​​the surface with measurement light, the measurement area comprising a plurality of surface sections. According to the invention, the at least two light sources are configured to illuminate the surface sections in the measurement area along at least two illumination beam paths at different angles of incidence with respect to a surface normal of the surface. Also, the at least two light sources may be fixed with respect to the detector device.

[0015] The measuring apparatus further comprises a detector device having a detector array consisting of a plurality of detector pixels arranged to detect scattered light scattered at the surface. According to the invention, the detector device is provided with imaging optics, whereby a measurement area of ​​the surface is imaged onto the detector array, and the detector device is configured to capture at least two scattered light images of the illuminated measurement area at a predetermined viewing angle relative to a surface normal of the surface, the portion of the scattered light received by the detector pixels having a common spatial frequency in each case in each of the illumination beam paths.

[0016] The measurement apparatus also comprises an evaluation device configured to determine at least one roughness characteristic of the surface from the captured scattered light. According to the invention, the evaluation device is configured to determine at least one roughness characteristic for each surface section from the at least two scattered light images.

[0017] According to a second general aspect of the invention, the aforementioned object is achieved by a method for roughness and / or defect measurement of a surface section of a surface of an investigated sample, in particular a structured surface such as a burl end face of a holding device for holding a wafer.

[0018] A method for roughness and / or defect measurement includes illuminating a measurement area of ​​a structured surface with measurement light along at least two illumination beam paths at different angles of incidence relative to a surface normal of the surface, the measurement area encompassing a plurality of surface sections, capturing scattered light scattered by the surface using a detector device comprising a detector array consisting of a plurality of detector pixels, and determining at least one roughness characteristic of the surface from the captured scattered light.

[0019] According to the invention, the illumination comprises a directed illumination of the surface section in the measurement area along at least two illumination beam paths at different angles of incidence relative to the surface normal of the surface using at least two fixedly positioned light sources. A detector device provided with imaging optics, by means of which the illuminated measurement area of ​​the surface is imaged onto a detector array, captures at least two scattered light images of the illuminated measurement area at a predefined viewing angle relative to the surface normal of the surface, the portion of the scattered light received by the detector pixels having a common spatial frequency in each of the illumination beam paths in each case. At least one roughness feature is determined from the at least two scattered light images for each surface section.

[0020] The method according to the invention for roughness and / or defect measurement or an embodiment thereof is preferably carried out using a measurement device according to the first general aspect of the invention or an embodiment thereof.

[0021] The roughness measurement and / or defect detection is according to the invention performed on surface sections of a surface. Each surface section is a part of the surface from which scattered light is imaged onto individual detector pixels or groups of detector pixels. All surface sections considered are preferably of the same shape and surface area. The totality of the surface sections can completely or partially cover the surface to be investigated.

[0022] The roughness measurement and / or defect detection is preferably carried out on a structured surface. The term "structured surface" refers to any surface that is closed or discontinuous, i.e. preferably consists of partial surfaces with parts that are separated from one another and have different surface structures or textures. In the case of structured surfaces, the surface sections considered are particularly preferably parts of the structured surface that have the same surface structure or texture in each case. That is to say, surface sections are parts of a surface that are separated from one another and locally have the same surface structure or texture that differs from the surface structure or texture of the remaining surface.

[0023] For example, the investigated sample comprises a holding device for electrostatic holding of a component, on a surface of the holding device a plurality of protruding burls are arranged, the surface of the holding device forms a structured surface, and the surface sections comprise end faces of the burls.

[0024] Contrary to the method known from [1], according to the invention at least two light sources are arranged in such a way that at least two different illumination beam paths are formed, which have different angles of incidence relative to the surface normal of the surface. The measurement light is directed at the sample at the at least two different angles of incidence, each light source being arranged to illuminate at one of the angles of incidence. In particular, during the recording of the at least two scattered light images, the light sources are fixed relative to the detector device. The amplitude of the scattered light detected by the detector device is determined at each angle of incidence by the scattering of the measurement light at a scattering structure with a defined spatial frequency. In other words, at each angle of incidence, scattered light images with different spatial frequencies are detected. This allows for a larger spatial frequency range, in particular below 0.005 μm. -1 ~100μm -1 This advantageously enables roughness characteristics to be determined at spatial frequencies within a spatial frequency range of .

[0025] Including high spatial frequencies in determining the at least one roughness feature is particularly advantageous for roughness and / or defect measurements at the burl end faces of a holding device, e.g. an ESC, since structures with high spatial frequencies have particular influence on the adhesion properties of the burls and thus on the function and characteristics of the holding device, in particular preventing lateral instabilities when clamping the wafer.

[0026] Roughness and / or defect measurements are performed in a different way compared to the method described in [2] when the incidence angle is rigidly set. The movement of the light source for setting different incidence angles is avoided, which simplifies the construction of the measurement device and accelerates the roughness measurement. A further advantage of illumination using a rigidly set incidence angle is that the fixed incidence angle allows a fixed positioning of the beam trap, which is optionally provided. The beam trap absorbs the direct reflection from the surface and thus suppresses stray light due to light scattering of reflections in the measurement system. In the case of the method described in [2], the beam trap has to be transported with the reflections, which results in an increased complexity of the conventional measurement.

[0027] Also, unlike [1], the detector device is configured for image acquisition by providing imaging optics. Advantageously, simultaneous detection of the scattered light amplitudes of several surface sections is made possible, as a result of which the roughness and / or defect measurement of the entire surface is significantly accelerated and a higher lateral resolution can be achieved more easily. The high resolution is also advantageous with regard to surface section identification, defect identification and defect classification. In the case of the method described in [1], the illumination spot size has to be significantly reduced for this purpose, resulting in the already mentioned disadvantages (long measurement times, overlap of small scattering angles due to the need for focusing).

[0028] The term "roughness feature" refers to any quantity or data quantity, such as a function course, that quantitatively represents the roughness (non-uniformity of the surface height of the surface being investigated). The roughness feature is a measure of the non-uniformity due to topographical non-uniformity and / or defects. Advantageously, it is possible to distinguish different roughness features of a surface section. The at least one roughness feature to be identified can be selected, for example, depending on the requirements of a particular application. The roughness feature is preferably derived from an ARS value (ARS: angle resolved scattering) and / or a variable derived therefrom, such as an integrated scattering value and / or at least one PSD function (power spectral density function), from which the rms roughness (rms: root mean square) can be determined. Particularly preferably, the PSD function and / or the rms roughness are obtained as roughness feature.

[0029] The defects include contamination and / or material damage on the surface, in particular on the surface section under consideration. The defects are preferably identified from at least one roughness feature and / or angle resolved scattering.

[0030] According to an advantageous embodiment of the invention, the illumination device may be configured for time-controlling at least two light sources such that at least two scattered light images can be recorded temporally separated from each other, the evaluation device is configured for determining a PSD function of a surface section within the illuminated measurement area from the at least two scattered light images, and the illumination device is configured for calculating at least one roughness feature for each surface section from the scattering function, in particular an ARS function, of the surface section.

[0031] The time control of the at least two light sources such that at least two scattered light images are recorded sequentially and separated in time from one another comprises successively alternatingly activating one of the light sources in each case, such that the activated light source is on while all other light sources are inactivated (switched off or masked). The PSD function of each surface section is determined in a manner known per se from the amplitude of the scattered light images, as will be explained in more detail below. The calculation of at least one roughness feature for each surface section may for example comprise the calculation of a roughness value from the average PSD function of the surface section and / or the PSD variance within the surface section.

[0032] The PSD function may be determined for each detector pixel of the detector array of the detector device. If scattered light from a single surface section, for example from a burl, is received by multiple detector pixels, for example 2-20 detector pixels, an average scattered light amplitude may be calculated for each surface section and the PSD function may be determined from the average scattered light amplitude at different angles of incidence. Alternatively, the PSD functions calculated per pixel may be averaged. Averaging advantageously allows to reduce the amount of data to be processed. Using multiple detector pixels per surface section also provides advantages for identifying defects.

[0033] Alternatively or additionally, according to a further advantageous embodiment of the invention, the illumination device can be configured for intensity control, whereby it is possible to record at least two scattered light images with a weighted irradiance (energy per surface area) of directed illumination by the at least two light sources at each surface section. Preferably, a weighted irradiance is provided for the at least two scattered light images to be recorded simultaneously at each surface section or at a group of surface sections.

[0034] The weighted irradiance of directed illumination is provided at each of the surface sections by at least two light sources in that the irradiance, e.g., the illumination intensity of the light sources, is set such that the light sources correspond to uniform hemispherical illumination. The weighted irradiance (E) is, in particular, calculated as sin(θ i )cos(θ i ) and is set to be proportional to θ i is the angle of incidence of the illumination of the respective surface section. In this case, the evaluation device is configured to determine the integrated scattering (S function or S value) or in particular the total scattering (TS function or TS value, equal to the integrated scattering in the scattering angle range 2 to 85°) of the structured surface in the illuminated measurement area from the at least two scattered light images and to calculate at least one roughness feature for each surface section from the integrated scattering over the surface section.

[0035] Embodiments of the present invention comprising intensity control and simultaneous recording of at least two scattered light images have the particular advantage that scattered light images can be recorded using just one single instance of illumination, thus significantly shortening measurement periods.

[0036] Intensity control of the illumination device is achieved in particular in that the light sources are operated with different energies, in particular different powers, pulse lengths and / or pulse width modulations, and / or that a predefined reducer is provided in the illumination beam path, in particular in the light source, such that each of the surface sections is illuminated with a weighted irradiance, the reducer preferably comprising a beam-shaping optical element and / or a position-variable reducer such as a grey filter.

[0037] Embodiments of the present invention with time control and with intensity control can be combined, thereby advantageously increasing the information content of the measurement.

[0038] The number of light sources can be advantageously selected depending on the concrete conditions of use of the invention, in particular depending on the desired speed of the measurement. Two light sources are sufficient to provide at least two illumination beam paths, for example to construct a PSD function. A larger number of light sources can be advantageous to increase the accuracy of the measurement, but also leads to an increase in the duration of the measurement, including data processing. The number of light sources of the illumination device is preferably selected to be in the range of 3 to 70 or more, for example up to 100. In a practical application of roughness measurements in a holding device, for example 10 to 60 light sources are provided. In a suitable configuration of light sources, for example end-emitting optical fibers, it is also possible to provide more light sources, for example up to 100 or more.

[0039] The light source is preferably arranged perpendicular to the surface of the sample to be investigated in at least one azimuthal plane, advantageously resulting in a compact configuration. Alternatively, the light source may be arranged differently, in particular on multiple sides beside the detector device. In particular, illumination along the illumination beam path may be achieved at different azimuthal angles, in particular in multiple azimuthal planes, achieving advantages in the identification or suppression of interfering local reflections or high dispersion coefficients that may be superimposed on nearby measurement points. In addition, this arrangement simplifies the identification of anisotropic surface structures and defects.

[0040] According to a further advantageous embodiment of the invention, the illumination device is configured to generate the measurement light to be at the same wavelength in all illumination beam paths, in which case the construction of the measurement arrangement and the determination of the at least one roughness characteristic are simplified.

[0041] According to an alternative embodiment of the invention, the illumination device is configured to generate the measurement light to be of different wavelengths and / or different polarizations in all radiation beam paths. The different wavelengths are provided, for example, by using light sources with appropriate spectral emission. To provide the different polarizations, for example, adjustable polarizing beam splitters and / or polarizing filters are used. Advantageously, an additional degree of freedom is provided in the selection of the spatial frequencies detected in the illumination beam path. It also becomes easier to detect whether the scattering structures are surface defects or surface structures or are non-topographically scattering, and thus defects can be better identified. If defects are identified, a decision can preferably be made regarding a subsequent cleaning of the surface. Alternatively, if surface structures are identified, a post-processing of the surface, for example of the burl end face, can preferably be performed. Such a possibility of differentiation advantageously allows a significant reduction in post-processing process times and therefore in costs.

[0042] According to a preferred variant of the use of different wavelengths in the illumination beam paths, the illumination device is configured to generate measurement light in at least one of the illumination beam paths such that it has at least two different wavelengths and / or such that it has at least two different polarizations.

[0043] The evaluation device is particularly preferably configured to determine at least two PSD functions, in particular having equal spatial frequencies (i.e. common spatial frequencies or spatial frequency ranges), and / or at least two instances of integral scattering, in particular having the same spatial frequency range, of a surface section in the illuminated measurement area, and to identify structures on the surface section from the at least two PSD functions and / or the at least two instances of integral scattering. The structures on the surface section can be obtained from the deviation, for example the difference or quotient, of the at least two PSD functions and / or the at least two instances of integral scattering. Thereby, the distinction between defects and surface structures is advantageously further simplified.

[0044] According to a further embodiment of the invention, the evaluation device is configured to identify surface sections in the measurement area, in particular in the scattered light images. The determination of the at least one roughness feature may in particular be achieved from at least two scattered light images for each surface section. Alternatively, only a subgroup of the surface sections to be evaluated may be provided. The identification of the surface sections in particular in the scattered light images improves the accuracy of the determination of the at least one roughness feature, and undesired interferences due to possible scattered light from edge structures of the surface sections or due to diffraction effects can be detected and prevented.

[0045] Advantageously, image regions in the scattered light image can be associated with surface patches (deterministic structures) and distinguished from stochastic structures such as defects or impurities. Calculation of the at least one roughness feature may be limited to the surface patches, and information may be averaged over individual surface patches.

[0046] The identification of the surface compartment can be achieved by image recognition methods to which the scattered light image is subjected. According to a particularly preferred embodiment of the invention, the evaluation device is configured to identify the surface compartment by using neural networks and / or machine learning methods. This embodiment of the invention is characterized by a combination of an imaging scattered light measurement method tuned to very high spatial frequencies and a learning numerical method. Advantageously, this makes it possible to increase the reliability of the identification of the surface compartment even when defects are present.

[0047] The evaluation device is preferably configured to identify structures on the surface section from the at least two scattered light images by using a digital image mask. The digital image mask comprises, for example, predefined reference images of predefined shapes and / or dimensions that are characteristic of defects expected on the surface. The defects comprise, for example, foreign body contamination such as particles or fibers, or material damage such as edge breaks or cracks. Comparing the image mask with the scattered light images allows the defects to be identified and, optionally, classified (e.g. dust, fiber, edge break, etc.) and / or sized. The use of a digital image mask advantageously allows for accelerating the evaluation of the scattered light images.

[0048] The image mask may in particular make it easier to only partially evaluate the scattered light image, for example to calculate roughness from the scattered light data only on surface sections, especially on burls. In the case of particle identification, it is possible that the scattered light data of the particles can be extracted using the image mask in order to use other evaluation algorithms for size determination as well. The creation of the mask and the classification of the defects can be achieved by machine learning methods, in particular neural networks. The evaluation of the extracted scattered light data can be carried out using models known per se, the choice of the model being made depending on the type of defect identified. For example, individual particles can be assessed from the ARS data using the Bobbert Vlieger model for size determination, defects can be assessed by size determination by pixel counting / surface determination and / or burls can be assessed by surface roughness calculation from the ARS data using the Rayleigh Rice model.

[0049] According to a further preferred embodiment the imaging optics of the detector device are telecentric optics, which offers advantages in terms of the construction of the measurement apparatus, since no complex separate optical elements are required.

[0050] According to the invention, it is sufficient that the scattered light image is recorded by the detector device at a fixed viewing angle relative to the surface of the sample being investigated. The viewing angle of the detector device preferably extends parallel to the surface normal of the surface of the sample being investigated. Telecentric optics are advantageous in that the viewing angle or scattering angle is identical for each pixel and therefore the spatial frequency is also the same for all pixels. This is also advantageous in that lateral dimensions can be measured with greater precision by viewing the sample preferably perpendicularly over the entire field of view. Also, the solid angle is the same for all pixels, which is advantageous in terms of calibration.

[0051] Further advantages, especially with regard to observing the burl end faces of the holding device, are provided if the illumination device is configured for directed illumination of the surface section at a shallow angle of incidence in the range of 88°-75° using measurement light with a wavelength λ smaller than 300 nm, and the evaluation device is configured for determining a skewness parameter of the surface section. The skewness parameter is a statistical quantitative measure of the type and degree of topographical asymmetry of the individual end faces from their average value. Determining the skewness parameter makes it possible to determine certain interaction characteristics of the wafer with the burls, which are decisive for the lateral distortion of the wafer.

[0052] According to a further advantageous embodiment of the invention, at least one further detector device is provided, comprising an imaging optics and a detector array with a plurality of detector pixels, the at least one further detector device being configured to capture at least two scattered light images of the illuminated measurement area at at least one further field of view angle relative to the surface normal of the surface. For example, at least one second detector device is provided, which forms a larger field of view angle relative to the sample normal of the sample to be investigated than the first detector device. Thus, it is advantageously possible to detect structures with even higher spatial frequencies.

[0053] A further embodiment includes at least one illumination, the angle of incidence of which is selected such that the direct specular reflection falls on at least one additional detector. Advantageously, the measurement of the additional detector allows the local reflection coefficient of the sample to be determined, which increases the accuracy of the evaluation. To determine the reflection coefficient at an angle of incidence parallel to the surface normal, a beam splitter can be used in the optical path of the detection beam path, which allows both viewing the sample parallel to the surface normal and illuminating the sample parallel to the surface normal.

[0054] A further advantage of the present invention is that the illumination device is not limited to a particular type of light source. The term "light source" refers to any form of light emitting device arranged to illuminate the sample being investigated at a given angle of incidence.

[0055] The light source is preferably an active light emitting element, such as a light emitting diode or a laser, whose light emission is directed towards the sample. Light emitting diodes or lasers have the advantage of the availability of components with high illumination intensity, the availability of narrow band emitters, and directed light emission. Light emitting diodes may be advantageous due to their incoherence, since no interfering speckle pattern is generated on the sample. The laser may include, for example, a laser diode or a solid-state laser.

[0056] Alternatively, the light source includes a passive light emitting element (illuminated element) such as an optical fiber, in particular a fiber end of an optical fiber, or a reflector, in combination with a light emitting diode or laser. Several or all of the optical fibers may be coupled to a common light emitting diode or a common laser. In these variations, advantages are obtained with respect to a dense arrangement of several light sources, e.g. fiber ends, to set several angles of incidence.

[0057] The illumination device may comprise the same or different variations of light sources with different angles of incidence, for example a laser and / or light emitting diode that directly illuminates the sample may be combined with an optical fiber coupled to the same or another laser and / or light emitting diode.

[0058] Each light source is preferably provided with illumination optics, which provide advantages with regard to the orientation of the illumination towards the sample and / or the formation of an illuminated measurement area on the surface of the sample. If the light source comprises a light emitting diode or an optical fiber, the illumination optics is preferably a collimating lens. If the light source comprises a laser, the illumination optics is preferably a combination of an expander lens and a collimating lens. Alternatively, mirror optics can be used to reduce the inherent scattered light of multiple instances of illumination or to improve aberrations and transmission or reflection losses over a wider spectral range of illumination.

[0059] The scattered light image can cover the entire surface of the sample to be investigated. Alternatively, the scattered light image can cover a part of the surface of the sample to be investigated, and the entire surface can be captured by repetition of measurements on different parts. In the latter case, according to a further preferred embodiment of the invention, the sample holder for receiving the sample and / or the assembly consisting of the illumination device and the detector device are movable relative to each other, whereby the illumination device can be arranged to illuminate different measurement areas (parts) of the structured surface, the scattered light images of the different measurement areas can be captured by the detector device, and the measurement device is configured to perform repeated roughness measurements on surface sections in the different measurement areas of the structured surface. Each part comprises at least two, typically at least 20 surface sections. In case of roughness measurements on a holding device, one part covers for example approximately 50 burl end faces.

[0060] The roughness measurement can advantageously be combined with further investigation and / or processing methods. For example, according to a preferred variant of the invention, selected surface sections, particularly those with increased roughness, can be further deliberately investigated using interferometric, topographical and / or spectroscopic measurement methods, in particular interferometry, confocal microscopy, fluorescence spectroscopy or Raman spectroscopy. The processing of the sample can include, for example, a surface treatment of the sample, such as preparation and / or polishing in the measurement area. The further observation and / or processing methods are preferably carried out in sequence after the roughness measurement according to the invention. Alternatively, observations prior to the roughness measurement are also possible.

[0061] Particularly preferably, at least one subsequent second measurement method, preferably an interferometric, topographical and / or spectroscopic measurement method such as interferometry (in particular white light interferometry), confocal microscopy, fluorescence spectroscopy or Raman spectroscopy, is provided to purposefully measure each identified defect or structure again. Surface sections with increased roughness and / or increased deviations between at least two PSD functions and / or at least two integrated scatterings of the surface section are preferably investigated, since these deviations may indicate defects or especially contamination. The additional information and data that is made available as a result can then be used to obtain more accurate information (e.g. pass / fail) on structures and defects or to train and improve the evaluation accuracy of neural networks from the scattered light measurement data. Also, other product characteristics known from previous or subsequent measurements using other measurement methods, such as wear resistance, can be correlated with the data of the roughness measurement according to the invention, for example by adjusting or training the evaluation algorithm or neural network. In addition, it is also possible to expand the detectable spatial frequency range as a result.

[0062] In the case of a topography measurement method, at least one further measurement may be performed using a detector device and a modified focal plane is set. As a result, the topography of the surface can be reconstructed. For example, the local inclination angle of a surface structure or section relative to the lateral extent of the sample in a common measurement area in which the roughness measurement according to the invention is performed can be determined. As a result, it is advantageously possible to determine the local angles of incidence and scattering with high precision and to calculate at least one roughness feature, improving the accuracy of the evaluation.

[0063] Preferably, focus variation measurements (also referred to as the "shape from defocus" method) and / or Makyoh image measurements (Makyoh topography measurements) may be provided as topography measurement methods. In the case of focus variation measurements, preferably performed in dark field, a variation (scanning) of the focal plane of the detector device and a detection of the focal plane are performed where the individual image textures are the clearest, from which the height profile of the surface, in particular the local surface slope, is determined. In the case of Makyoh images, preferably performed in bright field, at least one camera image is acquired, preferably out of focus. The local sample curvature affects the intensity of the out-of-focus image, which makes it possible to calculate the height profile, in particular the local surface slope. For this calculation, for example, a homogeneous incident illumination distribution in the measurement area and a constant local reflection coefficient are assumed. Alternatively, the incident illumination distribution and the local reflection coefficient can also be calculated by calibrating the detector device using a sample with known reflection.

[0064] The features disclosed in relation to the measuring device and its embodiments also constitute preferred features of the method according to the invention and vice versa. Thus, the above-mentioned aspects and inventive and preferred features, in particular with regard to the construction of the measuring device and the dimensions and composition of the individual components described in relation to the measuring device, also apply to the method. The preferred embodiments, variants and features of the invention described above may be combined with one another.

[0065] Further details and advantages of the invention will now be explained with reference to the accompanying drawings. [Brief description of the drawings]

[0066] [Figure 1] 1 illustrates generally the features of a preferred embodiment of a measurement apparatus and method for roughness and / or defect measurement with time control of a light source; [Diagram 2] 2 shows a schematic diagram of a modified embodiment of the measurement apparatus according to FIG. 1, comprising a plurality of detector devices; [Diagram 3] 4 shows diagrammatically a further modified embodiment of the measurement device with intensity control of the light source; [Figure 4] 4 shows a schematic diagram of the function of the measuring device according to FIG. 3; [Diagram 5] 10A-10C are schematic diagrams illustrating angles and directions referenced when determining at least one roughness characteristic; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0067] Features of embodiments of the invention are described below with reference to roughness and / or defect measurements on a sample in the form of a plane ESC1 (part of which is shown diagrammatically) by way of example. It is emphasized that the application of the invention is not limited to measurements on the plane ESC1, but rather also includes measurements on other holding devices with burls extending over the carrier surface, as well as on other planes or curved components, such as optical mirrors or lenses or joining plates. In case of measurements on curved components, directional specifications such as surface normals refer locally to the respective directions in the current measurement area of ​​the measurement. When describing the determination of at least one roughness feature, reference is made to angles, directions and further variables shown in FIG. 5.

[0068] The drawings are schematic diagrams illustrating the features of embodiments of the invention: details of the described arrangements, such as the carrier of the light source, the camera housing or the sample holder, are not shown insofar as these are known per se from the prior art.

[0069] By way of example, reference is made to an embodiment of the invention in which the scattered light from the sample is detected in reflection, i.e. the illumination and detector devices are arranged in a common half-space on one side of the sample. The invention can be applied in a corresponding manner when the scattered light from the sample is detected in transmission, i.e. the illumination and detector devices are arranged in different half-spaces on either side of the sample.

[0070] The upper part of Fig. 1 shows diagrammatically a measurement apparatus 100 for roughness and / or defect measurements on an ESC1, which comprises an illumination device 10, a detector device 20 with a camera, an evaluation device 30 and a sample holder 40. The configuration of the evaluation device 30 as well as the method implemented by the evaluation device 30 for determining at least one roughness characteristic, including determining the roughness and / or identifying defects, are illustrated in the lower part of Fig. 1.

[0071] The evaluation device 30 comprises a computer unit configured to execute a data processing program for processing the scattered light image data recorded by the detector device 20. The evaluation device 30 is preferably also provided with a control unit for controlling the measurement arrangement 100. Alternatively or additionally, a further control unit (not shown) may be provided separate from the evaluation device 30 for controlling the measurement arrangement 100.

[0072] The ESC1 has the shape of a plane plate, on whose surface a number of burls 1A, each with an end face 2, are arranged. In the detailed view, the ESC1 is shown diagrammatically with only a few burls 1A. In practice, a dense arrangement of tens of thousands of burls 1A, for example with a spacing of less than 3000 μm, whose end faces 2 each have a diameter of less than 0.5 mm, is provided, spanning the carrier surface of the ESC1 for wafers. The ESC1 extends in a reference plane (here the XY plane) and the surface normal of the ESC1 extends perpendicular to the reference plane (here the Z direction). The end faces 2 are preferably located in a common plane parallel to the reference plane. The focal depth and axial chromatic aberration of the camera preferably allow sharp imaging at different wavelengths in both planes simultaneously. Alternatively, by refocusing, it is possible to use the camera to record both planes simultaneously and sharply. Optionally, the topography and / or local variations in the height of the surface of ESC1 may be detected, for example, by focus variation measurements and / or Makyoh image measurements.

[0073] The ESC1 is placed on a sample holder 40 (schematically shown), which for example comprises an xy stage for holding and moving the ESC in the XY plane. The position of the ESC1 relative to the illumination device 10 and the detector device 20, and in particular the measurement area 3 illuminated by the illumination device 10 within the entire surface of the ESC1, can be set by the sample holder 40.

[0074] The illumination device 10 comprises a group of light sources 11A, 11B, 11C, 11D, each of which comprises a light-emitting diode, for example of type M300L4 (340 nm wavelength, 53 mW power, manufacturer Thorlabs GmbH). The light sources 11A, 11B, 11C, 11D are fixed on a common carrier and / or housing (not shown) and each is provided with an illumination optics (collimation optics) 12. Optionally, at least one light source, and preferably all light sources, can be provided with a displaceable reducer, such as a gray filter (see FIG. 3). The reducer can be retracted and extended in the beam path of the respective light source in order to increase the dynamics of the measurement.

[0075] The illumination optics 12 of the light sources 11A, 11B, 11C, 11D each include a collimation lens for forming a collimated illumination beam path LA, LB, LC, and LD toward the ESC1. The light sources 11A, 11B, 11C, 11D and associated illumination optics 12 are arranged such that each of the illumination beam paths LA, LB, LC, and LD is incident at a different angle of incidence θ with respect to the surface normal of the surface of the ESC1. i and are arranged to be fixed so as to form a common measuring area 3. Due to the illumination optics 12 used, the illumination in the illumination beam paths LA, LB, LC and LD is typically circular or elliptical and larger than the considered measuring area 3. The measuring area 3 is for example rectangular or square in shape adapted to the shape of the detector array 21 of the detector device 20, the measuring area 3 having in particular a side length in the range of 0.1 cm to 10 cm.

[0076] Although four light sources are shown by way of example, in practice further light sources can be provided, for example ten or more light sources in total. The angles of incidence can be arranged to be distributed, for example, in the range of 2° to 75°. Within this range, the angles of incidence are preferably distributed to allow as uniform a scanning of the detected spatial frequency as possible. Thus, according to equation (1) (see below), a smaller angle of incidence leads to a higher density of light sources. If scattering structures with specific spatial frequencies are expected on the sample, the angles of incidence can preferably be set according to the range of expected spatial frequencies.

[0077] The light sources 11A, 11B, 11C, 11D together with the associated illumination optics 12 have a fixed position and orientation relative to the detector device 20 and thus also relative to the XY plane. This fixed position and orientation is set at least during operation of the measurement apparatus 100 and for the duration of all scattered light image recordings in order to determine the at least one roughness feature. For the purpose of adapting to corrected measurement conditions, the light sources 11A, 11B, 11C, 11D together with the associated illumination optics 12 may be releasably fixed on a carrier or in a housing and may be adjustable.

[0078] The illumination device 10 and the sample holder 40 are connected to a control unit of the evaluation device 30 (see double-headed arrow) or to a separate control unit, which provides time control or optionally intensity control of the light sources 11A, 11B, 11C, 11D and the positioning of ESC1, as will be explained in more detail below.

[0079] The detector device 20 comprises a camera with a detector array 21, for example of the OrcaFusion type (manufacturer Hamamatsu), and imaging optics 22, which are preferably telecentric optics. The detector device 20 has a fixed viewing angle θ sThe field of view is preferably 0°, i.e. the optical axis of the detector device 20 extends parallel to the surface normal (Z direction). Alternatively, the field of view may deviate from 0° (see for example FIG. 2).

[0080] A detector array 21, such as a CMOS matrix, comprises a number of detector pixels for spatially resolved recording of scattered light when the measurement area 3 is illuminated along the illumination beam paths LA, LB, LC and LD. The output signal of the detector array 21 is a number of scattered light images 4A, 4B, 4C, 4D, ... of the same measurement area 3, the number of these images being at least equal to the number of light sources in an embodiment with time control of the illumination devices. When recording an illumination series, for example in order to increase the reliability of the identification of deterministic structures or defects, the number of scattered light images 4A, 4B, 4C, 4D is preferably equal to an integer multiple of the number of light sources. The illumination series can be configured, for example, with different integration times in order to increase the dynamics and thus for example to obtain HDR (High Dynamic Range) images. The scattered light images 4A, 4B, 4C, 4D are fed to an evaluation device 30, by which a method for determining at least one roughness feature is performed.

[0081] According to a first embodiment of the invention, a time control of the light sources 11A, 11B, 11C, 11D is provided in such a way that for each angle of incidence, scattered light images 4A, 4B, 4C, 4D, ... are recorded separated in time from one another. During the recording of the scattered light images 4A, 4B, 4C, 4D, ... the position of the detector device 20 remains unchanged. The light sources 11A, 11B, 11C, 11D are successively activated by the evaluation device 30. In each case one of the light sources is activated for a predefined illumination period, for example 40 ms, while all the other light sources are deactivated. In each illumination phase using one of the light sources, one scattered light image is recorded and transmitted to the evaluation device 30. For example, a 25×25 mm 2Ten scattered light images 4A, 4B, 4C, 4D, . . . of the measuring field can be recorded within just 0.5 seconds, these images having different illumination beam paths LA, LB, LC, LD, .

[0082] Each scattered light image is an amplitude image characterized in that the scattered light amplitudes recorded individually at the detector pixels are determined in each case by light scattering at a structure such as a surface patch 2, a particle 5 or a fiber 6, and have a common spatial frequency f according to [6]. f = [sin(θ s )-sin(θ i )] / λ (1) (See Fig. 5. θ s : viewing angle or scattering angle, θ i : incident angle, λ: wavelength of illumination.) The spatial frequency f can alternatively be calculated by:

[0083]

number

[0084] f x and f y The following applies to: f x =(sinθ s cosΦ s -sinθ i ) / λ and f y =(sinθ s sinΦ s ) / λ (3) Φ s extends perpendicular to the surface normal and has an incidence angle θ i is the azimuthal angle with respect to an axis lying in the plane of incidence subtended by the direction of illumination and the surface normal (see FIG. 5). Therefore, the spatial frequency f x and f y are related to the coordinates of the scattering geometry defined by the direction of the incident beam. If the incident beam is not aligned on the X-axis of the sample, i.e., around the surface normal, at an azimuthal angle Φ iIf the image is rotated by about 100 degrees, the spatial frequency is the sample spatial frequency f x and f y Rotating to transforms it as follows: f X =f x cosΦ i +f y sinΦ i ,f Y =-f x sinΦ i +f y cosΦ i

[0085] The scattered light amplitude in the scattered light image is defined by the angle-resolved scattering ARS according to [6]. ARS(θ s ,Φ s ,θ i )=ΔP s (θ s ,Φ s ,θ i ) / P i ΔΩ (4) ΔΩ represents the solid angle, and ΔP s represents the scattering power, and P i represents the incident power (see Figure 5).

[0086] Using the evaluation device 30, PSD functions of the surface sections 2 in the illuminated measurement area 3 are determined from the scattered light images 4A, 4B, 4C, 4D, ... and these functions are output as roughness features of the associated surface sections 2 or at least one derived roughness feature of the associated surface sections 2, such as an rms value, is calculated from each PSD function.

[0087] Alternatively, the integrated scattering S of the surface section 2 within the illuminated measurement area 3 is determined from the scattered light images 4A, 4B, 4C, 4D, ... and at least one derived roughness feature of the associated surface section 2, such as an rms value, is calculated from each S value. Alternatively or additionally, the evaluation device 30 identifies and characterizes defects on the ESC surface from the scattered light images 4A, 4B, 4C, 4D, ...

[0088] In detail, in step S1 (bottom of FIG. 1), using image recognition methods available per se, optionally using neural networks, it is determined whether the scattering structure is a deterministic structure, i.e. a surface patch 2. If not, in step S2, it is determined by image recognition whether the scattering structure contains a defect. In this case, a classification according to expected defects such as particles 5, fibers 6, material breaks or cracks is performed in order to later characterize the relevant defects more accurately. Step S2 therefore comprises several sub-steps, of which FIG. 1 shows, purely by way of example, the sub-step of identifying whether the defect is formed by a particle.

[0089] When using neural networks for image recognition, the neural networks can be continuously adapted by machine learning. For this purpose, information from the topography review process Rev (see below) and / or from other local and / or global measurements (such as the wear resistance of the sample) can be used. As a result, it is also possible to predict subsequent product properties that are only indirectly related to the measurements. In addition, as a result, it is possible to continuously improve the accuracy of the derivation of local product properties (such as the surface roughness of the structure, or the size of the defects).

[0090] If, after steps S1 and S2, the scattering structure has not been identified as a surface patch 2 or a defect, the associated data may be discarded and the image recording may be repeated and / or the scattering structure may be approached individually and investigated using a different measurement method and / or a higher spatial resolution.

[0091] In steps S1 and S2, if a deterministic structure or defect is identified by checking the quantitative degree of error with a sufficient level of confidence that no further image recording is required for subsequent calculation of at least one roughness feature, the light source can be controlled accordingly so that no further illumination is performed in order to save measurement time.

[0092] After the identification of the surface sections 2 in step S1, these are cut out by using a digital mask in the scattered light images 4A, 4B, 4C, 4D, ... (selecting the structures according to their local position on the sensor matrix) in steps S3 and S4, then in steps S5 and S6, respectively, to determine the ARS value and from there to calculate the PSD function of the scattering structures on the surface sections 2 taking into account the field geometry (solid field angle Ω of the detector device) in steps S7 and S8. The ARS value is obtained directly from the detected scattered light amplitude, i.e. the output signal of the detector pixels.

[0093] The relationship between the PSD function of each pixel in a selected structure, depending on the spatial frequency f and the ARS value, is given by the following equation (see for example [3]):

[0094]

number

[0095] In equation (5), Q is an optical factor that contains information about the scattering geometry (angle of incidence and scattering, light polarization) and material properties (complex refractive index) of the sample being investigated ([6], [8]). Alternatively, Q can also be approximated by the reflection coefficient, which can be determined in a locally resolved manner from a reflectance measurement of the sample by detecting the direct reflection using the measurement methods described ([6], [8]).

[0096] The PSD function can be determined by measurements at at least two support points, i.e. at different spatial frequencies f or different angles of incidence, and then extrapolation to the entire spatial frequency range of interest. Extrapolation is advantageously possible because for polished and coated surfaces, the PSD function generally extends continuously and without abrupt jumps, and therefore only a few sampling points are sufficient.

[0097] The rms roughness σ, as a roughness characteristic of the relevant surface section, can be calculated from the PSD function according to the following formula:

[0098]

number

[0099] (See, for example, [4].) Alternatively, the roughness may be calculated, for example, from the average PSD function within a surface section.

[0100] After identification of defects 5 in step S2, these defects are cut out in steps S9 and S10 and then in step S11 from the image data by using a digital mask in the scattered light images 4A, 4B, 4C, 4D, ... to calculate, for example, the number of detector pixels covering the defect and the particle size from the imaging properties of the detector device. Alternatively, the identified small particles may be calculated from the scattered light images using a particle scattering model such as, for example, [7].

[0101] Alternatively or in addition to the above-mentioned use of a digital mask adapted to the deterministic structures to be cut out (surface sections 2), the analysis mask in the evaluation device 30 may also be adapted to cover the image sections between the deterministic structures.

[0102] 1 shows diagrammatically that at least one review process Rev can be performed using the evaluation device 30 in order to identify further information related to the surface patch or defect. The at least one review process Rev is performed depending on the result of the identification of the at least one roughness feature, preferably by topographical measurement methods such as white light interferometry and / or AFM. Alternatively or additionally, at least one review process Rev can be provided that provides additional or more precise information based on the method according to the invention, where for example further scattering angles and / or further wavelength and / or phase information are evaluated.

[0103] The roughness and / or defect characteristics within the measurement area 3 are recorded and optionally output as results. To measure the entire surface of the ESC1, the ESC1 together with the sample holder 40 is repeatedly moved relative to the illumination device 20 and the detector device 30 and measured at each position.

[0104] Figure 2 shows a modified variant of the embodiment according to figure 1, in which in addition to the detector device 20, a further detector device 20A is provided with a detector array 21A and an imaging optical element 22A. By way of example, an illumination device 10 is shown with two light sources 11A, 11B, each with an imaging optical element 12, but in practice more light sources can be provided. The light sources 11A, 11B illuminate the measurement area 3 alternately and successively at two different angles of incidence. An evaluation device (not shown in figure 2) is provided for processing the scattered light images, as explained with reference to figure 1. By way of example, a tilt of the detector array 21A is shown, which can be provided for uniform focus (Scheimpflug principle).

[0105] The first detector device 20 is arranged with a detector array 21 and imaging optics 22 at a field angle of 0°, as described with reference to FIG. 1, so as to record at least two scattered light images. The further detector device 20A is arranged at a different field angle, for example 60°, relative to the surface normal of the surface. According to the described time control, the light sources 11A, 11B are alternately activated to record scattered light images at different angles of incidence. The first detector device 20 and the further detector device 20A record two scattered light images of the illuminated measurement area 3 at different field angles for each angle of incidence. Thereby it is advantageously possible to detect scattering features with high spatial frequency. The recording of the scattered light images by the first detector device 20 and the further detector device 20A is preferably performed simultaneously.

[0106] Alternatively or additionally, the detector devices may have different spectral sensitivities. This advantageously allows the use of illumination wavelengths in a larger spectral range. Also, multiple illuminations with different wavelengths may be performed simultaneously, whereby a separation of wavelengths occurs due to the different spectral sensitivities of the detector devices. Furthermore, beam splitting optics would allow the use of multiple detector devices with the same viewing angle but with different spectral or polarization sensitivities. This kind of separation is advantageous for improving the measurement speed or for improving the sensitivity to individual structures or defects, or for improving the ability to distinguish between different structures or defects.

[0107] According to a second embodiment of the invention illustrated in Figures 3 and 4, scattered light images are recorded simultaneously and intensity control of the light sources 11A, 11B, 11C is provided such that the illumination device 20 provides different illumination intensities within the measurement area 3 via the illumination beam path.

[0108] Direct calculation of the rms roughness σ of each pixel from the total scattering TS by an evaluation device (not shown in FIG. 3) is performed according to the following formula:

[0109]

number

[0110] (See [5].) TS is the total hemispherical scattering power P TS and the incident power P i Calculated from (TS=P TS / P i ), where R represents the reflection coefficient of the surface being investigated. The total scattering is calculated for polar scattering angles θ at least in the angular range of 2° to 85°. s , taken according to ISO 13696. However, the roughness measurements described here apply equally well to smaller angular ranges, in which case the measurement variable is conventionally referred to as integrated scatter.

[0111] The detector device 20 is used to obtain a scattered light image having a scattered light amplitude proportional to the total or integrated scattering, where the irradiance E of individual instances of illumination is calculated as E=ΔP i A weighting G of / ΔA is performed, where ΔP i represents the power incident on a surface element ΔA in the measurement area 3. The surface element ΔA may correspond to a detector pixel. As a result, all the instances of illumination can illuminate the sample (ESC1) simultaneously. The angular resolved information of the scattering distribution ARS(θ s ,φ s ) is lost and so the PSD function can no longer be calculated, but the proportionality to the total scattering means that time control of the light sources 11A, 11B, 11C is no longer required and thus the roughness measurement can be performed with a single simultaneous image recording upon illumination at all angles of incidence. Advantageously, measurement times can be significantly reduced as a result.

[0112] The determination of the at least one roughness feature in a second embodiment with light source intensity control (FIG. 3) is based on a modification of the determination of the total scattering known from the prior art, as explained below.

[0113] In the prior art, total scattering is determined by quasi-perpendicular illumination and direct hemispherical detection. Hemispherical detection is typically achieved using, in technical terms, a Koblenz or Ulbricht sphere [8], whereby light is integrally collected and deflected to the detector. Alternatively, total scattering is determined by numerical integration from the measured data of angle-resolved scattering, in the case of quasi-perpendicular light incidence and scanning of the scattered light hemisphere [8].

[0114]

number

[0115] In the case of an isotope scattering surface, the ARS is rotationally symmetric about the surface normal when viewed vertically by the camera. Then, the azimuthal angle φs It is sufficient to determine the angle resolved scattering ARS at , where

[0116]

number

[0117] In contrast, when determining total scattering in the second embodiment of the present invention, the incident and detection directions are swapped. When the propagation directions are swapped, the ARS becomes ARS(θ s )=ARS(θ i )cos(θ i ) / cos(θ s ) After rearrangement, it becomes:

[0118]

number

[0119] The scattering distribution ARS of the sample is detected by each round of illumination at angles θ and Φ, or only θ in the one-dimensional case. s = 0°, using the detector device 20 to look vertically, and 0>θ1>θ N And the incidence angles θ1 to θ N (see FIG. 4A) and scanning the ARS with N illuminations with Φ=0, the integral of TS is calculated according to the following formula:

[0120]

number

[0121] where Δθ is the representative full angle of illumination of interest (see FIG. 4a) or the step angle between multiple illuminations.

[0122] To achieve a measurement signal proportional to TS, the signal ARS N It is clear that the measurement signals ARS(θ NThis weighting of the irradiances is carried out in the measurement method by weighting the individual irradiances. The weighting coefficients are then calculated according to the following formula: G(θ N )=cosθ N sinθ N Δθ N (12)

[0123] This is the irradiance E = ΔP of each light source for each illumination of the surface element ΔA in the measurement area 3. i With respect to / ΔA, the following equation is obtained: E(θ N )=E0G=E0cosθ N sinθ N Δθ N (13) Here, E0 is the irradiance, preferably selected so that the overall illumination of the surface elements within the measurement area 3 results in a favorable signal-to-noise ratio upon camera detection. N If multiple light sources are used in each case and the light sources are arranged at different azimuth angles, the light sources should preferably have the same irradiance, and the total irradiance E(θ N ) also corresponds to Equation 13.

[0124] To calculate the TS values ​​from the detector measurement signals of samples with the described illumination weighting, calibration samples with known TS values ​​are used, as known from the prior art.

[0125] For scattering angles around the specular reflection, the ARS typically extends very steeply, so in this region it is advantageous to keep the solid angle as small as possible to reduce measurement uncertainty. At wide angles to the specular reflection, the ARS extends more shallowly and the scattering power is significantly lower. In this region it is advantageous to choose the scan spacing and solid angle to be large.

[0126] In the measurement area (3), each angle θ NIn order to provide uniform illumination of the sample (see FIG. 4B), beam-shaping optics or position-variable reducers 13, e.g. grey filters, or corresponding operation of light sources with different energies, i.e. different powers, pulse lengths or pulse width modulation may be provided.

[0127] The embodiment described in Figure 3 allows for measuring rms roughness having a spatial frequency spectrum determined by the minimum and maximum angles of incidence and the wavelength λ1 of the light source (see Equation 1). To extend the spatial frequency range, additional illumination wavelengths λ N can be used which also simultaneously illuminates the sample with an existing wavelength.

[0128] Since the calculation of the rms roughness is wavelength-dependent (see above), in this case the illumination intensities of the multiple illuminations with different wavelengths are also adapted to one another. Since the sample-specific reflection coefficient also depends on the measurement wave, this too is preferably corrected by wavelength adjustment. When using multiple measurement wavelengths, the following additional weighting factors are applied for the individual illuminations:

[0129]

number

[0130] If the detector device 20 also has a spectral sensitivity, this is corrected accordingly.

[0131] In the following preferred features of the present invention are summarized.

[0132] A measurement apparatus (100) configured for roughness and / or defect measurements on a plurality of surface sections (2) of a surface of a sample (1) to be investigated comprises an illumination device (10) with at least two light sources (11A, 11B, 11C, 11D) arranged to illuminate a measurement area (3) of the surface with measurement light, a detector device (20) with a detector array (21) with a plurality of detector pixels arranged to capture scattered light scattered at the surface, and an evaluation device (30) configured to determine at least one roughness characteristic of the surface from the captured scattered light, wherein the at least two light sources (11A, 11B, 11C, 11D) are configured to illuminate the measurement area (3) along at least two illumination beam paths (LA, LB, LC, LD) at different angles of incidence with respect to a surface normal of the surface, and wherein the at least two light sources (11A, 11B, 11C, 11D) are configured to illuminate the measurement area (3) along at least two illumination beam paths (LA, LB, LC, LD) at different angles of incidence with respect to a surface normal of the surface, , 11B, 11C, 11D) can be fixed relative to a detector device (20), the detector device (20) being provided with imaging optics (22) arranged to image a measurement area (3) of the surface onto the detector array (21), the detector device (20) being configured to capture at least two scattered light images (4A, 4B, 4C, 4D) of a surface section (2) within the measurement area (3) illuminated at a predetermined field of view angle relative to a surface normal of the surface, the portion of the scattered light received by the detector pixels and formed in each case by the illumination in one of the illumination beam paths (LA, LB, LC, LD) having a common spatial frequency in each case, and the evaluation device (30) is configured to determine at least one roughness feature of the surface section (2) from the at least two scattered light images (4A, 4B, 4C, 4D).

[0133] The illumination device (10) may be configured to time-control the at least two light sources (11A, 11B, 11C, 11D) such that at least two scattered light images (4A, 4B, 4C, 4D) can be recorded successively and temporally separated from one another, and the evaluation device (30) is configured to determine a power spectral density function (PSD function) of the surface sections (2) in the illuminated measurement area (3) from the at least two scattered light images (4A, 4B, 4C, 4D), and the evaluation device (30) is configured to calculate at least one roughness feature for each surface section (2) from the PSD function or from the integrated scattering of the surface section (2).

[0134] The illumination device (10) may be configured for intensity control such that at least two scattered light images (4A, 4B, 4C, 4D) can be recorded in each of the surface sections (2) with a weighted irradiance of directed illumination by at least two light sources (11A, 11B, 11C, 11D), the weighted irradiance of directed illumination being provided in that the irradiance is set such that the illumination by the light sources corresponds to uniform hemispherical illumination, and the evaluation device (30) is configured to determine an integrated scattering of the structured surface within the illuminated measurement area (3) from the at least two scattered light images (4A, 4B, 4C, 4D), and the evaluation device (30) is configured to calculate at least one roughness feature for each surface section (2) from the integrated scattering of the surface section (2).

[0135] Preferably, at least one of the following features is provided: the number of light sources (11A, 11B, 11C, 11D) of the illumination device (10) is selected to be in the range of 3 to 100, the illumination device (10) is configured to generate the measurement light to be of the same wavelength in all illumination beam paths (LA, LB, LC, LD), and the illumination device (10) is configured to generate the measurement light to be of different wavelengths and / or different polarizations in the illumination beam paths (LA, LB, LC, LD). If the illumination device (10) is configured to generate the measurement light to be of different wavelengths and / or different polarizations in the illumination beam paths (LA, LB, LC, LD), the illumination device (10) is preferably configured to generate the measurement light to be of at least two different wavelengths and / or at least two different polarizations in at least one of the illumination beam paths (LA, LB, LC, LD).

[0136] The illumination device (10) may be configured to generate measurement light to be at different wavelengths in the illumination beam paths (LA, LB, LC, LD), and the evaluation device (30) is configured to determine at least two PSD functions and / or at least two integrated scatterings of a surface section (2) within the illuminated measurement area (3), and the evaluation device (30) is configured to identify a structure on the surface section (2) from the at least two PSD functions and / or the at least two integrated scatterings.

[0137] Preferably, the evaluation device (30) is configured to identify the surface section (2) in the measurement area (3) by using a digital image mask. Preferably, the evaluation device (30) is configured to identify the surface section (2) by using neural networks and / or machine learning methods and / or the evaluation device (30) is configured to identify structures on the surface section (2) from the at least two scattered light images (4A, 4B, 4C, 4D) by using a digital image mask.

[0138] The imaging optics (22) of the detector device (20) may be telecentric optics and / or the field of view of the detector device (20) may extend parallel to the surface normal.

[0139] Preferably, the illumination device (10) is configured for directed illumination of the surface section (2) at a shallow angle of incidence in the range of 88° to 75° using measurement light having a wavelength λ smaller than 300 nm, and the evaluation device (30) is configured for determining a distortion parameter of the surface section (2).

[0140] At least one further detector device (20A) may be provided comprising an imaging optical element (22A) and a detector array (21A) having a plurality of detector pixels, the at least one further detector device (20A) configured to capture at least one scattered light image (4A, 4B, 4C, 4D) of the illuminated measurement area (3) at at least one further field of view angle relative to a surface normal of the surface.

[0141] Preferably, the at least two light sources (11A, 11B, 11C, 11D) comprise at least one of a light emitting diode, a laser, and an optical fiber coupled to at least one light emitting diode and / or at least one laser, and / or each light source is provided with illumination optics (12).

[0142] Optionally, the sample holder (40) for receiving the sample (1) and / or the assembly consisting of the illumination device (10) and the detector device (20) are movable relative to each other, whereby the illumination device (10) can be positioned to illuminate different measurement areas of the structured surface, scattered light images (4A, 4B, 4C, 4D) of the different measurement areas can be captured by the detector device (20), and the measurement apparatus (100) is configured to perform repeated roughness measurements on surface sections (2) within the different measurement areas of the structured surface.

[0143] The illumination device (10) and detector device (20, 20A) may be configured to capture scattered light from the sample (1) in reflection and / or transmission.

[0144] The method for roughness and / or defect measurement on a surface section (2) of a surface of a sample to be investigated comprises the steps of illuminating a measurement area (3) of the surface with measurement light, the measurement area (3) encompassing a plurality of surface sections (2), capturing scattered light scattered at the surface using a detector device (20) having a detector array (21) with a plurality of detector pixels, and determining at least one roughness characteristic of the surface from the captured scattered light, the illumination being at different incidences with respect to a surface normal of the surface along at least two illumination beam paths (LA, LB, LC, LD) using at least two fixedly positioned light sources (11A, 11B, 11C, 11D). The invention relates to a method for determining a roughness characteristic of a surface section (2) in a measurement area (3) at a predetermined viewing angle, the method comprising the steps of: (a) providing a directed illumination of a surface section (2) in a measurement area (3) at a predetermined viewing angle, the detector device (20) being provided with imaging optics (22) whereby the measurement area (3) of the surface is imaged onto the detector array (21), the detector device (20) capturing at least two scattered light images (4A, 4B, 4C, 4D) of the illuminated measurement area (3) at a predetermined viewing angle relative to a surface normal of the surface, the portion of the scattered light received by the detector pixels in each of the illumination beam paths (LA, LB, LC, LD) having a common spatial frequency in each case, and at least one roughness characteristic being determined for each surface section (2) from the at least two scattered light images (4A, 4B, 4C, 4D).

[0145] Time control of the at least two light sources (11A, 11B, 11C, 11D) may be provided such that at least two scattered light images (4A, 4B, 4C, 4D) are recorded temporally separated from one another, a power spectral density function (PSD function) of a surface section (2) within the illuminated measurement area (3) is determined from the at least two scattered light images (4A, 4B, 4C, 4D), and at least one roughness feature is calculated for each surface section (2) from the PSD function or the integrated scattering of the surface section (2).

[0146] Intensity control may be provided such that at least two scattered light images (4A, 4B, 4C, 4D) are recorded at each of the surface sections (2) with a weighted irradiance of directed illumination by at least two light sources (11A, 11B, 11C, 11D), the weighted irradiance of directed illumination being provided in that the irradiance is set so that the light sources produce uniform hemispherical illumination, and an integrated scattering of the structured surface within the illuminated measurement area (3) is determined from the at least two scattered light images (4A, 4B, 4C, 4D), and at least one roughness feature is calculated for each surface section (2) from the integrated scattering of the surface section (2).

[0147] The measurement light of all illumination beam paths (LA, LB, LC, LD) may be generated to be at the same wavelength.

[0148] The measurement light of the illumination beam paths (LA, LB, LC, LD) may be generated to be of different wavelengths and / or of different polarizations. Preferably, the measurement light of at least one of the illumination beam paths (LA, LB, LC, LD) is generated to be of at least two different wavelengths and / or of at least two different polarizations, and / or at least two PSD functions and / or at least two integrated scatterings of the surface section (2) in the illuminated measurement area (3) are determined, and a structure on the surface section (2) is obtained from the at least two PSD functions and / or the at least two integrated scatterings.

[0149] Identification of the surface patch (2) in the at least two scattered light images (4A, 4B, 4C, 4D) may be provided. Preferably, the surface patch (2) is identified by using neural networks and / or machine learning methods and / or structures on the surface patch (2) are identified from the at least two scattered light images (4A, 4B, 4C, 4D) by using digital image masks.

[0150] The surface section (2) may be illuminated at a shallow angle of incidence in the range of 88° to 75° using measurement light having a wavelength λ smaller than 300 nm, and the skewness parameters of the surface section (2) are determined from the scattered light images (4A, 4B, 4C, 4D).

[0151] Preferably, at least one further detector device (20) may be provided comprising an imaging optical element (22) and a detector array (21) having a plurality of detector pixels, and at least one scattered light image (4A, 4B, 4C, 4D) of the illuminated measurement area (3) at at least one further field angle relative to the surface normal of the surface is captured.

[0152] A number of different measurement areas of the structured surface may be illuminated and scattered light images (4A, 4B, 4C, 4D) of the different measurement areas are acquired by the detector device (20), and the roughness measurement at the surface section (2) is repeated within the different measurement areas of the structured surface.

[0153] Preferably, selected surface sections, in particular those having increased roughness and / or increased deviation between at least two PSD functions and / or at least two integrated scatterings of the surface section (2), are purposefully investigated using interferometric, topographical and / or spectroscopic measurement methods, in particular interferometry, confocal microscopy, fluorescence spectroscopy or Raman spectroscopy.

[0154] The surface or a selected surface section may be subjected to focus variation measurements and / or Makyoh image measurements to determine the height profile, and in particular the local surface slope, of the surface.

[0155] Preferably, the sample to be investigated (1) comprises a holding device for holding a component, a plurality of protruding burls being arranged on a surface of the holding device, and the surface section (2) comprises the end faces of the burls.

[0156] Light scattered from the sample (1) can be captured in reflection and / or transmission.

[0157] The features of the invention disclosed in the above description, in the drawings and in the claims may be important individually, in combination or in subcombinations for the realization of various embodiments of the invention.

Claims

1. A measuring device (100) configured for roughness and / or defect measurements on a plurality of surface sections (2) of a surface of a sample (1) to be investigated, comprising: an illumination device (10) having at least two light sources (11A, 11B, 11C, 11D) arranged to illuminate a measurement area (3) of said surface with measurement light; a detector device (20) having a detector array (21) with a plurality of detector pixels arranged to capture scattered light scattered at the surface; an evaluation device (30) configured to determine at least one roughness characteristic of the surface from the captured scattered light, the at least two light sources (11A, 11B, 11C, 11D) are configured to illuminate the measurement area (3) along at least two illumination beam paths (LA, LB, LC, LD) at different angles of incidence with respect to a surface normal of the surface, the at least two light sources (11A, 11B, 11C, 11D) can be fixed relative to the detector device (20); the detector device (20) is provided with imaging optics (22) arranged to image the measurement area (3) of the surface onto the detector array (21); the detector device (20) is configured to capture at least two scattered light images (4A, 4B, 4C, 4D) of a surface section (2) within the illuminated measurement area (3) at a predetermined viewing angle relative to the surface normal of the surface, the portion of the scattered light received by the detector pixel and formed in each case by the illumination in one of the illumination beam paths (LA, LB, LC, LD) has a common spatial frequency in each case, the evaluation device (30) is configured to determine the at least one roughness characteristic for the surface section (2) from the at least two scattered light images (4A, 4B, 4C, 4D). A measuring device (100).

2. the illumination device (10) is configured for time-controlling the at least two light sources (11A, 11B, 11C, 11D) such that the at least two scattered light images (4A, 4B, 4C, 4D) can be recorded successively and separated in time from one another, the evaluation device (30) is configured to determine a power spectral density function (PSD function) of the surface patch (2) within the illuminated measurement area (3) from the at least two scattered light images (4A, 4B, 4C, 4D), 2. The measuring apparatus according to claim 1, wherein the evaluation device (30) is configured to calculate the at least one roughness feature for each surface section (2) from the PSD function or from the integrated scattering of the surface section (2).

3. the illumination device (10) is configured for intensity control such that the at least two scattered light images (4A, 4B, 4C, 4D) can be recorded at each of the surface sections (2) with a weighted irradiance of the directed illumination by the at least two light sources (11A, 11B, 11C, 11D), the weighted irradiance of the directed illumination is provided in that the irradiance is set such that the illumination by the light source corresponds to a uniform hemispherical illumination; the evaluation device (30) is configured to determine an integrated scattering of the structured surface within the illuminated measurement area (3) from the at least two scattered light images (4A, 4B, 4C, 4D), 3. The measuring apparatus according to claim 1 or 2, wherein the evaluation device (30) is configured to calculate the at least one roughness feature for each surface section (2) from the integrated scattering of the surface section (2).

4. the number of light sources (11A, 11B, 11C, 11D) of the lighting device (10) is selected to be in the range of 3 to 100; The illumination device (10) is configured to generate the measurement light of all the illumination beam paths (LA, LB, LC, LD) to be at the same wavelength; and the illumination device (10) is configured to generate the measurement light in the illumination beam paths (LA, LB, LC, LD) to be of different wavelengths and / or different polarizations; 4. The measuring device according to claim 1, further comprising at least one of the following features:

5. the illumination device (10) is configured to generate the measurement light in the illumination beam paths (LA, LB, LC, LD) to be of different wavelengths and / or different polarizations, 5. The measuring apparatus of claim 4, wherein the illumination device (10) is configured to generate the measuring light in the illumination beam path (LA, LB, LC, LD) to be at least two different wavelengths and / or at least two different polarizations.

6. The illumination device (10) is configured to generate the measurement light in the illumination beam paths (LA, LB, LC, LD) to be at different wavelengths, the evaluation device (30) is configured to determine at least two PSD functions and / or at least two integrated scattering functions of the surface section (2) within the illuminated measurement area (3), 6. The measuring arrangement according to claim 4 or 5, wherein the evaluation device (30) is configured to determine a structure on the surface section (2) from the at least two PSD functions and / or the at least two integrated scattering functions.

7. 7. The measuring apparatus according to any one of the preceding claims, wherein the evaluation device (30) is configured to identify the surface section (2) within the measurement area (3) by using a digital image mask.

8. 8. The measuring apparatus according to claim 7, wherein the evaluation device (30) is configured to identify the surface section (2) by using neural networks and / or machine learning methods.

9. 9. The measuring apparatus according to claim 7 or 8, wherein the evaluation device (30) is configured to identify structures on the surface section (2) from the at least two scattered light images (4A, 4B, 4C, 4D) by using a digital image mask.

10. the imaging optics (22) of the detector device (20) are telecentric optics; and / or 10. The measuring apparatus according to claim 1, wherein the field of view of the detector device (20) extends parallel to the surface normal.

11. the illumination device (10) is configured for directed illumination of the surface section (2) at a shallow angle of incidence in the range of 88° to 75° using measurement light having a wavelength λ smaller than 300 nm; 11. The measuring apparatus according to any one of the preceding claims, wherein the evaluation device (30) is configured to determine a distortion parameter of the surface plot (2).

12. At least one further detector device (20A) is provided, comprising an imaging optics (22A) and a detector array (21A) having a plurality of detector pixels, 12. The measuring apparatus according to claim 1, wherein the at least one further detector device (20A) is configured to capture at least one scattered light image (4A, 4B, 4C, 4D) of the illuminated measurement area (3) at at least one further viewing angle relative to the surface normal of the surface.

13. the at least two light sources (11A, 11B, 11C, 11D) comprise at least one of a light emitting diode, a laser, and an optical fiber coupled to at least one light emitting diode and / or to at least one laser; and / or 13. Measuring device according to any one of the preceding claims, wherein each light source is provided with illumination optics (12).

14. a sample holder (40) for receiving the sample (1) and / or the assembly consisting of the illumination device (10) and the detector device (20) are movable relative to one another, whereby the illumination device (10) can be positioned to illuminate different measurement areas of the structured surface, The scattered light images (4A, 4B, 4C, 4D) of the different measurement areas may be captured by the detector device (20); The measuring device (100) according to any one of the preceding claims, wherein the measuring device (100) is configured to perform repeated roughness measurements on the surface sections (2) in the different measurement areas of the structured surface.

15. A method for roughness and / or defect measurements on a surface section (2) of the surface of a sample to be investigated, comprising the steps of: - illuminating a measurement area (3) of the surface with measurement light, the measurement area (3) encompassing the plurality of surface sections (2); capturing scattered light scattered at said surface using a detector device (20) having a detector array (21) with a plurality of detector pixels; determining at least one roughness characteristic of the surface from the captured scattered light; said illumination comprises a directed illumination of said surface section (2) in said measurement area (3) along at least two illumination beam paths (LA, LB, LC, LD) at different angles of incidence with respect to a surface normal of said surface, using at least two fixedly positioned light sources (11A, 11B, 11C, 11D); The detector device (20) is provided with imaging optics (22) by means of which the measurement area (3) of the surface is imaged onto the detector array (21); the detector device (20) captures at least two scattered light images (4A, 4B, 4C, 4D) of the illuminated measurement area (3) at a predetermined viewing angle relative to the surface normal of the surface; the portion of the scattered light received by the detector pixels has a common spatial frequency in each of the illumination beam paths (LA, LB, LC, LD); said at least one roughness feature is determined for each surface section (2) from said at least two scattered light images (4A, 4B, 4C, 4D); A method comprising:

Images